1
00:00:02,120 --> 00:00:04,639
[Auto-generated transcript. Edits may have been applied for clarity.]
Well, ladies and gentlemen, my name is Richard Ovenden.

2
00:00:04,640 --> 00:00:10,969
I am both Bodley's Librarian and head of gardens, libraries and museums here in the University of Oxford.

3
00:00:10,970 --> 00:00:18,500
And it's my enormous pleasure to welcome you to this event, the second of our

4
00:00:18,500 --> 00:00:26,990
celebratory conferences that we've had around the ARCHiOx project and this project,

5
00:00:28,310 --> 00:00:33,950
the second phase of which comes to an end really today

6
00:00:33,950 --> 00:00:42,680
with this conference has been a collaboration with our close friends and colleagues, Factum Foundation in Madrid.

7
00:00:43,040 --> 00:00:49,370
And you will hear from and will see many of them, during the course of this event this afternoon.

8
00:00:49,760 --> 00:00:54,860
And, all I can say is that that has been an absolutely

9
00:00:54,860 --> 00:01:04,280
fantastic experience to work with such a talented and innovative group of colleagues,

10
00:01:04,280 --> 00:01:16,399
working in the the same space, but with a very different set of skills that they bring to complement those which we have here in the library

11
00:01:16,400 --> 00:01:20,720
and, I think, in the cultural heritage sector as a whole.

12
00:01:21,230 --> 00:01:27,290
And of course, none of this would be possible without the immense generosity of the Helen Hamlyn Trust.

13
00:01:27,290 --> 00:01:39,290
And I'd just like to say for the record, a huge vote of thanks and to express my very sincere gratitude to them for their confidence

14
00:01:39,680 --> 00:01:47,210
in us, in the Bodleian and in the Factum Foundation to take this project forward and to develop,

15
00:01:47,570 --> 00:01:53,059
a set of technologies, approaches and use cases which enable

16
00:01:53,060 --> 00:02:02,780
us to really transform the way in which we're able to interrogate the objects in our collections for the sake of creating new knowledge.

17
00:02:05,300 --> 00:02:08,530
And our thing today is seeing the unseen.

18
00:02:08,540 --> 00:02:12,260
And that's exactly what we've been doing over the past couple of years.

19
00:02:13,060 --> 00:02:18,260
And that's why we're excited to share the findings of the work that we've been doing,

20
00:02:18,260 --> 00:02:29,060
both the technologies and the techniques which we've developed. Across our GLAM institutions here in Oxford, that's gardens, libraries and museums,

21
00:02:29,270 --> 00:02:32,030
we have 32 million collection items,

22
00:02:32,390 --> 00:02:45,170
and ARCHiOx is helping us give new ways of both extracting the data from those collections and understanding how those objects were created,

23
00:02:45,590 --> 00:02:51,290
what they were made from, how, in some cases, when and where they were created.

24
00:02:51,650 --> 00:02:55,910
But also, and I think particularly for me this has been the most exciting thing,

25
00:02:56,150 --> 00:03:01,730
the human interactions with those collections after they were first created, after they were first formed.

26
00:03:02,150 --> 00:03:12,260
How can we learn how humans have responded to, intervened in, and left their marks on

27
00:03:12,950 --> 00:03:18,560
books, documents, artefacts, objects, all of those kinds of things.

28
00:03:18,990 --> 00:03:28,940
ARCHiOx helps us understand that in ways which are truly new, I think, and I hope you'll, if you haven't seen it before,

29
00:03:28,940 --> 00:03:33,710
you'll come away from this event with a new perspective on those issues.

30
00:03:35,520 --> 00:03:42,420
The extraordinary findings from the first phase of the project, which were announced in this very room about two years ago.

31
00:03:42,810 --> 00:03:49,860
included the discovery of the reading practices of an early medieval religious woman, Eadberg.

32
00:03:50,340 --> 00:03:56,550
And these have been matched in phase two with the discovery of William Blake's first foray into engraving.

33
00:03:57,610 --> 00:04:04,570
Given the close collaboration between the Bodleian and Oxford's four outstanding university museums,

34
00:04:04,690 --> 00:04:11,770
we were keen to experiment in this second phase to test the utility for ARCHiOx with museum collections.

35
00:04:12,580 --> 00:04:16,210
We were not disappointed, ladies and gentlemen, and you will see more about this

36
00:04:16,420 --> 00:04:24,760
this afternoon. We reveal previously unseen marks of creation from Raphael, Michelangelo and Giulio Romano in the Ashmolean.

37
00:04:25,180 --> 00:04:31,870
We explored the details on naturally made objects in the Natural History Museum, from meteorites to fossilised frogs,

38
00:04:32,440 --> 00:04:41,070
and have experimented with the built heritage in making high fidelity records of the 15th century fan vaulting on the ceiling of the Divinity School,

39
00:04:41,080 --> 00:04:47,110
the heart of the Bodleian. My own personal show and tell.

40
00:04:47,140 --> 00:04:57,410
Recording aspects of ARCHiOx continue to be extraordinary, as we hope to demonstrate to you this afternoon. And together,

41
00:04:57,410 --> 00:05:06,470
the Bodleian and Factum have been working to make using the technology as easy as possible within the environments of a library, archive or museum.

42
00:05:07,070 --> 00:05:11,180
The team, led by specialists in our digital library operation here in the Bodleian,

43
00:05:11,660 --> 00:05:16,310
in the Bodleian, have also been busy integrating ARCHiOx with the IIIF standard.

44
00:05:16,580 --> 00:05:25,510
Indeed, building out that standard to accommodate it so that the data that ARCHiOx creates can be easily and interoperable,

45
00:05:25,580 --> 00:05:31,700
shared from digital platforms which are IIIF compliant,

46
00:05:31,700 --> 00:05:38,630
and that is now encompassing a huge array of the library and museum sector across the globe.

47
00:05:40,390 --> 00:05:49,600
Of course. Our colleagues in the imaging studio, several floors below where we're sat have also had their own practices,

48
00:05:49,840 --> 00:05:59,469
and indeed their spaces, transformed by ARCHiOx and the team themselves led, absolutely superbly by John Barrett, have

49
00:05:59,470 --> 00:06:04,510
really grabbed hold of this opportunity with both hands and have

50
00:06:04,510 --> 00:06:09,040
had the pleasure not only of refining the techniques and working with scholars to do that

51
00:06:09,310 --> 00:06:18,310
to explore those use cases, but also increasingly to share it with other professionals in the library and museum sector.

52
00:06:18,310 --> 00:06:23,770
And it's wonderful to have a number of our colleagues, here this afternoon.

53
00:06:26,330 --> 00:06:29,880
Finally, allow me to read the horoscope of the project.

54
00:06:29,900 --> 00:06:36,770
If you would. We have another phase of work planned to further extend and enhance the technology,

55
00:06:37,070 --> 00:06:41,390
to refine the workflows at the institutional end of things,

56
00:06:41,630 --> 00:06:47,180
and especially to build a community of practice as more institutions acquire

57
00:06:47,180 --> 00:06:52,820
the technology and start to use it at an industrial scale to document their collections.

58
00:06:53,510 --> 00:07:00,350
The Bodleian, working with our colleagues at Factum, want to create the space in which the instructions,

59
00:07:00,920 --> 00:07:06,890
the approaches that we develop can be shared with

60
00:07:07,130 --> 00:07:13,130
other institutions as they join the ARCHiOx community as they acquire

61
00:07:13,400 --> 00:07:18,830
the hardware, the software, and indeed the desire to use the technology.

62
00:07:19,280 --> 00:07:26,990
So that we can support each other in taking this extraordinary opportunity forward.

63
00:07:28,630 --> 00:07:29,200
It's now.

64
00:07:29,350 --> 00:07:37,870
Unfortunately, the university, in its great wisdom, has decided this afternoon will be its meeting of the University's governing Council.

65
00:07:38,110 --> 00:07:43,930
And I must run probably literally from these words to that meeting.

66
00:07:44,290 --> 00:07:51,910
But I will join you later in the afternoon. But as I do so, it's now my great pleasure to hand over to my partner in crime here,

67
00:07:52,090 --> 00:07:58,280
Adam Lowe, the inspirational founder and leader of Factum Arte and Factum Foundation in Madrid.

68
00:07:58,300 --> 00:08:13,830
Ladies and gentlemen, please join me in welcoming Adam. 
Thank you very much.

69
00:08:14,220 --> 00:08:20,220
So many people turning out for the second of the ARCHiOx summaries

70
00:08:20,280 --> 00:08:23,729
about what we're doing. John, can I do that from here? 

71
00:08:23,730 --> 00:08:39,360


72
00:08:39,370 --> 00:08:50,420
Got it. For the second of the show and tells and summaries of what's gone on in the last year and a half

73
00:08:50,840 --> 00:08:56,780
here at the Bodleian and also in Madrid in Factum's studios.

74
00:08:57,140 --> 00:09:01,850
So I always love the sort of serendipity of what happens on these kind of occasions.

75
00:09:02,360 --> 00:09:10,159
And as I was walking in, Michael Rich was talking to someone I haven't seen for 2 or 3 years who's

76
00:09:10,160 --> 00:09:18,110
studying graffiti in the Temple of Karnak, who's just moved to study graffiti in Abydos.

77
00:09:18,380 --> 00:09:24,950
And is incredibly exciting about finding different languages on the surface of the walls in Abydos.

78
00:09:24,950 --> 00:09:34,190
And really, I wanted to begin with a real thank you to Helen Hamdan's Trust for believing in us for so long.

79
00:09:34,520 --> 00:09:37,610
And the image here is actually from this morning.

80
00:09:38,390 --> 00:09:48,890
That's an exaggeration. It was from yesterday, but it's in the Greek tomb in the Ipogeo Cristallini where we're using a vertical,

81
00:09:48,890 --> 00:09:54,710
the first proper use of the vertical Selene to record graffiti.

82
00:09:55,280 --> 00:10:00,440
These are Greek tombs in Naples from the fourth century BC.

83
00:10:00,890 --> 00:10:06,290
And this wall has never been recorded and it's absolutely covered in graffiti.

84
00:10:06,470 --> 00:10:14,390
On the left you can see Ferdy Saumarez Smith, who leads Factum Foundation in London, in the polychrome chamber.

85
00:10:14,390 --> 00:10:17,810
So this is the best Greek polychrome from the fourth century.

86
00:10:18,110 --> 00:10:22,250
And the Medusa's head is what we're recording today.

87
00:10:22,610 --> 00:10:31,930
So, I mean, what started as a dream to say we could build a scanner that has these properties.

88
00:10:31,940 --> 00:10:37,610
So I'm not going to try and go through them. But this is one you can do on your phones, that you can read it later.

89
00:10:38,090 --> 00:10:46,340
But basically the Selene scanner does a lot of things that enable us to focus on the materiality of objects.

90
00:10:46,730 --> 00:10:58,490
And it's really this system that John has been trialling, developing, pioneering, using and sharing downstairs and communicating to

91
00:10:58,910 --> 00:11:04,680
as Richard mentioned, a growing community of practice.

92
00:11:04,700 --> 00:11:08,030
So really the idea is to spread this knowledge.

93
00:11:08,480 --> 00:11:13,340
But I thought I'd just begin with a very quick thing and I'll do this one very fast.

94
00:11:13,340 --> 00:11:18,830
But Factum started in 2001 and we started developing the technologies.

95
00:11:19,310 --> 00:11:22,580
And these are a number of things. This one stops in 2021.

96
00:11:22,580 --> 00:11:30,650
Of course, it goes on beyond there. But developing the hardware and the software has been central to what we're focussed on in Madrid.

97
00:11:31,040 --> 00:11:41,000
And then applying that technology and making sure that it's used and focused to different targets and different end uses.

98
00:11:41,210 --> 00:11:46,770
And it's really that, that I'm most proud of about what's happened here at ARCHiOx.

99
00:11:48,970 --> 00:11:51,660
In the British Museum at the moment,

100
00:11:51,670 --> 00:12:02,020
they have become part of the community of practice and have a Selene which is currently recording cuneiform tablets.

101
00:12:02,800 --> 00:12:09,400
So, when they're flat like this, it's relatively easy for the Selene to record it.

102
00:12:09,400 --> 00:12:13,860
But most cuneiform tablets have curvature and have information on the edges

103
00:12:13,870 --> 00:12:19,989
so now we're working with them to try and see how we can extend, what it is.

104
00:12:19,990 --> 00:12:25,450
But if you can see this, the greatest thing we've done is to

105
00:12:25,660 --> 00:12:34,000
find a way through digital communication to make these objects physically available as if you're holding them,

106
00:12:34,360 --> 00:12:41,950
as if you're changing the light, as if you're actually with the object and can handle it and look at it.

107
00:12:41,950 --> 00:12:45,700
So digitality is no longer just taking a photograph.

108
00:12:46,330 --> 00:12:51,220
It's actually making something physical so you can physically engage with it.

109
00:12:51,990 --> 00:12:59,080
And I think the work on cuneiform is presenting lots of challenges that we are trying to solve.

110
00:12:59,080 --> 00:13:06,690
Now, John Barrett, as Richard mentioned, has been recording in the, and John 

111
00:13:06,700 --> 00:13:13,960
this is ready to introduce you, not to steal your thunder. But that actually, this is a type of

112
00:13:16,040 --> 00:13:19,790
fossil of a frog. So I'm going to do my show and tell.

113
00:13:19,880 --> 00:13:26,270
These we'll pass around the room. Taking the data John recorded,

114
00:13:26,540 --> 00:13:34,220
and outputting it with the qualities of the fossil and then removing the colour so you can see the fossil in different ways.

115
00:13:34,520 --> 00:13:41,750
But I think this sense of handling an object is really the key to

116
00:13:43,270 --> 00:13:46,720
what excites me about what Factum is doing with the Bodleian.

117
00:13:48,790 --> 00:13:50,470
When we started talking last year,

118
00:13:50,680 --> 00:14:00,520
Jo story and Jess Hodgkinson gave a talk about vellum. And out of that talk came a meeting with Alain Fouad George

119
00:14:00,520 --> 00:14:11,110
who came to Madrid and said, would we focus on recording the vellum used for early Qur'ans.

120
00:14:11,440 --> 00:14:16,810
And so John recorded this, which is a tiny detail showing the surface of the vellum.

121
00:14:17,320 --> 00:14:22,020
And again, if you take the same image and look at it with raking light

122
00:14:24,120 --> 00:14:33,120
you can appreciate the way the pigment's applied, the way the object, the physicality of the object, which is not possible really within conventional photography.

123
00:14:35,890 --> 00:14:39,910
Developing the software to go with the hardware is critical.

124
00:14:39,970 --> 00:14:44,860
And Carlos Bayod is going to talk later about this project,

125
00:14:44,860 --> 00:14:52,149
which was recording some incredibly deformed books in Lakenhal, which are actually, I mean for me

126
00:14:52,150 --> 00:15:01,630
I just love the books is the truth. They are the cloth samples made in Lakenhal of the famous black cloth that they were dying.

127
00:15:01,960 --> 00:15:08,320
And these are the books to prove every yarn of cloth that came from Lakenhal, so they could say, no, this one's fake,

128
00:15:08,320 --> 00:15:14,830
this is not Dutch black cloth. So all the books contain these little wedge shaped samples of black.

129
00:15:15,310 --> 00:15:19,600
And no one could really use them because the books are in such bad condition.

130
00:15:20,140 --> 00:15:25,000
But we wrote the software that after we'd scanned the pages, could flatten them and you'll hear more about that later.

131
00:15:28,530 --> 00:15:34,620
I think Carlos will also introduce what can be done with the data in different types of outputs.

132
00:15:34,620 --> 00:15:38,219
So making exact facsimiles is one.

133
00:15:38,220 --> 00:15:44,280
And the data being recorded here enables us to do extraordinary facsimile work.

134
00:15:45,840 --> 00:15:50,190
Building the community of practice means we have to go into other collections.

135
00:15:50,190 --> 00:15:56,510
And so last year, the Helen Hamlyn Trust gave us the funds to

136
00:15:56,610 --> 00:16:02,310
have two Selene scanners that we could place into different collections for short periods of time.

137
00:16:02,610 --> 00:16:12,840
And it's proved incredibly positive. So right now, the Selene is in the Instituto Valencia de Don Juan in Madrid.

138
00:16:13,440 --> 00:16:21,930
And it's recording a lot of the Islamic cloths and manuscripts and objects there.

139
00:16:22,260 --> 00:16:27,569
And there's also one at the moment in the Bruschettini collection in Milan

140
00:16:27,570 --> 00:16:37,860
which has been recording a wide range of objects. From this portolan chart, but to a whole lot of different fabrics especially.

141
00:16:38,280 --> 00:16:49,020
And it's there, Mongol fabrics that are really opening the way to show what can be done with this system: starting to do

142
00:16:49,350 --> 00:16:54,690
recordings and digital restorations of a whole range of different fabric.

143
00:16:55,670 --> 00:17:04,400
But as I try and sort of focus the day, I want to basically say most of the work here is happening with the Selene.

144
00:17:05,240 --> 00:17:11,270
The hardware and software and what you're going to hear about this afternoon from a whole

145
00:17:11,270 --> 00:17:17,030
group of different people who've been using it with Raphael drawings with

146
00:17:17,420 --> 00:17:22,610
the extraordinary copperplate that's been attributed to William Blake.

147
00:17:22,850 --> 00:17:32,120
And other case studies. But also we're doing photogrammetry and high resolution recording of,

148
00:17:32,120 --> 00:17:40,009
for example, the whole of the Divinity Schools and it's those that Richard was talking

149
00:17:40,010 --> 00:17:44,790
so again, just to pass around. There's some of these.

150
00:17:46,660 --> 00:17:54,200
Three and a half day period

151
00:17:54,590 --> 00:18:04,830
where we recorded the whole of the Divinity School ceiling, which in March, is going to become, a VR display

152
00:18:04,850 --> 00:18:08,870
so you'll be able to see it close up. But for me, I love the idea.

153
00:18:08,870 --> 00:18:18,500
These are actually greatly reduced. So, the one Deb is handling is actually about 50 to 40cm across, so they're reduced down.

154
00:18:18,950 --> 00:18:21,650
But the fact that with a 35 millimetre camera,

155
00:18:22,220 --> 00:18:30,410
the right skills and the right software you can now produce physical objects is something that never ceases to amaze me.

156
00:18:30,890 --> 00:18:37,660
So the ability, one of them is this head that's recorded

157
00:18:37,760 --> 00:18:45,260
as a three dimensional form and explaining the importance of the ceiling can now take on a whole new character.

158
00:18:45,830 --> 00:18:53,610
We can look at it in different ways. We can show it as a VR.

159
00:18:55,250 --> 00:18:59,000
The way we're doing the VR. So this is actually, you'll see it on the screen,

160
00:18:59,030 --> 00:19:03,940
normally you would be immersed in it, but it's using the

161
00:19:03,950 --> 00:19:10,460
new Finnish headsets that were developed for pilots where you can be right close to the surface.

162
00:19:11,060 --> 00:19:15,470
There are no hand controls. It's all eye tracking. You physically move around it.

163
00:19:16,360 --> 00:19:25,330
Put the colour on. You can take the colour off. You can suddenly start to engage in a totally different way with the object you are looking at.

164
00:19:26,640 --> 00:19:33,000
Anyways, you can experience the Tomb of Seti I, better,

165
00:19:34,210 --> 00:19:42,850
differently, from the experience of the original object because you're gravity free, so you could be close to the walls at any point in the two.

166
00:19:43,330 --> 00:19:47,080
So you can do in the VR what you would have to do

167
00:19:47,500 --> 00:19:55,719
with scaffolding normally. The final thing I want to do, because we mentioned the

168
00:19:55,720 --> 00:20:02,920
Natural History Museum and increasingly the recording of, or the merging of nature and culture.

169
00:20:03,340 --> 00:20:10,930
So just again as I was coming in here talking to Jo story about vellum and, you know, I could go on for hours now,

170
00:20:10,930 --> 00:20:18,730
talking about vellum and actually watching how it tells you that the data we're recording can tell you about animal diseases,

171
00:20:18,850 --> 00:20:25,510
herding, a whole lot of subjects that wouldn't normally come into a conversation about recording cultural heritage.

172
00:20:26,200 --> 00:20:29,560
I want to show this image because this is where we are right now.

173
00:20:29,710 --> 00:20:35,260
On the left is a photogrammetry of a little bug, let's say a centimetre across.

174
00:20:36,040 --> 00:20:43,270
And on the right is what we're experimenting with now, which is called Gaussian splatting.

175
00:20:43,930 --> 00:20:49,030
So if I can show you this little video to end on. There are about,

176
00:20:50,620 --> 00:20:55,059
how many, several billion specimens that are

177
00:20:55,060 --> 00:21:03,520
in natural history museums around Europe that are past their shelf life and are increasingly brittle and falling to bits.

178
00:21:04,000 --> 00:21:12,250
And from these specimens, you can tell so much information about evolution, about change, about

179
00:21:13,210 --> 00:21:16,760
character of the insect

180
00:21:16,760 --> 00:21:20,840
and what's happened to it. If you can record them in three dimensions.

181
00:21:21,890 --> 00:21:26,540
It's an enormous task and it's very complicated, but Gaussian splatting,

182
00:21:26,540 --> 00:21:30,820
and here you're watching a video. But if you were a researcher, you could control this.

183
00:21:30,830 --> 00:21:36,260
It's a three dimensional image. It's not a three dimensional model as such.

184
00:21:36,650 --> 00:21:39,860
And so you can see every hair, every detail.

185
00:21:40,340 --> 00:21:50,030
And it's this kind of research that's really pushing the envelopes, that are not only showing how technology can transform the formation of knowledge,

186
00:21:50,390 --> 00:21:56,420
but it can even break the professional barriers that have kept different knowledges apart.

187
00:21:56,900 --> 00:22:04,190
And for us, it's really this that ARCHiOx is pushing and developing.

188
00:22:04,760 --> 00:22:12,670
So, I know we have to keep on target and on track today, so I'm going to hand over to John.

189
00:22:13,120 --> 00:22:17,200
Oh no, Judith, are you going to do this or do I? Sorry, Judith is going to do this.

190
00:22:25,240 --> 00:22:28,719
Thank you Adam for setting the scene for us. So hello again

191
00:22:28,720 --> 00:22:33,850
everyone. I'm Judith Siefring. I'm head of Digital Collections Discovery here at the Bodleian, so I'm working on the digital library side.

192
00:22:34,330 --> 00:22:42,250
So for this first session, we're going to, sort of, take a deep dive into the research findings being enabled by this first phase,

193
00:22:42,250 --> 00:22:46,360
second phase of ARCHiOx, and then we'll have a brief Q&A.

194
00:22:46,630 --> 00:22:52,270
And then after that, we'll take a break around 2:45 for some tea and coffee out in Blackwell Hall.

195
00:22:52,870 --> 00:22:57,309
So, first of our three speakers in this first section will be John.

196
00:22:57,310 --> 00:23:01,180
So John Barrett is senior photographer for the Bodleian Libraries.

197
00:23:01,690 --> 00:23:06,340
And since 2005, he's provided photographs of Bodleian originals for numerous publications.

198
00:23:06,820 --> 00:23:10,630
His work involves the development of new methods of recording Special Collections material.

199
00:23:11,020 --> 00:23:17,110
And he's the author of Imaging Guidelines for Digitisation of Rare and Special Materials, a document that was commissioned by the Bodleian.

200
00:23:17,110 --> 00:23:22,299
And so he's very much been the lead for the Bodleian for ARCHiOx.

201
00:23:22,300 --> 00:23:25,570
And I will hand over to him to talk us through some of the findings from phase two.

202
00:23:26,260 --> 00:23:29,760
Thank you, Judith. Good afternoon everyone.

203
00:23:29,800 --> 00:23:36,610
I'm delighted to be with you today and to be able to share some of the amazing recordings which we've made for this exciting project.

204
00:23:37,590 --> 00:23:42,300
And it's a project which has involved making 3D recordings of our cultural heritage material.

205
00:23:43,140 --> 00:23:48,450
Prior to ARCHiOx, very little attention had been paid to the three dimensional surface of our collections.

206
00:23:48,780 --> 00:23:54,929
But through the recordings we've made for the project, we've come to realise that the materiality of manuscripts, books,

207
00:23:54,930 --> 00:23:59,640
printing plates and artwork can tell us a huge amount about who made them,

208
00:24:00,000 --> 00:24:03,990
how they were made, and about people's interactions with them since they were made.

209
00:24:06,920 --> 00:24:12,440
ARCHiOx, the Analysis and Recording of Cultural Heritage in Oxford began in January 2022.

210
00:24:12,920 --> 00:24:17,690
The project has been funded through the generous support of the Helen Hamlyn Trust,

211
00:24:18,320 --> 00:24:23,809
and the aim of the project has been to record originals from across Oxford University's

212
00:24:23,810 --> 00:24:27,800
Institutions using technology created by the Factum Foundation.

213
00:24:30,440 --> 00:24:32,120
This is what that technology looks like.

214
00:24:32,420 --> 00:24:39,930
It's named the Selene and it's a 3D recording system, which was designed and developed by Factum's Head of Technology Jorge Cano.

215
00:24:40,950 --> 00:24:46,890
All of the recordings, which I'll share during this first session of today's conference, have been made with this incredible tool.

216
00:24:47,790 --> 00:24:57,690
The Selene uses a principle called photometric stereo to acquire both colour and 3D data through the capture and processing of just four photographs.

217
00:25:00,110 --> 00:25:05,960
Here we can see the four photographs in the left column, each lit at intervals of 90 degrees.

218
00:25:06,860 --> 00:25:14,660
Through analysing the position of the highlights and shadows, the four photographs can be processed in order to create additional 2D image files,

219
00:25:14,900 --> 00:25:19,460
which encode the direction and height of the surface at each pixel.

220
00:25:22,000 --> 00:25:29,260
The Bodleian was the first institution to use the Selene and through our collaboration with Factum Foundation, we have assisted in its development.

221
00:25:30,510 --> 00:25:37,920
During this second phase of ARCHiOx, the recording system has not only been used at the Bodleian, but also at the Ashmolean Museum,

222
00:25:38,280 --> 00:25:44,250
the Oxford University Museum of Natural History, and to record material from the History of Science Museum.

223
00:25:46,600 --> 00:25:55,690
At the end of 2022, I presented this recording of what was considered to be an early 18th century printing plate. Being small and heavily corroded,

224
00:25:55,930 --> 00:25:58,120
the design was almost impossible to see.

225
00:26:00,020 --> 00:26:05,690
But through making a recording of a three dimensional surface, I demonstrated how the design could be revealed.

226
00:26:06,590 --> 00:26:13,910
The image on the right is called a shaded render. It's similar to making a plaster impression of the surface and then relighting it with a torch.

227
00:26:14,630 --> 00:26:19,670
This introduces shadows and highlights to reveal the surface without the distraction of its colour.

228
00:26:22,200 --> 00:26:30,920
I also showed the recording of the reverse of the same plates. Revealing markings which were impossible to see in the conventional image:

229
00:26:31,430 --> 00:26:41,230
musical notes and words. After mirroring the image and adding a little of the colour to make it easier to read.

230
00:26:42,420 --> 00:26:48,640
I showed how we could annotate over the markings in order to reveal the entire design.

231
00:26:50,610 --> 00:26:54,270
Which we discovered was a musical score inspired by Psalm nine.

232
00:26:56,620 --> 00:27:02,110
But today I can share that not only do we know what the score looks like, but also what it sounds like.

233
00:27:06,190 --> 00:27:11,180


234
00:27:21,200 --> 00:27:29,090
The piece which you've just heard is the first set of Psalms for Three Voices by William Child, dating to 1639.

235
00:27:29,480 --> 00:27:33,950
And this recording has allowed researcher Chiara Betti to conclude that this small

236
00:27:33,950 --> 00:27:38,840
printing plate may be the earliest surviving engraved music plate in the UK.

237
00:27:41,700 --> 00:27:47,010
Just as with other research imaging techniques which have become commonplace over the last few decades,

238
00:27:47,370 --> 00:27:51,900
photometric stereo is proving to be immensely powerful as a tool for discovery.

239
00:27:55,070 --> 00:28:00,980
As well as being able to produce images capable of showing features which can't be easily seen or photographed,

240
00:28:01,490 --> 00:28:17,490
it can also be used to make detailed measurements. We can interact with the images, relighting the renders from any direction or height.

241
00:28:17,880 --> 00:28:22,020
We can capture large originals in their entirety, not just small samples.

242
00:28:22,440 --> 00:28:25,590
And not only are these recordings highly effective and versatile,

243
00:28:25,920 --> 00:28:32,880
they are also extremely quick to make. The capture and processing of an A5 size tile takes less than a minute.

244
00:28:36,370 --> 00:28:44,050
The Selene is particularly useful in revealing the tool markings left by authors and artists throughout the centuries and across continents.

245
00:28:47,140 --> 00:28:54,250
And in almost all cases, it's the tools which don't leave a visible trace that are most successfully captured rather than those which do.

246
00:28:57,010 --> 00:29:02,170
Though a feather quill may leave a faint scratch on the surface of a page of parchment or paper,

247
00:29:04,160 --> 00:29:06,560
the surface reveals very little of the text.

248
00:29:06,800 --> 00:29:12,830
However, in this case, the ruled horizontal lines which were not present in the colour image can be seen clearly.

249
00:29:15,660 --> 00:29:21,630
The previous example was a recording of the Bodleian's earliest map of the world, dating from the 11th century.

250
00:29:22,170 --> 00:29:28,170
The volume was made in Milan, and is a version of a work by Bede, the English monk and Anglo-Saxon scholar.

251
00:29:31,040 --> 00:29:33,529
Though the nib of a pen is difficult to capture in the surface

252
00:29:33,530 --> 00:29:41,510
recording, the Selene is incredibly effective in recording markings made using a sharp tool, such as a stylus or divider.

253
00:29:44,030 --> 00:29:49,250
Here we have a detail from an astronomical manuscript which is covered with preparatory stylus incisions.

254
00:29:49,490 --> 00:29:57,710
It's just that they're very, very difficult to see. Until we make a render of the surface and exclude the colour.

255
00:30:01,570 --> 00:30:06,730
Evidence of the use of sharp scribal tools has not only been found in European manuscripts.

256
00:30:06,940 --> 00:30:13,810
Markings made with reed pens and iron styluses have been recorded from manuscripts from our South Asian collections too.

257
00:30:15,670 --> 00:30:19,690
We can see an example of this in this recording of a Sinhalese palm leaf.

258
00:30:22,130 --> 00:30:27,710
Where the outline of this deer has been both incised and then pricked as a result of copying.

259
00:30:33,040 --> 00:30:37,430
But of course, the practice of writing with sharp tools dates back far further. Incised

260
00:30:37,450 --> 00:30:43,450
text has been recorded from a variety of substrates for this project, from clay to wood and wax.

261
00:30:46,080 --> 00:30:53,280
Such as the second century tablet, which had never been imaged in any way until earlier this year when the following recording was made.

262
00:30:56,560 --> 00:30:59,800
A new 3D recording of one of several wax tablets.

263
00:31:02,950 --> 00:31:09,160
Photometric stereo imaging allows us to view brushstrokes and pigment layers in ways that we've never been able to before.

264
00:31:11,480 --> 00:31:17,780
We can see these as distinct layers on the surface of the page at a resolution of 4 million pixels per square inch.

265
00:31:18,020 --> 00:31:22,220
The recording reveals the composition and condition of these pigments.

266
00:31:22,520 --> 00:31:28,940
What we're looking at is a detail from a copy of an Egyptian manuscript compiled almost exactly a thousand years ago.

267
00:31:33,450 --> 00:31:40,680
The example shown here being the waqwaq tree whose branches or fruits became the heads of men, women, and monstrous animals.

268
00:31:43,880 --> 00:31:47,420
And finally we have tools used for writing and drawing on metal:

269
00:31:47,570 --> 00:31:55,460
the burin and the etching stylus. These were the tools used for the production of hundreds of printing plates held in the Bodleian's collections.

270
00:31:57,270 --> 00:32:01,440
The latter having been used in the production of this heavily oxidised copper plate.

271
00:32:01,980 --> 00:32:07,470
It's strange shape suggests that it would have been used to print hand fans during the 18th century.

272
00:32:10,100 --> 00:32:19,560
And this is what the recording of the surface has revealed. With the design revealed, I started to consider what the fan might have looked like.

273
00:32:20,070 --> 00:32:29,660
Here we can see my initial prototype. And a final coloured version, allowing us to see the design as it might have appeared perhaps 280 years ago.

274
00:32:32,870 --> 00:32:38,420
The recreation of the fan demonstrates how 3D recordings could be used in the production of facsimiles.

275
00:32:39,200 --> 00:32:45,380
In 2022, the Factum Foundation made a facsimile of the Bodleian's Gough Map of Great Britain.

276
00:32:45,830 --> 00:32:51,290
And since then, we have recorded dozens of other amazing maps and atlases from our collections.

277
00:32:54,400 --> 00:32:59,470
Here is a tile from the recording of the Gough map, which reveals tiny indentations.

278
00:33:02,270 --> 00:33:07,550
Which were used as a technique to copy the position of geographical features from a precursor map.

279
00:33:10,960 --> 00:33:15,400
We've since discovered that this was a method used in the production of other important maps,

280
00:33:15,610 --> 00:33:18,970
like this 16th century edition of the Al-Idrisi world map.

281
00:33:19,840 --> 00:33:24,520
The render reveals pricking around the coastlines and in the corners of islands.

282
00:33:32,880 --> 00:33:37,530
In this example, we have a page of manuscript by explorer and cartographer William Baffin.

283
00:33:40,000 --> 00:33:44,500
And again, the same copying technique is evident in his Atlas of 1619.

284
00:33:47,260 --> 00:33:53,650
The frontispiece of the same volume has been planned in great detail with prepatory stylus before adding ink to the page.

285
00:33:56,290 --> 00:33:59,320
With the colour removed, this preparatory stage becomes clear.

286
00:34:02,900 --> 00:34:05,240
You can see the same technique in his elevations.

287
00:34:07,470 --> 00:34:12,900
Where stylus markings have served as a form of shorthand, in this case marking the position of trees.

288
00:34:16,590 --> 00:34:21,060
We can also see how Baffin's final design sometimes deviates from his preparatory work.

289
00:34:21,660 --> 00:34:24,690
Here we can see how the position of the word 'north' has been altered.

290
00:34:27,020 --> 00:34:31,250
And on the left we can see a decoration absent in the final design.

291
00:34:33,660 --> 00:34:37,890
But throughout centuries we've attempted not only to map the earth, but also the heavens.

292
00:34:38,430 --> 00:34:42,210
The volume you see here is from the collections of the History of Science Museum.

293
00:34:44,310 --> 00:34:49,260
At least some of the devices from this manuscript were likely to have been created in the 1550s.

294
00:34:51,880 --> 00:34:58,300
The colour layer removed from the 3D recording, we can see how each and every line was meticulously incised.

295
00:35:00,030 --> 00:35:03,420
The volume features seven instruments, including this astrolabe.

296
00:35:06,270 --> 00:35:13,050
By isolating each component, it's been possible to produce an animation, giving us a sense of how the astrolabe might operate.

297
00:35:14,210 --> 00:35:17,470
The animation we see here was produced by my colleague Dylan Schirmacher.

298
00:35:18,140 --> 00:35:23,570
And we can see, while the rete turns, that we can blend from the colour to the 3D render.

299
00:35:32,700 --> 00:35:38,850
Having made the previous recording, Stephen Johnston of the History of Science Museum invited me to record this brass astrolabe.

300
00:35:39,770 --> 00:35:45,230
It was made in Germany in 1521 and includes three double sided plates.

301
00:35:45,650 --> 00:35:53,930
We can see these change in this animation. These are positioned behind the moving parts and contain markings which specify latitude.

302
00:35:55,530 --> 00:36:03,600
Just as with the recordings of the Bodleian's copper printing plates, the 3D renders reveal the markings on this astrolabe with exceptional clarity.

303
00:36:09,620 --> 00:36:16,970
As we saw from the William Baffin Atlas, stylus incisions are extremely hard to spot but can be recorded very effectively with the Selene.

304
00:36:17,900 --> 00:36:23,660
There are several reasons for making these markings, and through these recordings we can better understand their purpose.

305
00:36:26,630 --> 00:36:28,400
During the previous ARCHiOx conference.

306
00:36:28,520 --> 00:36:36,590
A series of recordings from this eighth century Acts of the Apostles were shared by Leicester University PhD candidate Jessica Hodgkinson.

307
00:36:38,940 --> 00:36:43,740
The Selene recording of this lower margin revealed an inscription made in dry point.

308
00:36:46,150 --> 00:36:50,830
Totally absent from the colour recording and impossible to see under conventional lighting.

309
00:36:53,340 --> 00:36:56,690
What was clear and complete from the recording was the name Eadberg,

310
00:36:56,730 --> 00:37:04,400
preceded by a cross. But as well as discovering incised text in the margins,

311
00:37:05,620 --> 00:37:11,770
the Selene was also successful in revealing informal illustrations which captured the attention of the media.

312
00:37:14,400 --> 00:37:19,050
Mysterious figures, whose outlines may have been drawn around a finger or thumb.

313
00:37:21,820 --> 00:37:25,740
Jessica has continued to find evidence of dry point inscriptions while working

314
00:37:25,740 --> 00:37:30,000
as Special Collections curatorial assistant at New College here in Oxford.

315
00:37:32,180 --> 00:37:39,260
In June of this year, Jess and I recorded a folio from this small 16th century manuscript which was made for King Henry VIII.

316
00:37:41,880 --> 00:37:45,000
Here we see a section of the lower margin.

317
00:37:47,390 --> 00:37:49,880
And this is the recording of the surface of the same page.

318
00:37:50,240 --> 00:37:55,760
Jessica's research suggests that the first letter appears to be a characteristically Tudor 'H'.

319
00:37:58,380 --> 00:38:06,210
And towards the end of the dry point addition, the Roman numeral VIII can also be identified, followed by a final calligraphic flourish.

320
00:38:07,140 --> 00:38:10,590
It's impossible to be certain that this inscription is the work of Henry VIII,

321
00:38:10,860 --> 00:38:15,090
but given its position beneath an initial enclosing a crowned Tudor rose,

322
00:38:15,420 --> 00:38:21,900
there is plenty of evidence which suggests that this could indeed be his signature discovered after five centuries.

323
00:38:25,050 --> 00:38:28,170
And these kinds of markings have been revealed throughout the project.

324
00:38:28,590 --> 00:38:32,190
Here we have an example from a 13th century manuscript from Abingdon.

325
00:38:35,160 --> 00:38:39,120
It's a manuscript which researcher Luise Morawetz invited me to record.

326
00:38:39,750 --> 00:38:43,500
And we had no idea that among the numerous dry points and inscriptions,

327
00:38:44,650 --> 00:38:48,340
we would find a beast lurking in the lower flyleaf.

328
00:38:51,180 --> 00:38:56,100
Stylus editions not only reveal readers interactions with manuscripts.

329
00:38:56,130 --> 00:38:59,550
Sometimes the markings are evidence of processes in their production.

330
00:39:01,740 --> 00:39:11,160
From this 11th century manuscript, we see how incised text served as a reminder for the addition of rubrication in a later scribal stint.

331
00:39:14,070 --> 00:39:18,060
From the render, you might be able to make out the text in the left margin.

332
00:39:19,940 --> 00:39:26,900
Which reads virginiia. And this has been adapted to virginitatis in red ink in the uppermost line.

333
00:39:30,550 --> 00:39:35,890
In this 16th century Arabic volume, a stylus has been used for planning the page layout.

334
00:39:38,000 --> 00:39:42,620
Perhaps these could be instructional markings informing another scribe of their intentions.

335
00:39:44,150 --> 00:39:47,570
Or simply a technique used to ensure a neat presentation.

336
00:39:51,500 --> 00:39:55,700
Here we have another example of how research imaging can assist in understanding

337
00:39:55,700 --> 00:40:00,020
the processes involved in the manufacture of manuscripts and artworks.

338
00:40:00,560 --> 00:40:06,410
This Moralised Bible was produced in Paris between 1226 and 1275.

339
00:40:08,550 --> 00:40:12,720
The frontispiece shows God holding a divider creating heaven and earth.

340
00:40:15,520 --> 00:40:19,210
Eight medallions from each page were created using the same tool.

341
00:40:22,660 --> 00:40:26,440
By using a range of imaging techniques. Photometric stereo,

342
00:40:27,660 --> 00:40:33,500
transmitted light, ultraviolet, infrared, and hyperspectral imaging.

343
00:40:35,270 --> 00:40:39,020
We can, in a sense, turn back the clock and see how the manuscript was made.

344
00:40:39,350 --> 00:40:41,420
So beginning with a blank sheet of parchment.

345
00:40:43,330 --> 00:40:49,690
The first stage would have involved tracing a rough design around a model or earlier volume with a broad, rounded tool.

346
00:40:52,560 --> 00:40:55,950
Basic under drawing would then have been added in a carbon based ink.

347
00:40:57,750 --> 00:41:03,210
Improvements would then have been added. Followed by the burnished gold leaf.

348
00:41:05,750 --> 00:41:10,100
3D recordings then help us determine the order in which the colour pigments were applied.

349
00:41:10,340 --> 00:41:14,590
Firstly, red led. Dark azurite.

350
00:41:16,910 --> 00:41:20,240
And a copper pigment was used for the turquoise of the crucifix.

351
00:41:22,310 --> 00:41:26,830
A lighter azurite. Followed by an organic red,

352
00:41:26,980 --> 00:41:34,490
most likely derived from insects. And finally a brown ink, completing the details in the faces.

353
00:41:37,010 --> 00:41:44,300
The Bodleian's volume contains 1784 miniatures, each one measuring just 7.5cm in diameter.

354
00:41:45,140 --> 00:41:51,110
Two sets of these three part manuscripts were produced, which must have taken an enormous amount of time and dedication.

355
00:41:55,520 --> 00:42:00,380
As we've seen, the recordings made with the Selene are entirely complementary with existing imaging techniques.

356
00:42:00,710 --> 00:42:06,950
In this example, we see a Coptic Gnostic text recorded as an albedo or colour image without shadow,

357
00:42:07,340 --> 00:42:15,410
a photometric stereo render, a transmitted light image where the manuscript is illuminated from beneath, and an infrared recording.

358
00:42:15,810 --> 00:42:20,540
These can be layered and blended to allow us to reveal the design more clearly.

359
00:42:23,060 --> 00:42:29,690
There's no right or wrong approach. A composite of multiple layers may provide a researcher with valuable new information.

360
00:42:32,380 --> 00:42:40,080
Working side by side of a researcher is mutually beneficial, as it ensures that the final render is made with the benefit of context.

361
00:42:43,030 --> 00:42:44,860
To conclude this first presentation,

362
00:42:45,020 --> 00:42:52,450
I'd like to take this opportunity to shed some light on some originals which have not received as much attention as many of our books and manuscripts.

363
00:42:53,290 --> 00:42:57,430
Though they cover the walls of every corridor and reading room throughout our historic libraries,

364
00:42:57,940 --> 00:43:06,030
in many cases, we know very little about them. And yet many of the Bodleian's portraits have an interesting story to tell.

365
00:43:06,900 --> 00:43:10,350
Here we have a curious case of the back of a painting not matching the front.

366
00:43:10,950 --> 00:43:18,360
The wooden panel bears a label stating that the portrait depicts Sir Thomas More, the 16th century statesman and martyr.

367
00:43:21,520 --> 00:43:24,760
It's a portrait that was a particular interest to this man. Nowell

368
00:43:24,790 --> 00:43:30,100
Myres, the Bodleian's librarian from 1948 to 1965.

369
00:43:30,310 --> 00:43:35,620
Like his father, Myres was an archaeologist and historian and, perhaps in the spirit of uncovering the past,

370
00:43:35,800 --> 00:43:40,420
he embraced x-ray imaging in order to examine portraits from the Bodleian's collection.

371
00:43:41,540 --> 00:43:47,869
In 1955, Myres wrote to paintings conservator Horace Buttery to inform him that an x-ray

372
00:43:47,870 --> 00:43:52,250
of the Thomas Moore portrait had revealed a perfectly preserved head of a lady.

373
00:43:53,030 --> 00:44:00,170
He goes on to write, 'Sir Thomas More of course, is quite worthless and I think we must have him off and chercher la femme.'

374
00:44:03,020 --> 00:44:07,340
Thankfully, before Buttery removed the overpainting, the portrait was photographed.

375
00:44:07,790 --> 00:44:11,210
This poor quality black and white contact print is all that remains.

376
00:44:11,990 --> 00:44:17,810
The photograph reveals that the overpainting is a copy of Hans Holbein's portrait from 1527.

377
00:44:20,610 --> 00:44:27,890
So what did Myres and Buttery uncover? The original female portrait is not of significant quality,

378
00:44:28,220 --> 00:44:33,710
but evidence from a partial inscription suggests that she may have been painted in the 16th century.

379
00:44:34,580 --> 00:44:37,970
But is there evidence which remains on the surface from the overpainting?

380
00:44:41,390 --> 00:44:42,500
As this animation plays,

381
00:44:42,500 --> 00:44:50,450
we can see in the Selene recording a series of S-shaped forms extending from the sitter's left ear to the left side of her forehead.

382
00:44:51,050 --> 00:44:58,430
And if we then fade to the 1950s photograph with the overpainting, we see that these are in fact the links of Thomas More's chain of office.

383
00:45:02,620 --> 00:45:10,590
Perhaps Myres was delighted or perhaps he felt short changed. When overlaid in exactly the same position, we can compare the two portraits.

384
00:45:12,970 --> 00:45:17,410
Whether or not we consider Myres to have been right to request the removal of the overpainting,

385
00:45:17,620 --> 00:45:23,200
what is less ambiguous is that he should have ensured that it was properly recorded before being erased forever.

386
00:45:23,950 --> 00:45:31,630
The Thomas More portrait could have been photographed in colour. Kodachrome was introduced 20 years earlier, in 1935.

387
00:45:32,110 --> 00:45:38,350
However, nearly 70 years after the overpainting was removed, we can have a go at recreating it digitally.

388
00:45:40,450 --> 00:45:47,680
Here the original frame has been reinstated and the colour has been added to the monochrome photograph, using Holbein's portrait as a guide.

389
00:45:49,680 --> 00:45:54,090
This is perhaps how the Thomas Moore portrait would have appeared from the 17th century.

390
00:45:57,200 --> 00:46:04,310
This small panel painting of Charles I bears a resemblance to the portrait painted by Anthony van Dyck in 1633.

391
00:46:05,640 --> 00:46:12,780
Recording that we're looking at is not a conventional photograph. Instead, it's a combination of both a colour image and a render of the surface.

392
00:46:13,500 --> 00:46:18,870
Though unconventional, the image provides far more information regarding the materiality of the painting.

393
00:46:22,540 --> 00:46:28,030
And using a high resolution recording system is especially important when recording a small portrait like this one.

394
00:46:28,630 --> 00:46:33,880
The Selene even allows us to record the colour and surface data through the domed glass.

395
00:46:35,850 --> 00:46:40,140
Something which would be either challenging or impossible to do with a laser based system.

396
00:46:41,040 --> 00:46:44,380
Here we see the colour image, render,

397
00:46:46,330 --> 00:46:51,159
and finally the composite. At first

398
00:46:51,160 --> 00:46:59,050
this 1897 portrait of Thomas Alcock might appear to be a pencil sketch, but it is in fact a collotype print.

399
00:47:01,210 --> 00:47:05,200
As you might expect to see, we have no tool markings in the surface,

400
00:47:05,560 --> 00:47:08,620
only the chain lines and laid lines embedded in the paper.

401
00:47:11,270 --> 00:47:16,360
Being able to record light sensitive material like this print both quickly and without

402
00:47:16,520 --> 00:47:21,560
subjecting it to bright lighting has ensured that it has not been damaged during capture.

403
00:47:27,220 --> 00:47:31,120
And finally we have another portrait featuring an unidentified female sitter.

404
00:47:31,600 --> 00:47:38,830
In this case, she was revealed in 1838 when an overpainted portrait of Mary, Queen of Scots was removed.

405
00:47:40,650 --> 00:47:46,770
Before removing the overpainting, the portrait, as it appeared in the early 19th century, was copied not once

406
00:47:47,930 --> 00:47:56,640
but twice. This is a recording which I made in 2022, prior to her much needed conservation treatment.

407
00:47:58,630 --> 00:48:06,160
And this is how she appeared earlier this year when I recorded her again halfway through restoration and with the varnish having been removed.

408
00:48:09,500 --> 00:48:17,150
Recorded at over a million pixels per square inch. The recording even allows us to see the woodgrain in the panel on which the painting was made.

409
00:48:18,050 --> 00:48:23,960
We can see how the paints have cracked and flaked. We can exclude the colour and concentrate only on the surface.

410
00:48:25,300 --> 00:48:31,870
And we can use other types of shaders like this Atlas shader, which reveals the surface using a spectrum of colour.

411
00:48:36,790 --> 00:48:41,680
Not only have the recordings proved to be useful for the conservator in trying to untangle the painting's history,

412
00:48:42,040 --> 00:48:47,410
they've also ensured that the painting is documented before, during, and after its treatment.

413
00:48:48,010 --> 00:48:53,200
This demonstrates how photometric stereo imaging can be used both as a technique for discovery,

414
00:48:53,560 --> 00:48:58,150
and also as an effective means of ensuring the preservation of our collections material.

415
00:48:58,780 --> 00:49:03,520
And I'll be focusing on the theme of preservation in the next session.

416
00:49:15,100 --> 00:49:22,780
Wonderful. Thank you, John. So for the second, part of this session, we're going to focus on some new research.

417
00:49:23,290 --> 00:49:30,190
I'd like to introduce Mark Crosby, who is an associate professor in English and the director of the Digital Centre at Kansas State University.

418
00:49:30,790 --> 00:49:36,730
He has published on William Blake and his circle, and most recently co-edited a collection on Blake's manuscripts.

419
00:49:37,150 --> 00:49:42,970
Mark is currently preparing a catalogue of Blake's apprenticeship materials in the Bodleian Library and the Society of Antiquaries.

420
00:49:43,380 --> 00:49:47,320
And Mark is going to talk to us about his findings on copper plates.

421
00:49:59,840 --> 00:50:05,540
I should start by saying the last time I was in this building giving a talk on Blake, I went on for 75 minutes.

422
00:50:05,810 --> 00:50:08,180
I will not tax your attention for that long! I've got 20 minutes.

423
00:50:08,660 --> 00:50:14,660
So to start, before I dive into the copper plates, I'd like to present some historical context to begin with.

424
00:50:15,160 --> 00:50:31,400
In 1809, the antiquarian, first director of the Society of Antiquaries, Richard Gough, bequethed his vast collection to the Bodleian Library.

425
00:50:32,370 --> 00:50:33,870
Included in this collection

426
00:50:35,150 --> 00:50:45,500
were the materials relating to his mammoth publication, Sepulchral Monuments in Great Britain, which was published between 1786 and 1796.

427
00:50:47,770 --> 00:50:51,100
The publishing, this particular volume.

428
00:50:53,330 --> 00:50:56,780
This particular volume contains engravings by James Basire.

429
00:50:58,620 --> 00:51:02,759
The engravings were made along with the preparatory drawings in the 1770s,

430
00:51:02,760 --> 00:51:11,070
between 1773 and 1779 / 1780, which coincided with William Blake's apprenticeship.

431
00:51:11,430 --> 00:51:18,960
Now, before Blake became an apprentice to Richard Gough, he spent four years at Henry Pars' Drawing School on the strand.

432
00:51:19,290 --> 00:51:23,490
Now the Drawing School was seen as the most pre-eminent drawing school in London at the time,

433
00:51:23,850 --> 00:51:30,360
and it was considered a prerequisite to attending Saint Martin's Academy, which would be the precursor to the Royal Academy.

434
00:51:31,020 --> 00:51:36,690
So if you wanted to attend Saint Martin's and then the Royal Academy, you would have to attend Pars' Drawing School.

435
00:51:37,170 --> 00:51:42,630
Now, Blake demonstrated a precocious talent for drawing, so much so that his father,

436
00:51:43,050 --> 00:51:49,140
a Hosier by trade, wanted to apprentice Blake to a painter of eminence.

437
00:51:49,470 --> 00:51:57,660
However, the fees involved in the apprenticeship were so much that Blake instead requested to be apprenticed to an engraver.

438
00:51:57,990 --> 00:52:03,570
Now, the first engraving master that Blake's father approached was a chap called William Wynne Ryland.

439
00:52:04,440 --> 00:52:11,370
When Blake was introduced to William Wynne Ryland, Blake reportedly said to his father, I do not like that man's face.

440
00:52:11,610 --> 00:52:14,040
He will live to be hanged.

441
00:52:15,180 --> 00:52:24,720
Ten years later, William Wynne Ryland was convicted of forging East India Company bills and hanged by the neck at Tyburn.

442
00:52:25,530 --> 00:52:30,870
Some say Blake is a prophet and a visionary. Maybe this is the evidence for such things.

443
00:52:33,840 --> 00:52:41,430
Instead, Blake was apprenticed to James Basire on the 4th of August 1772 for the princely sum of £52.

444
00:52:41,700 --> 00:52:47,640
The apprenticeship involved spending seven years living with the Basire household.

445
00:52:48,150 --> 00:52:55,230
So he'd moved from his home in Poland Street to Basire's studio in Great Queen Street, just near Lincoln's Inn.

446
00:52:55,770 --> 00:52:59,910
And here we have the documentation from the Stationers Records in London,

447
00:53:00,150 --> 00:53:06,780
where you have William Blake signing away his life for seven years at the tender age of 14.

448
00:53:06,840 --> 00:53:14,829
Now, there is plenty of evidence we have for Blake and what he was doing during his apprenticeship.

449
00:53:14,830 --> 00:53:21,940
And some of this evidence is first hand, that is, people who knew Blake, who worked with Blake or friends with Blake during his life.

450
00:53:22,210 --> 00:53:26,620
So we can infer that they got this evidence from Blake directly.

451
00:53:27,220 --> 00:53:36,070
In particular, what happened in 1773 is that Blake was sent to Westminster Abbey to make a series of preparatory drawings of

452
00:53:36,070 --> 00:53:42,850
the medieval royal tombs and effigies of wall paintings that then were to be engraved for Gough's publication.

453
00:53:45,390 --> 00:53:53,100
And this is Basire's studio in Great Queen Street. And this is a representation of the engraving studio. Now as an apprentice

454
00:53:53,340 --> 00:53:59,580
the first years would have involved some of the menial tasks like cleaning the studio, preparing the copper plates.

455
00:53:59,880 --> 00:54:05,790
But clearly Blake was such a good draughtsman that Basire entrusted him, just one year into his apprenticeship,

456
00:54:06,000 --> 00:54:09,720
to going make these preparatory drawings for this very significant publication

457
00:54:10,080 --> 00:54:18,300
of Gough's. We also have evidence that Blake, by 1773, could engrave.

458
00:54:18,420 --> 00:54:23,460
Now, what you're seeing here are the only two states of what we think of today

459
00:54:24,480 --> 00:54:28,740
(and maybe this will have to be revised after this talk) of Blake's earliest engravings.

460
00:54:29,040 --> 00:54:40,080
They are engravings after a figure from Michelangelo's The Death of the Crucifixion of Saint Peter, from engraving by Francesco Salviati.

461
00:54:40,590 --> 00:54:49,260
Now the one on your left contains an inscription that reads 'engraved when I was a beginner at Basire's'.

462
00:54:49,260 --> 00:54:56,820
And then the later one, actually gives us the date 'engraved by W Blake, 1773, from an old Italian drawing'.

463
00:54:57,570 --> 00:55:00,840
The state on the right is the later state,

464
00:55:00,840 --> 00:55:05,610
the second state. So it's the state on the left, which is the first state which we can date to 1773.

465
00:55:05,820 --> 00:55:08,760
Now, the fascinating thing about this, by one year into his apprenticeship,

466
00:55:09,090 --> 00:55:13,049
Blake already had mastered the visual language of line engraving. That he could

467
00:55:13,050 --> 00:55:19,080
cover a copper plate with the series of incised linear patterns to recreate form,

468
00:55:19,080 --> 00:55:25,230
tone, and depth. In other words, by 1773, Blake could engrave.

469
00:55:26,120 --> 00:55:35,810
Now when he went to Westminster Abbey, he spent between 1773 to possibly 1777 making a series of preparatory drawings.

470
00:55:36,240 --> 00:55:39,380
To give you a sense of where he was during this period.

471
00:55:39,800 --> 00:55:50,000
In 1774 he starts in the Confessor's chapel, this is just behind the high altar, making drawings of the royal effigies and tombs.

472
00:55:51,170 --> 00:55:59,990
The following year, he moves onto the great Cosmati pavement in front of the high altar, faces to the north

473
00:56:00,260 --> 00:56:09,320
and starts sketching the tombs of, the recently revealed tombs of King Sebert, supposedly the original founder of Westminster Abbey,

474
00:56:09,710 --> 00:56:12,830
as well as the medieval wall paintings that were there.

475
00:56:12,980 --> 00:56:21,590
Now, these medieval wall paintings at the time were believed to have been executed by a contemporary of Giotto, a chap called Calvinatti.

476
00:56:22,930 --> 00:56:30,100
He did an about turn to face the opposite direction, and did a series of preparatory sketches of the other royal tombs on the other side of the

477
00:56:30,460 --> 00:56:37,000
Cosmati pavement, and then moved to Saint Edmund's Chapel to sketch the tombs there.

478
00:56:37,360 --> 00:56:46,189
In total he did over 50 drawings. Now these drawings, are all signed J Basire, J Basire del. 

479
00:56:46,190 --> 00:56:49,150
The Latinate formula for delineating he is the author of them.

480
00:56:49,300 --> 00:56:56,050
Now, this was commonplace practice for the master of the studio to sign any work coming out of the studio.

481
00:56:59,840 --> 00:57:06,290
And what I've been doing with John over the last year and a half is looking at both the drawings and the copper plates.

482
00:57:06,650 --> 00:57:11,270
Now, it's been accepted in Blake's scholarship that

483
00:57:11,270 --> 00:57:14,840
many of these drawings are attributed to Blake because of the circumstantial evidence.

484
00:57:15,290 --> 00:57:24,160
And the six copper plates of the portraits of kings and queens have also been accepted as attributed to Blake.

485
00:57:24,600 --> 00:57:30,950
So part of the investigation that we've been conducting is whether or not the other copper plates after

486
00:57:30,950 --> 00:57:35,450
Blake's preparatory drawings are also or could also be attributed to Blake.

487
00:57:35,750 --> 00:57:45,830
So one of the first things we did, or I asked John to do, was to use the Selene technology to look at the six oval copper plate portraits,

488
00:57:46,220 --> 00:57:52,980
to see if we could find the consistencies in terms of the depth of lines that were etched in there.

489
00:57:53,000 --> 00:57:56,420
And what we found is across all six copper plate portraits,

490
00:57:56,630 --> 00:58:03,950
there is a consistency of depth, roughly 18 microns deep, which is a quarter of the width of a human hair,

491
00:58:03,950 --> 00:58:13,220
so very shallow. And that's consistent across all six copper plates, which would suggest the same hand was responsible for the etching of these.

492
00:58:14,450 --> 00:58:28,649
Press play. And we focus on particular areas of, sort of densely hatched areas under the eyes, around the ears and, as I said,

493
00:58:28,650 --> 00:58:36,780
all of these measurements prove to be very consistent, which again suggests that the same hand is responsible for them.

494
00:58:39,040 --> 00:58:44,710
Again, all of the engravings that come from this studio are signed by him.

495
00:58:44,980 --> 00:58:49,600
Now, there are question marks over the rest of the engravings, in large part because

496
00:58:50,360 --> 00:58:55,760
the first volume of Sepulchral Monuments wasn't published until 1786, which is

497
00:58:57,020 --> 00:59:05,930
seven years after Blake completes his apprenticeship, and there's also one engraving that bears the date 1783, which again would post date Blake's 

498
00:59:06,910 --> 00:59:13,209
involvement in this particular plate. However, we know from the Gough correspondence and his 

499
00:59:13,210 --> 00:59:22,120
manuscript autobiography here in the Bodleian that he started working on this project in 1771 or 1772,

500
00:59:22,330 --> 00:59:29,620
and commissioned Basire to start the process of making the prepatory drawings and the engravings in 1773.

501
00:59:29,860 --> 00:59:37,660
That's to say the engravings, some of the engravings were done much earlier than the publication date of the volume in 1786.

502
00:59:39,410 --> 00:59:45,080
Now there are a couple of plates in particular that require our attention.

503
00:59:45,620 --> 00:59:51,409
Um, what I was interested in doing is looking at the rectos and the versos of the copper plates. Now,

504
00:59:51,410 --> 00:59:55,040
on many of the copper plates on the rectos you have the printed image.

505
00:59:55,310 --> 01:00:00,530
And now on the versos, there were some extraordinary markings. On this plate in particular,

506
01:00:00,980 --> 01:00:10,290
what we have is a series of practice work. Now for an engraver to learn how to engrave

507
01:00:10,440 --> 01:00:14,160
they needed to use the engraving tools and a copper plate.

508
01:00:14,790 --> 01:00:22,110
And what you're looking at now, is the verso of a copper plate filled with practice work.

509
01:00:23,410 --> 01:00:27,240
Basire's studio was famous for its linear style.

510
01:00:28,280 --> 01:00:33,560
A series, a visual language that uses linear patterns to recreate form and tone.

511
01:00:33,860 --> 01:00:38,540
And this is an apprentice practising those patterns on a copper plate.

512
01:00:40,740 --> 01:00:46,380
So what we have is a consistency in depth of line and spacing of line.

513
01:00:46,710 --> 01:00:55,680
This demonstrates the apprentice engraver was practising putting pressure on a burin to gouge out these lines at a consistent depth,

514
01:00:55,830 --> 01:01:03,900
and a consistent width. The consistent width gives you a consistent tone, and the consistent depth again indicates a consistent pressure

515
01:01:05,000 --> 01:01:10,420
when creating these patterns. In addition to straight lines, we also have hatching,

516
01:01:10,690 --> 01:01:18,040
one of the main aspects of a linear approach is the use of hatching to create, again, depth and tone.

517
01:01:18,310 --> 01:01:21,910
And here we have rectilinear areas. The one on your left is

518
01:01:22,960 --> 01:01:27,580
an apprentice practising how to create hatching repeating patterns.

519
01:01:27,820 --> 01:01:35,860
There's also some swirls and what could be seen as sort of curved lines and potentially sort of vegetative,

520
01:01:35,860 --> 01:01:44,350
shapes that correspond with some of the copper plates attributed to Blake that contain similar curved lines.

521
01:01:44,710 --> 01:01:50,230
In other words, an apprentice practising these curved lines by using a Burin or an etching needle,

522
01:01:50,530 --> 01:01:55,030
are evidence on the rectos of these copper plates.

523
01:01:55,150 --> 01:01:56,020
Versos, sorry.

524
01:01:56,320 --> 01:02:04,930
In the case of these vegetative lines, it sort of anticipates Blake's own use of interlinear vegetation, most famously in his illuminated books.

525
01:02:04,930 --> 01:02:14,410
And here's an example from the Bodleian copy of Songs of Innocence which Blake produced exactly ten years after he finished his apprenticeship.

526
01:02:14,860 --> 01:02:17,200
It's probably his most accessible work.

527
01:02:17,560 --> 01:02:24,160
And if you're interested and need a Christmas Present, the Bodleian have a wonderful edition for sale right now in the bookstore.

528
01:02:26,970 --> 01:02:33,250
So one of the things that the Selene technology has allowed us to do is to 

529
01:02:34,520 --> 01:02:39,710
get information, render the information in a way that is just impossible with the naked eye.

530
01:02:40,130 --> 01:02:45,110
And so in addition to the squiggly lines and straight lines and hatching marks,

531
01:02:45,350 --> 01:02:51,829
there's also a series of semicircles which could only have been made with a dry point compass.

532
01:02:51,830 --> 01:02:57,560
And John has been able to detect the shadow impressions of where the compass

533
01:02:57,590 --> 01:03:04,130
would have set as Blake started to experiment with these semicircles.

534
01:03:04,430 --> 01:03:13,810
Why is that sort of significant? Well, we find in the finished engravings, a series of swirls that were used with drypoint compass

535
01:03:13,910 --> 01:03:20,660
and Blake himself about four years after he completed his apprenticeship and went into business as a printmaker,

536
01:03:20,840 --> 01:03:25,220
he uses a compass in architectural designs. And this is from 1784.

537
01:03:25,570 --> 01:03:29,450
And Blake's use of a compass you can see at the top and bottom on the left hand side.

538
01:03:31,980 --> 01:03:37,080
Compasses and circles also feature as visual motifs throughout Blake's works.

539
01:03:37,080 --> 01:03:39,000
Most famously, this is called the Ancient of Days,

540
01:03:39,000 --> 01:03:46,290
you may be familiar with this particular image. Where you have the Urizenic God figure circumscribed by

541
01:03:47,410 --> 01:03:51,370
a globe of some sort, holding the dividers,

542
01:03:51,370 --> 01:03:55,900
or most Blake scholars interpret this as the golden compass mentioned in Milton's Paradise Lost.

543
01:03:56,380 --> 01:04:01,209
Compasses also feature regularly in his watercolour paintings. He did 537 

544
01:04:01,210 --> 01:04:07,660
watercolour illustrations to Young's Night Thoughts, several of which include these compasses

545
01:04:09,130 --> 01:04:17,080
as well as the Illuminated Book, Jerusalem, his longest and most complex work in illuminated printing, includes a compass as well.

546
01:04:17,770 --> 01:04:20,810
And most famously, this image you may be familiar with

547
01:04:20,850 --> 01:04:25,060
if you've been to the British Library, you have this Paolozzi's fabulous sculpture.

548
01:04:25,460 --> 01:04:32,500
This is in the Tate Britain at the moment. This is Newton using a compass to circumscribe an arc within a triangle.

549
01:04:35,530 --> 01:04:45,100
On the back of another copper plate, we noticed some other incisions that were less structured, more spontaneous, and perhaps playful.

550
01:04:45,220 --> 01:04:47,770
Now, what you're looking at here is a plate mark,

551
01:04:48,070 --> 01:04:54,820
and what the apprentice engraver has done is decide to put some hatching in the first 'o' of London.

552
01:04:54,820 --> 01:04:59,260
To give you an idea, the size of the 'o' is about 1.5mm.

553
01:04:59,440 --> 01:05:06,880
So he's working at a sort of very minute scale in a very controlled way, using some form of magnification.

554
01:05:09,630 --> 01:05:14,790
There's also this. This is a short shafted, heavily fletched arrow.

555
01:05:15,890 --> 01:05:20,360
Was made with a type of burin called a scorper that has a flat end.

556
01:05:20,900 --> 01:05:25,100
It's 18mm long, so while you can see it with the naked eye,

557
01:05:25,280 --> 01:05:29,150
it's very difficult to see much of the details.

558
01:05:29,840 --> 01:05:33,170
Why is that important for Blake? Well Blake loved an arrow.

559
01:05:33,560 --> 01:05:38,540
He loved more than arrows. In his pictorial work as well as his poetic work,

560
01:05:38,780 --> 01:05:44,420
arrows feature prominently as a motif. Indeed, it's a motif that sort of reappears again and again.

561
01:05:44,750 --> 01:05:47,809
This is from Blake's illustrations to Milton's Paradise Lost.

562
01:05:47,810 --> 01:05:51,380
This is Satan meeting his son, otherwise known as death.

563
01:05:51,710 --> 01:05:54,050
Armed both with arrows.

564
01:05:54,590 --> 01:06:01,970
This is another illustration from Blake's Paradise Lost, where you have the son with his seven arrows defeating the rebel angels.

565
01:06:02,300 --> 01:06:11,630
Sending them all to the fiery lake. In his Night Thoughts, we have the reappearance of the sun with his bow pulling arrows out of quivers,

566
01:06:11,840 --> 01:06:18,710
and then death appears as well in the same set of illustrations, holding a short shafted, heavily fletched arrow.

567
01:06:18,860 --> 01:06:28,339
Blake also engraved 43 of these illustrations, so he did 537 watercolours illustrating Young's poem,

568
01:06:28,340 --> 01:06:35,390
of which then he engraved 43 that were published in a doomed attempt to enter into the large illustrated book market.

569
01:06:35,660 --> 01:06:35,840


570
01:06:35,840 --> 01:06:42,590
Sadly, the market fell through because Britain decided to go to war with Napoleon. Actually the other way around, Napoleon declared war on Britain.

571
01:06:44,200 --> 01:06:49,630
And then it his poetry. There are too many examples to name, but I'm going to give you the most famous.

572
01:06:50,180 --> 01:06:54,879
This is the preface to Milton, the hymn now known as Jerusalem, where we have the lines:

573
01:06:54,880 --> 01:06:59,620
'Bring me my bow of burning gold. Bring me my arrows of desire.'

574
01:07:00,010 --> 01:07:09,280
So if we're looking at a work by Blake here, this could be the first instance of Blake using this motif around the age of 15,

575
01:07:09,280 --> 01:07:14,920
a motif that reappears both textually and visually throughout the rest of his life.

576
01:07:15,490 --> 01:07:22,180
The other motif we found on this plate was almost impossible to see, largely because

577
01:07:23,300 --> 01:07:31,720
the existence of what's called repoussage. Repoussage is a way that an engraver uses to correct mistakes or to amend a plate.

578
01:07:31,740 --> 01:07:38,660
So if something needs amending, the copper plate is flipped over and they take a small ball point hammer and knock out the mistake.

579
01:07:39,630 --> 01:07:48,050
So what we have here is a tiny face. Engraved, incised, into the back of the plate underneath the repoussage.

580
01:07:51,320 --> 01:07:55,220
There we go. Now this is, the scale again is minute.

581
01:07:55,430 --> 01:07:59,030
This is, from top to bottom, about 18mm.

582
01:07:59,360 --> 01:08:06,410
What the engraver has done, he's hatched the eyes to give a sort of expressive character.

583
01:08:06,710 --> 01:08:12,260
And it's a face. But it's a face that I sort of, I think is a doodle,

584
01:08:13,260 --> 01:08:18,150
rather than a fully formed face. In other words, during the apprentice's long day

585
01:08:18,360 --> 01:08:23,310
where he's doing these very repetitive actions of making these linear marks.

586
01:08:23,640 --> 01:08:27,720
There are times that clearly he needs to entertain himself.
587
01:08:28,050 --> 01:08:31,860
So he flips over the plates and he starts doodling.

588
01:08:33,520 --> 01:08:41,710
In other words, this is personal and private art that for the last 250 years, we haven't been able to access.

589
01:08:41,860 --> 01:08:45,670
The Selene technology now makes it possible to see this face

590
01:08:47,470 --> 01:08:51,879
made by an apprentice 250 years ago. If this is Blake,

591
01:08:51,880 --> 01:08:57,580
what you're looking at now is Blake's first portrait, albeit a crude portrait, but Blake's first portrait.

592
01:08:58,840 --> 01:09:05,370
Thank you very much. 
Wonderful.

593
01:09:05,370 --> 01:09:11,519
Thank you so much. Thank you to all our speakers also for running so beautifully to time.

594
01:09:11,520 --> 01:09:15,180
So we have around ten minutes 10-15 minutes for a Q&A.

595
01:09:15,750 --> 01:09:19,470
We have Niki at the back with a microphone so I can see one hand up here already.

596
01:09:19,480 --> 01:09:23,400
Can I invite Angelamaria and Mark, to come up and sit with John for some questions?

597
01:09:27,710 --> 01:09:31,970
Hi. Two quick questions for Adam, actually.

598
01:09:32,540 --> 01:09:42,920
Can the software behind Selene be reverse engineered take a digital flat image and

599
01:09:45,040 --> 01:09:49,210
bring out the implied three dimensionals actually within the image.

600
01:09:52,650 --> 01:09:57,090
I can take this Adam if you like. Yes, it does already.

601
01:09:57,120 --> 01:10:05,070
So what you're doing is you're capturing both a colour image without texture at all, and also the 3D without the colour.

602
01:10:05,520 --> 01:10:09,180
So it does both things automatically at the same time.

603
01:10:09,690 --> 01:10:15,600
And can that be printed? 
Yes, absolutely. 
Possibly a question for both of you.

604
01:10:15,930 --> 01:10:23,759
Um, in the Tutankhamun exhibition, currently on in Cairo, do some of Adam's Factum Arte

605
01:10:23,760 --> 01:10:27,560
works actually on show?

606
01:10:32,920 --> 01:10:38,970


607
01:10:55,380 --> 01:11:03,840
Proposal is to do a show in Gem that will look at the Hall of Beauties in its current condition from the tomb of Seti I,

608
01:11:04,260 --> 01:11:10,950
and then explain why it looks like that and then take you out through the recoloured

609
01:11:11,250 --> 01:11:16,830
Hall of Beauties that we made for the exhibition in Basel in 2017.

610
01:11:17,160 --> 01:11:21,900
So there's a lot of information on Factum Foundation's website about that.

611
01:11:25,350 --> 01:11:33,770
And that's, more questions

612
01:11:37,830 --> 01:11:50,170
I was just picking up on you,very nice, phrase, using a dry point to think discreetly.

613
01:11:50,480 --> 01:11:53,110
Yeah, which I really like, that's really helpful, that's very good.

614
01:11:53,590 --> 01:12:02,530
Um, and I was also thinking about how the use of raking light is something that links everything and all of the,

615
01:12:02,830 --> 01:12:09,340
examples that we've seen so far back from the, the eighth century manuscripts all the way through to, to to your thing.

616
01:12:09,670 --> 01:12:17,800
and I was wondering if it's, it's I was thinking more about the raking light and because we're thinking a lot about that in terms of how,

617
01:12:17,950 --> 01:12:20,650
monuments, things would have been envisaged in,

618
01:12:20,770 --> 01:12:28,809
in churches or tombs with flickering light and how that would have acted as a real,dynamic force to the way,

619
01:12:28,810 --> 01:12:35,080
the things on the wall were, were seen to move as the light moved around them and people moved with, with the lighting.

620
01:12:35,410 --> 01:12:42,910
But I was also wondering if there's any sense in which perhaps of the paper rather than, the membrane, but then perhaps with that to whether,

621
01:12:43,240 --> 01:12:54,150
you could use transmitted light so light from behind to,make these tiny, indentations more visible.

622
01:12:54,160 --> 01:12:58,390
Is it just the raking light, or could we have light behind? I don't know if they've done any experiments with that.

623
01:12:58,460 --> 01:13:03,900
No, no, not the to pick up the to be engraved lines is how I like calling them.

624
01:13:03,910 --> 01:13:10,750
yes, you need rake, you need Selene and John if you want to get the best results.

625
01:13:10,750 --> 01:13:14,470
But,in in the absence of John, you need

626
01:13:15,520 --> 01:13:20,340
dark, you need the dark. You know, we think that the truth is in the light,

627
01:13:20,380 --> 01:13:25,400
the truth is in the dark. You need a source of raking

628
01:13:25,420 --> 01:13:32,950
light, a good torch. and these, you know, if you don't have Selene these marks, not everything,

629
01:13:32,950 --> 01:13:39,670
so, for example, one of the real discoveries, I think, is to anatomical,

630
01:13:40,060 --> 01:13:46,240
notations that we've seen and in a way, we could pick them up, you know, pick them in any way.

631
01:13:46,240 --> 01:13:53,820
but the transmitted light, that is something different, it just highlights the, in the structure of the paper, I see,

632
01:13:53,890 --> 01:14:02,710
watermarks and, the laid lines and and so on, but, um, but even transmitted light, if you're drawing these lines.

633
01:14:02,740 --> 01:14:07,110
this is where Selene comes, very useful because,

634
01:14:07,120 --> 01:14:13,309
you you can pick up all these elements with raking light if the drawing has, you know, drawn down on something else.

635
01:14:13,310 --> 01:14:16,990
so, but for blind, blind side of dry point,

636
01:14:16,990 --> 01:14:23,060
yes, absolutely, raking light. I thought this print was quite nice.

637
01:14:23,070 --> 01:14:27,970
I hadn't seen this before, so this is Bandinelli. This is, yes, this is a design of Bandinelli.

638
01:14:28,410 --> 01:14:35,010
And we see some some work happening here with a light at a low angle and a sort of flickering light, in the dark.

639
01:14:35,010 --> 01:14:41,520
Yeah, there's no window here. So they're in the dark and they are using um, yes, a candle light and, and.

640
01:14:42,180 --> 01:14:49,800
yeah, so I think when you, when you see an original with which has been, um, annotated with blind styluses,

641
01:14:50,400 --> 01:14:54,000
you don't really get a sense of how difficult it is to see these markings.

642
01:14:54,450 --> 01:14:57,240
obviously with the recordings that we've made, they're very, very visible.

643
01:14:57,750 --> 01:15:03,870
but you wonder how the, the makers of, of this artwork and these manuscripts were able to be guided by them.

644
01:15:03,870 --> 01:15:12,540
And I think this is the this is the key, um, that they were working from, from a low light, which perhaps it was, was moving as well.

645
01:15:26,280 --> 01:15:29,640
I shouldn't ask a question, but I'm going to...Mark.

646
01:15:30,090 --> 01:15:34,680
When they were doing the, copperplate, etchings and stuff.

647
01:15:34,710 --> 01:15:39,360
Would that have been a similar light format? Would they've been doing it under candlelight or what?

648
01:15:39,360 --> 01:15:46,139
They've been working on the daylight because you said about how tiny the actual markings were.

649
01:15:46,140 --> 01:15:53,570
So. Yeah. I suppose the traditional, literature, on engraving in the 18th century was daylight.

650
01:15:53,600 --> 01:15:58,700
Natural light and engraving desk were situated next to large windows.

651
01:15:58,850 --> 01:16:04,280
when Blake moved to the south coast, the only time he left London was for a three year period,

652
01:16:04,430 --> 01:16:10,100
in 1800, where he lived in a ship's cottage made of ship's timbers.

653
01:16:10,100 --> 01:16:14,990
And his engraving studio was on the bottom floor next to a large window.

654
01:16:15,470 --> 01:16:24,830
So the idea was that he could engrave using natural light. Having said that, um, there is there is some evidence that again,

655
01:16:24,830 --> 01:16:29,450
this is slightly circumstantial evidence that Blake would engrave after his drawings

656
01:16:29,450 --> 01:16:34,040
and the phrase used in this documentary evidence is especially in winter.

657
01:16:34,940 --> 01:16:40,520
So the suggestion here is he, wouldn't use artificial light, candle lights to engrave as well.

658
01:16:46,220 --> 01:16:50,360
One more question, if anyone's brave enough to ask?

659
01:16:51,140 --> 01:17:02,940
Yes, do. John, this might be a bit leftfield.

660
01:17:03,060 --> 01:17:06,720
Has this technology got a forensic application?

661
01:17:08,580 --> 01:17:16,709
Yes, it probably does. I mean, I'll go into some forensic analysis, which is which has happened on more of a volumetric recording,

662
01:17:16,710 --> 01:17:20,310
which was recorded using a different principle called photogrammetry.

663
01:17:20,760 --> 01:17:26,790
So we'll focus on that in the next session. Um, but I'm sure there are absolutely applications.

664
01:17:27,090 --> 01:17:32,760
I mean, we know that, for instance, the pressure of a pen through a page can be transferred to a page underneath.

665
01:17:33,180 --> 01:17:38,190
I've recorded those markings before. So yes, there's all manner of different applications.

666
01:17:38,550 --> 01:17:41,830
Thanks. Thank you.

667
01:17:41,840 --> 01:17:46,700
I think we'll leave the session here. Um, we have a half hour break.

668
01:17:47,030 --> 01:17:54,830
in Blackwell Hall. There's refreshments available in the cordoned off area, and then we'll be back to start again around sort of quater past three.

669
01:17:55,370 --> 01:17:59,449
so I hope we'll see you all back here for more amazingness after the break.

670
01:17:59,450 --> 01:18:10,529
But for now, can you join me in thanking all of our wonderful speakers.

671
01:18:10,530 --> 01:18:13,830
So thank you everyone for coming back after coffee.

672
01:18:14,430 --> 01:18:23,440
and now we're going into the second session, which will last until about sort of 4:20, when we'll have time for a Q&A.

673
01:18:23,460 --> 01:18:26,790
So we've got three more talks, in this second session.

674
01:18:26,790 --> 01:18:32,400
And to begin with, I will invite John back,to tell us more about findings in phase two.

675
01:18:34,540 --> 01:18:41,380
Thank you Judith. In the first session, I shared a selection of recordings made using a principle called photometric stereo.

676
01:18:42,040 --> 01:18:49,900
The Selene recording system, which uses this principle, is a perfect fit for recording the surface of library material, it being mostly flat.

677
01:18:50,680 --> 01:18:55,660
Photometric stereo has numerous advantages over other 3D recording techniques.

678
01:18:55,990 --> 01:19:01,150
It's incredibly fast and highly efficient, allowing us to record at very high resolutions,

679
01:19:01,720 --> 01:19:05,740
but it's only capable of recording the surface of our collections, not the volume.

680
01:19:06,580 --> 01:19:11,800
In order to record the volume, it's necessary to use a different technique, which I'll describe in this session.

681
01:19:18,460 --> 01:19:22,090
What are the most important objectives for ARCHiOx has been to create recordings which

682
01:19:22,090 --> 01:19:27,760
assist in preserving our collections by embracing both these 3D recording techniques.

683
01:19:27,950 --> 01:19:33,580
We've been able to record everything from a tiny fossilised wasp to an entire historic building.

684
01:19:36,210 --> 01:19:41,290
At the previous ARCHiOx conference, I shared this Selene recording of a Japanese scroll,

685
01:19:41,620 --> 01:19:48,670
which demonstrated how a photometric stereo recording can highlight conservation concerns such as insect damage,

686
01:19:48,970 --> 01:19:51,700
historic repairs, and flaking pigment.

687
01:19:54,690 --> 01:20:01,920
This conventional image of a painting from our South Asian manuscripts may not appear to show features of concern to a conservator.

688
01:20:03,480 --> 01:20:07,620
But the render of the surface reveals areas of creasing and pigment loss.

689
01:20:08,220 --> 01:20:12,660
This demonstrates how photometric stereo renders can assist in condition mapping.

690
01:20:15,470 --> 01:20:24,500
Making recordings of extremely fragile originals like this unfired clay letter seal from the 2nd century B.C. has ensured that they are protected.

691
01:20:25,630 --> 01:20:27,790
Showing a facsimile in place of the original,

692
01:20:27,940 --> 01:20:34,630
has allowed students to have a hands-on experience without the need to remove the original seal from its box.

693
01:20:36,570 --> 01:20:41,910
And not only do we make a facsimile of the seal, but also the lost matrix, which was used to make it.

694
01:20:42,570 --> 01:20:48,450
The designs from seven seals from the same collection were recorded using the Selene and combined.

695
01:20:50,820 --> 01:20:57,000
A 3D print of the surface was then made onto a flexible material, which was wrapped around a cylindrical core.

696
01:20:57,720 --> 01:21:05,670
The facsimile can be used, as the original would have been, leaving an impression of the emblem of a Persian Prince named Arshama.

697
01:21:09,180 --> 01:21:17,130
Photographing seals can be challenging. Capturing a wax seal, like the 17th century example, would typically involve using a single raked light.

698
01:21:17,790 --> 01:21:20,700
While this technique may reveal the design successfully,

699
01:21:20,970 --> 01:21:27,030
it’s possible that the image may suffer from excessive contrast and un-even illumination across the surface.​

700
01:21:27,870 --> 01:21:31,440
These problems can be mitigated by making a photometric stereo recording. 

701
01:21:31,830 --> 01:21:37,200
Which has the additional benefit of allowing the end-user to relight the recording from any direction.

702
01:21:38,360 --> 01:21:41,510
As you can see, it's possible to change the viewing position.

703
01:21:41,840 --> 01:21:48,950
However, as the recording was made from a single position above the seal, we cannot turn the seal over and view the other side.​

704
01:21:51,780 --> 01:21:56,520
In this example, the 3D recording has been made using a technique called photogrammetry.

705
01:21:57,690 --> 01:22:04,950
While Selene acquires 3D data through capturing images illuminated from different positions and with the camera in a static position.

706
01:22:05,430 --> 01:22:08,340
Photogrammetry is a technique which works in the opposite way.

707
01:22:08,940 --> 01:22:15,270
A photogrammetry recording is made by photographing the object from multiple angles under even illumination,

708
01:22:15,540 --> 01:22:23,430
and combining the images into a three-dimensional model. In this example, the seal was photographed from 200 positions.

709
01:22:24,480 --> 01:22:28,350
This 13th century seal features two portraits of Henry III.

710
01:22:28,740 --> 01:22:32,280
The charter to which it was attached is a Confirmation of Magna Carta

711
01:22:35,230 --> 01:22:39,640
Here we can see the skeleton of the model, a mesh containing half a million triangles.

712
01:22:40,390 --> 01:22:45,280
Image files encoding colour, shadow and texture are then projected onto the mesh.

713
01:22:47,860 --> 01:22:50,020
To create a photogrammetry recordings

714
01:22:50,290 --> 01:22:58,060
ARCHiOx has used the PalaeoPi, a programmable turntable designed by Richard Allen, which synchronises rotation and capture.
715
01:22:59,170 --> 01:23:05,050
This is a highly efficient and relatively inexpensive solution for the photogrammetry of small originals,​

716
01:23:05,290 --> 01:23:08,140
like this 17th century embroidered Book of Psalms.

717
01:23:10,920 --> 01:23:19,590
Recordings like this one have been published online and can be explored in the browser on a desktop, laptop, tablet, or even a smartphone.

718
01:23:19,830 --> 01:23:25,260
We can rotate, we can zoom in, pan, relight the model as if holding the book in our hands.

719
01:23:29,830 --> 01:23:39,130
There are currently 76 photo parametric models available, but available to view online, which showcase some of the most beautiful book bindings.

720
01:23:39,940 --> 01:23:46,660
Here we have another 17th century embroidered book featuring a portrait of Charles II and his wife Catherine.

721
01:23:52,450 --> 01:23:56,920
But photometric recordings take a lot of time to make, sometimes several hours.

722
01:23:57,310 --> 01:24:02,620
They also produce large amounts of data in order to publish a photogrammetry model online.

723
01:24:02,770 --> 01:24:06,640
It's necessary to diminish the resolution and complexity of the recording.

724
01:24:07,390 --> 01:24:12,490
It's therefore very difficult to examine a small detail with as much clarity as you might like,

725
01:24:13,000 --> 01:24:17,200
such as the tiny peacock perched on top of the oval from this binding.

726
01:24:21,810 --> 01:24:24,209
Creating a high resolution recording of a detail

727
01:24:24,210 --> 01:24:31,110
using the Selene takes just a few seconds, with the peacock recorded at over 4 million pixels per square inch

728
01:24:31,410 --> 01:24:42,550
we can now examine each individual thread. As we've seen, both 3D recording methods have their own limitations and advantages.

729
01:24:43,330 --> 01:24:50,620
In this example, a photogrammetry model is useful as it allows us to see each face of this function or scientific binding.

730
01:24:51,130 --> 01:24:54,460
A French book dating from 1678.

731
01:24:58,270 --> 01:25:06,640
A note written by bookseller E.P. Goldschmidt, states that this volume is a rare and extraordinary book printed at a small town in Normandy,

732
01:25:06,850 --> 01:25:14,860
preserved in its original binding, which in fact comprises astronomical instruments without which no copy can be said to be complete.

733
01:25:16,040 --> 01:25:24,680
Before being presented to the Bodleian, and in 1978 this book belonged to the Broxbourne Library, compiled by book collector Albert Ehrman.

734
01:25:25,430 --> 01:25:31,250
Ehrman had a particular interest in bookbinding and named his library after his hometown in Hertfordshire.

735
01:25:33,790 --> 01:25:37,599
The Broxbourne Collection comprises some of the most beautiful book-bindings in the Bodleian's

736
01:25:37,600 --> 01:25:41,649
collections; here we have a tiny Portuguese book bound

737
01:25:41,650 --> 01:25:47,410
in the early 19th century, the gold-tooled Morocco case feature the Arms of Portugal,

738
01:25:48,400 --> 01:25:54,130
and while many of the book bindings which have been recorded for the project were chosen due to their aesthetic quality

739
01:25:54,520 --> 01:26:05,070
the next small volume is an exception. This pocket sized book was recorded instead because of the intriguing story surrounding its provenance.

740
01:26:05,970 --> 01:26:15,570
It was presented to the Bodleian in 1893, and catalogued by librarian E.W.B. Nicholson, as being found in Shelley's hand at his death.

741
01:26:16,740 --> 01:26:22,020
The volume was displayed with the caption ‘Shelley’s Sophocles which he had with him when drowned.

742
01:26:22,260 --> 01:26:25,590
Presented to the Bodleian Library, by Jane, Lady Shelley.

743
01:26:25,950 --> 01:26:29,670
the end edge is supposed to show the mark of Shelley’s thumb’.​

744
01:26:32,650 --> 01:26:38,560
Whether or not this story is true, there is good evidence to support the claim that this volume belonged to Percy Shelley.

745
01:26:39,370 --> 01:26:45,970
The Bodleian holds a matching pocket-sized edition of Aeschylus containing annotations in Shelley’s hand.

746
01:26:48,670 --> 01:26:55,450
Had the Sophocles volume not been discovered with Shelley’s body, perhaps it could have been among the volumes salvaged from his boat, the Ariel.

747
01:26:56,470 --> 01:27:03,790
What is undeniable is the presence of a thumb-sized indentation on the upper edge of this small, water-damaged book.

748
01:27:07,300 --> 01:27:15,310
In contrast, here we have one of the Bodleian's finest bindings an early 12th centaury Byzantine New Testament, bound in silver and ivory.

749
01:27:16,370 --> 01:27:23,720
Reproducing the reflective nature of metallic surfaces is challenging in both conventional photography and in photogrammetry.

750
01:27:24,980 --> 01:27:30,050
By selectively masking only the silver parts of the model and applying a metalness mask,

751
01:27:30,290 --> 01:27:36,080
my colleague, Dylan Schirmacher, has produced a very convincing 3D model of the original.

752
01:27:40,070 --> 01:27:43,010
And it's not just bindings that have been captured using photogrammetry.

753
01:27:43,550 --> 01:27:49,610
We've also recorded objects like this celestial globe, which was made by John Senex in 1728.  ​

754
01:27:50,600 --> 01:27:55,070
Senex was a prominent cartographer, astrologer and geographer to Queen Ann.

755
01:27:56,000 --> 01:27:59,000
Here Senex has mapped the stars according to the zodiac.

756
01:28:02,010 --> 01:28:05,730
Photogrammetry has also been used to assist in trying to solve a mystery.

757
01:28:06,270 --> 01:28:11,670
On the left we see a portrait bust of James Gibbs, the Architect of the Radcliffe Camera.

758
01:28:14,140 --> 01:28:18,580
The bust is a little tricky to reach, being located in the mezzanine of the upper reading room,

759
01:28:20,660 --> 01:28:25,610
from where it has been positioned, Gibbs has a wonderful view over the building which he designed.

760
01:28:28,270 --> 01:28:37,570
The regions of the bust, which were accessible, were photographed 300 times, and the images aligned to create a 3D model of over 5 million polygons.

761
01:28:40,170 --> 01:28:46,260
The request for the recording was made by the Bodleian’s former Exhibitions and Loans Conservator, Dana Josephson.​

762
01:28:50,010 --> 01:28:55,920
And the reason for making this recording is not because there is any doubt about who the bust depicts, or who sculpted it.

763
01:28:57,740 --> 01:29:04,310
We know this to be the work of John Rysbrack.  In fact, this is one of a pair of busts commissioned in 1726. ​

764
01:29:04,760 --> 01:29:10,130
The other bust, showing  Gibbs in contemporary dress  and a wig, is in the Victoria and Albert Museum.

765
01:29:13,750 --> 01:29:18,250
It's the identity of the sitter from this bust which is open to question.

766
01:29:19,150 --> 01:29:25,060
Now it also shares a home in the Radcliffe Camera. It is situated to the right of the doorway leading to the upper reading room.

767
01:29:27,700 --> 01:29:33,070
And despite bearing an inscription stating that the bust depicts Gibbs if it were carved from life.

768
01:29:34,470 --> 01:29:42,390
This would make it a very early specimen of a Greek Revival herm bust, which did not become popular in England until the 19th century.

769
01:29:44,600 --> 01:29:50,120
For this reason, and given that we don't know who sculpted it, we cannot be sure that this is in fact, Gibbs.

770
01:29:52,290 --> 01:30:00,690
What we do know is that the bust came to the library in 1927 and may once have been owned by the writer and art-historian Horace Walpole.

771
01:30:04,470 --> 01:30:08,430
Earlier this year, Dana Josephson commissioned Professor Caroline Wilkinson,

772
01:30:08,610 --> 01:30:15,840
Director of the Forensic Research Institute at Liverpool John Moores University to analyse the recordings of both busts.​

773
01:30:16,990 --> 01:30:18,880
Though the full report has yet to be received,

774
01:30:19,150 --> 01:30:26,800
Professor Wilkinson's summary suggests that the number of differences between the faces cannot be explained by the difference in ages depicted.

775
01:30:27,730 --> 01:30:34,420
She concluded that there was little evidence to support the assertion that the two pack, two busts depicts the same person.

776
01:30:37,750 --> 01:30:44,380
Not too far away in the Bodleian’s Old Schools Quad we have a statue which is often mistaken by tourists to be William Shakespeare.​
777
01:30:45,010 --> 01:30:49,840
In fact, it depicts William Herbert, the 16th and 17th century Earl of Pembroke.

778
01:30:50,560 --> 01:30:53,920
The statue originally stood in the outer court of Wilton House.

779
01:30:55,490 --> 01:31:03,200
This recording is a work in progress, and the model we see here was made by photographing the statue from 1200 different positions.

780
01:31:04,140 --> 01:31:09,750
But this is a comparatively modest effort compared to another recording, which was made last month.

781
01:31:11,900 --> 01:31:18,950
I'd been asked by colleagues to make a high resolution photograph of the entire ceiling of this room, the 15th century Divinity School.

782
01:31:20,030 --> 01:31:27,860
Given that the ceiling is at its highest point over 8 meters tall, another objective was to create a physical facsimile of a few of the bosses,

783
01:31:28,070 --> 01:31:32,660
so that visitors could have the opportunity to see these incredible carvings up-close.​

784
01:31:33,470 --> 01:31:38,420
Though ambitious, Adam Lowe suggested that the whole ceiling should be recorded in 3D.​

785
01:31:41,670 --> 01:31:47,100
Given the scale and complexity of the challenge, the recording was made by colleagues from the Factum Foundation.

786
01:31:47,790 --> 01:31:54,270
Cue photogrammetry experts, Imran Kahn, who photographed the ceiling from 16,000 positions at ground level.

787
01:31:56,500 --> 01:32:03,940
And Otto Lowe, who first recorded the entire room using LiDAR and then captured three bosses from the top of a tall tower.

788
01:32:05,750 --> 01:32:11,120
And I couldn't resist climbing the tower myself. Being able to get up-close to the carvings was a great privilege.

789
01:32:11,450 --> 01:32:16,730
From a height of eight meters, it's possible to see some of the carvings in a way which would be impossible from the ground.

790
01:32:17,180 --> 01:32:22,610
Like this cleric who appears to have been writing the same page for the last 540 years.

791
01:32:25,450 --> 01:32:29,230
From this perspective, we can also appreciate how delicate the bosses are.

792
01:32:29,410 --> 01:32:37,750
Some are almost completely hollow in structure. My colleagues at the Factum Foundation are continuing to work on this recording,

793
01:32:37,760 --> 01:32:42,470
with the final aim to be able to navigate the model in a virtual reality presentation.

794
01:32:43,730 --> 01:32:48,740
This initial preview gives us a sense of how the final and even higher definition recording would appear.

795
01:32:49,850 --> 01:32:51,840
In addition to the 3D presentation,

796
01:32:51,890 --> 01:32:59,090
an enormously high-resolution orthographic view has also been extracted from the photogrammetry and can be viewed online.

797
01:33:00,760 --> 01:33:07,210
Using this image, we can zoom into every detail of the ceiling at a resolution of 1000 pixels per square inch.

798
01:33:07,730 --> 01:33:12,640
And this is an incredible resource for the research of the 455 ceiling bosses.

799
01:33:13,510 --> 01:33:15,100
Each one appears to be unique.

800
01:33:15,880 --> 01:33:22,510
The extremely detailed elements, which can't be studied from the ground, can now be seen with clarity. From one of the bosses,

801
01:33:22,690 --> 01:33:26,110
it's even possible to count the feathers on the wings of the angels.

802
01:33:29,550 --> 01:33:32,460
Here we can see a boss featuring an Eagle and Child motif.

803
01:33:33,240 --> 01:33:39,030
And as we move past another showing a Tudor rose, we finally reach a very beautifully carved Green Man.

804
01:33:39,390 --> 01:33:43,200
One of the bosses which has been made into a physical facsimile.

805
01:33:46,820 --> 01:33:47,510
But to conclude,

806
01:33:47,510 --> 01:33:54,960
let's move from a recording of something enormous to a recording of something which is just a few millimetres in length. From this colour image.

807
01:33:55,010 --> 01:34:02,310
it's hard to make out what we're looking at. But from the render we can see the outline of a tiny fossilised wasp.

808
01:34:02,850 --> 01:34:10,290
And this is just one of dozens of specimens which were recorded at the Oxford University Museum of Natural History earlier this year.

809
01:34:12,880 --> 01:34:17,080
In order to make these recordings, a brand new portable version of the Selene was used.

810
01:34:19,330 --> 01:34:24,160
Taking just a few minutes to set up, recordings were made alongside curators and students.

811
01:34:26,110 --> 01:34:30,340
Through analysing the initial recordings together and understanding what they can reveal.

812
01:34:30,730 --> 01:34:36,970
They were able to support me in selecting specimens which would benefit from being recorded in this exciting new way.

813
01:34:39,900 --> 01:34:44,490
These are among the first recordings made with the Selene of objects from the natural world.

814
01:34:46,780 --> 01:34:48,340
And the results are very exciting.

815
01:34:51,410 --> 01:34:57,410
Fossils make incredible candidates for photometric stereo recording, and by being creative with the renders which we make.

816
01:34:58,960 --> 01:35:03,520
We can create representations which are both informative and also beautiful.

817
01:35:06,470 --> 01:35:10,670
This collection of fossilised leaves appear to have a flat, uniform surface.

818
01:35:12,680 --> 01:35:16,040
But the render allows us to see the topography as it truly is.

819
01:35:16,250 --> 01:35:26,150
Even the delicate veins are visible. Researchers at the museum were excited by the speed with which the recordings could be made and processed.

820
01:35:28,090 --> 01:35:38,860
And by the ways in which they could be analysed. Being able to study the specimens in this way may assist in identifying a given species.

821
01:35:39,640 --> 01:35:45,550
The renders can be easily shared, allowing for remote consultation with experts from other institutions.

822
01:35:49,740 --> 01:35:53,370
And as well as recording specimens from the Museums’ Earth collections,

823
01:35:53,760 --> 01:35:57,630
examples from their Life collections are equally fascinating and beautiful.​

824
01:36:00,460 --> 01:36:11,100
Here is a render which shows the tiny scales of a mantis. This composite render shows the surface of an unidentified snake skin.

825
01:36:12,000 --> 01:36:16,260
It was a specimen collected by the explorer and naturalist William John Burchell.

826
01:36:18,640 --> 01:36:24,700
Burchell described and collected over 63,000 specimens, most of which he gave to the British Museum,

827
01:36:25,090 --> 01:36:28,180
but several others were donated to Oxford’s Natural History Museum.

828
01:36:28,660 --> 01:36:32,230
These include a tortoise which he collected in 1814.​

829
01:36:35,430 --> 01:36:43,110
Here we have a colour recording of a Ulysses butterfly. And here's the render of the 3D surface.

830
01:36:46,720 --> 01:36:54,460
As is the case when recording dark regions of an oil painting, the surface details can be revealed very effectively in a 3D render.

831
01:36:55,180 --> 01:36:58,360
We can see how the scales on the left side of the lower wing,

832
01:37:00,130 --> 01:37:09,360
suddenly become visible. At the beginning of this presentation, I shared a high resolution recording of a macaw’s feather.

833
01:37:10,020 --> 01:37:13,230
In this example, we see a small African bird's wing.

834
01:37:15,880 --> 01:37:20,560
Every vane of this extremely delicate specimen can be seen in this atlas render.

835
01:37:23,970 --> 01:37:26,000
And this is the final recording I'd like to share.

836
01:37:26,270 --> 01:37:33,410
What we're looking at is a fragment from the Gibeon meteorite, which was collected in Namibia and fell to Earth in prehistoric times.

837
01:37:35,520 --> 01:37:39,299
Fine lines which have been recorded so successfully with the Selene, show

838
01:37:39,300 --> 01:37:47,010
the fragment's crystal composition. The patterns are the result of the meteorites having been slowly cooled from a molten state.

839
01:37:47,520 --> 01:37:51,660
This would have occurred in the vacuum of space when the solar system was first formed.

840
01:37:53,930 --> 01:37:56,240
Our collections are truly out of this world.

841
01:37:56,660 --> 01:38:02,450
Recording them as best as technology allows ensures that preservation and helps us to understand them more fully.

842
01:38:03,140 --> 01:38:07,490
These have been the aims of ARCHiOx, and will continue to use and develop this technology,

843
01:38:07,730 --> 01:38:13,250
so that both Oxford University and other institutions can benefit from their enormous potential.

844
01:38:15,240 --> 01:38:21,030
If you'd like to find out more about ARCHiOx, you'll find several articles on the Bodleian blog, which is called The Conveyor.

845
01:38:22,890 --> 01:38:28,530
I'd like to give special thanks to Dylan Schirmacher, who's my colleague who's worked on ARCHiOx with me,

846
01:38:29,100 --> 01:38:33,960
Elaine Anstee and Jorge Cano, who developed the Selene.

847
01:38:34,290 --> 01:38:45,349
Thank you very much. It's wonderful.

848
01:38:45,350 --> 01:38:50,149
John, thank you very much. So we've heard a lot about the work that's been happening at the Bodleian,

849
01:38:50,150 --> 01:38:56,120
and now we're going to move to the other half of the partnership to hear more about, um, work that's been ongoing in Factum.

850
01:38:56,420 --> 01:38:58,820
And welcome our next speaker, Carlos Bayod,

851
01:38:59,330 --> 01:39:06,230
who holds a PhD in art history and theory from Autonomous University of Madrid and a master's in architecture from Polytechnic University of Madrid.

852
01:39:06,830 --> 01:39:09,229
He is a project director at the Factom Foundation,

853
01:39:09,230 --> 01:39:13,969
and his work is dedicated to the development and application of digital technology to the conservation,

854
01:39:13,970 --> 01:39:18,650
study and dissemination of cultural heritage. With the Lucida 3D scanner,

855
01:39:18,980 --> 01:39:23,900
Carlos has recorded hundreds of artworks in the world's main museums, collections and archaeological sites.

856
01:39:24,350 --> 01:39:31,580
He's also a frequent speaker about facsimiles and museology, and has taught the Masters in Historic Preservation at Columbia University.

857
01:39:31,910 --> 01:39:36,290
So Carlos is going to talk us through some of the latest developments at Factum.

858
01:39:36,530 --> 01:39:39,910
Thanks Carlos. Thank you. Thank you Judith.

859
01:39:40,390 --> 01:39:41,140
Good afternoon.

860
01:39:43,090 --> 01:39:52,840
I would like to share today very quickly some examples of the type of work that we have been doing with the Selene scanner outside of Oxford.

861
01:39:53,350 --> 01:40:02,440
So at the same time as John was doing all these, beautiful recordings that have helped to, you know, bring these amazing discoveries,

862
01:40:03,040 --> 01:40:10,600
we have kept developing the system in order to adapt to the different requirements we have for, scanning,

863
01:40:10,960 --> 01:40:19,060
diverse range of artefacts, um, mostly from our base in Madrid, but also in museums all over the world.

864
01:40:19,570 --> 01:40:28,130
So here in the Bodleian, what happens is the Selene is fixed in the photography lab, and then objects come to be recorded.

865
01:40:28,570 --> 01:40:33,220
What we have been doing mostly is bringing the scanner to the institutions.

866
01:40:33,880 --> 01:40:41,080
So for example, when we were given the task of recording this important historic book in Madrid,

867
01:40:41,080 --> 01:40:46,750
in the Lázaro Galdiano museum, which is a book by Pacheco, Pacheco is the master, was the master of Velázquez.

868
01:40:48,570 --> 01:40:56,940
We just installed the Selene just for a couple days recording so we could capture as many pages as possible, both in colour and relief.

869
01:40:57,120 --> 01:41:04,170
We are talking about a historic book, not too different to Vasari's Lives of the Artists.

870
01:41:04,530 --> 01:41:14,760
Also, in this case, all the lives of Spanish artists, writers, play writers, etc. were portrayed that were drawn by Pacheco himself.

871
01:41:16,510 --> 01:41:22,270
Jorge Cano, has been developing and leading this work of creating the Selene,

872
01:41:22,270 --> 01:41:29,310
and also kept advancing both hardware and software in order to match new requirements and new challenges.

873
01:41:29,320 --> 01:41:36,610
So, for example, at the very beginning it was about controlling the scanner, literally writing lines of code.

874
01:41:37,570 --> 01:41:39,830
Probably at the beginning. This is what John had to do.

875
01:41:39,850 --> 01:41:47,410
Then over time, it's been about creating a proper user graphic user interface to make this operation more intuitively.

876
01:41:47,830 --> 01:41:52,880
So this is what we keep developing as we do the recordings.

877
01:41:52,900 --> 01:41:56,530
When we were working on this important book by Pacheco.

878
01:41:57,130 --> 01:42:01,770
Each page was unveiled in this manner.

879
01:42:01,780 --> 01:42:07,870
So as we have seen today, it's all about making accessible the materiality of books and other,

880
01:42:08,170 --> 01:42:13,780
artefacts through on screen visualisation, etc. but also,

881
01:42:13,780 --> 01:42:23,320
and I think it's important to highlight the Selene allows to obtain both colour and relief data that can be re materialised as facsimiles. As

882
01:42:24,340 --> 01:42:26,050
replicas that are physical,

883
01:42:26,620 --> 01:42:34,420
and this is what we did for creating a facsimile of this book that will facilitate researchers to have a more direct access without,

884
01:42:34,450 --> 01:42:38,110
putting in danger or in risk

885
01:42:38,320 --> 01:42:42,990
the original. This is especially important with objects like, for example,

886
01:42:42,990 --> 01:42:50,370
this series of important print blocks by Christophe Jagger in the um plant in the Plantin–Moretus Museum in Antwerp.

887
01:42:50,790 --> 01:42:54,740
Because this collection of printing blocks are actually toxic.

888
01:42:54,750 --> 01:43:02,489
So if you want to manipulate them, it's not possible to do it, let's say without proper, precaution measures including gloves,

889
01:43:02,490 --> 01:43:07,980
etc because of these specific treatment, they have to protect them from insects and so on.

890
01:43:08,610 --> 01:43:19,080
So in order to record this beautiful, originals, we have to develop what we can refer to as a portable version of the Selene.

891
01:43:19,680 --> 01:43:27,720
As you could see, the core of the system, which is the camera with four light units, can also be mounted on a tripod.

892
01:43:27,990 --> 01:43:37,260
And then with a bit more of work, it's possible to record smaller areas you know, more or less manual way.

893
01:43:45,450 --> 01:43:49,350
These type of recordings allow a new access to this printing blocks, as you can see.

894
01:43:50,010 --> 01:43:57,450
It would eventually allow a re materialisation of them or even new prints if necessary, if that was the desire by the museum.

895
01:43:58,140 --> 01:44:02,480
And then because we are talking about very dark elements

896
01:44:02,510 --> 01:44:10,370
it is only through this idea of removing the colour and allowing to see them, and actual bas reliefs or low relief objects,

897
01:44:10,610 --> 01:44:17,150
that we get to understand all their complexity, all these curly structures on the animal.

898
01:44:20,420 --> 01:44:28,100
This curly structures are somehow similar to what we found when we started working at De Lakenhal Museum in Leiden.

899
01:44:29,060 --> 01:44:33,590
This is a state of a series of sample books as Adam was showing us earlier.

900
01:44:34,040 --> 01:44:44,209
It's a collection of about 30 volumes from throughout the 18th century, and they contained a like an inventory of samples of fabrics

901
01:44:44,210 --> 01:44:54,020
so these like different types of cloth with different tints and for, because of the condition that they are in,

902
01:44:54,020 --> 01:45:01,910
it's very difficult, let's say, impossible for researchers to get to read what is written in this archival records.

903
01:45:02,600 --> 01:45:08,010
This is why we approach this work by combining two different recording methods.

904
01:45:08,030 --> 01:45:13,290
One, as you can see on the image, the actual portable Selene scanner,

905
01:45:13,310 --> 01:45:16,640
so we could get very close in to understanding the

906
01:45:17,680 --> 01:45:21,550
texture and surface qualities of this fabric samples.

907
01:45:22,120 --> 01:45:28,450
We can even get closer and eventually get a proper archive of these different surfaces,

908
01:45:29,350 --> 01:45:36,940
but also we understood the book as a volumetric object, almost like a sculpture that needed to be recorded with photogrammetry.

909
01:45:37,420 --> 01:45:46,060
So once we have recorded, let's say, a book open, in this particular page, we have this 3D model,

910
01:45:47,010 --> 01:45:51,420
and then we developed a method for virtually flatten these pages.

911
01:45:52,820 --> 01:46:02,510
Was having to do it on the original, it was possible to carry out a kind of virtual flattening so that the text would become readable,

912
01:46:03,460 --> 01:46:12,190
and then with a bit more, um, image processing, which would require more refinement, make some text really visible again by researchers.

913
01:46:13,720 --> 01:46:24,700
This idea of flattening volumetric objects and highlighting this double quality of both 3D and two and a half the qualities,

914
01:46:24,880 --> 01:46:31,390
is also something that we applied for recording this navigation globe, in this case in The Centraal Museum in Utrecht.

915
01:46:31,900 --> 01:46:37,870
Again, the combination of two technologies portable Selene recording, but also photogrammetry.

916
01:46:37,990 --> 01:46:42,910
So we are at the same time understanding this globe as a sphere, what it is,

917
01:46:43,540 --> 01:46:49,780
but also trying to get an even closer look to the skin of the object.

918
01:46:50,230 --> 01:46:55,690
So with the Selene scanner, we were focusing on certain details along the surface of this globe,

919
01:46:56,320 --> 01:47:04,210
and then we would combine those,  , techniques, those types of information to layers together on the same file.

920
01:47:04,540 --> 01:47:13,170
The way we did it is, trying to imagine how this globe was built in the first time, in the first the stage.

921
01:47:13,180 --> 01:47:24,340
So basically a kind of unwrapping or unfolding of the sphere surface through digital means so that it could be extended as a flat area.

922
01:47:25,880 --> 01:47:31,190
And then it's possible also to map one layer of data onto the other.

923
01:47:31,790 --> 01:47:37,100
What we can see on the image right now is like the central area which contains even higher detail,

924
01:47:37,520 --> 01:47:41,750
is the Selene information on top of the photogrammetry information.

925
01:47:43,380 --> 01:47:53,040
All this multi-layer file, this kind of stack of layers of information is what we provide to the museums so it can assist them in the process of,

926
01:47:53,760 --> 01:47:56,990
conserving the original or any treatment,

927
01:47:57,000 --> 01:47:59,700
it's going to happen on this particular object.

928
01:48:00,000 --> 01:48:05,760
And eventually, the same way as we unwrapped the surface of the sphere, it will be possible to use the same method

929
01:48:06,060 --> 01:48:09,750
if we wanted to re materialise this object. That's a facsimile.

930
01:48:12,240 --> 01:48:17,770
So we have been bringing the Selene scanner as a portable version to key institutions

931
01:48:18,160 --> 01:48:23,500
that somehow were in need of unveiling the three dimensionality of their collections.

932
01:48:24,500 --> 01:48:29,630
The important collections of original plates by Goya at Calcografía Nacional in Madrid.

933
01:48:30,620 --> 01:48:37,909
It's a relevant project because it is only very recent that they managed to remove a coating that was,

934
01:48:37,910 --> 01:48:44,840
in a way, covering the original copper of these plates. It was a complex process that now, for the first time in many years,

935
01:48:44,840 --> 01:48:52,160
it is possible to go to this place and see the surface of the plates a closer to the way Goya made them.

936
01:48:53,270 --> 01:49:01,790
And thanks to the Selene scanner, it's possible to understand, exactly what's the information contained in them.

937
01:49:01,970 --> 01:49:08,000
Copper plates as shiny, as these one are probably some of the most challenging and difficult objects to photograph.

938
01:49:08,240 --> 01:49:12,950
You don't really get a proper image of what is represented in there

939
01:49:13,130 --> 01:49:21,740
and as John was showing earlier, the advantage of saline is to remove for one moment any call or information if you want,

940
01:49:21,740 --> 01:49:27,430
and then focus on the actual relief data. Like what these images are showing.

941
01:49:27,880 --> 01:49:34,460
So thanks to the Selene, it was great opportunity to be able to read this,  

942
01:49:34,850 --> 01:49:43,990
plates eventually being able also to produce new additions, if that's, again, what the museum wanted.

943
01:49:44,950 --> 01:49:49,600
But also, we had a big challenge with objects like this once when we are talking about with metallic objects,

944
01:49:50,590 --> 01:49:59,140
there is always a bit of, um, excessive reflection that could affect the quality of the data from the Selene.

945
01:50:00,040 --> 01:50:05,200
And this is why we tried a new development, the new improvement in the system,

946
01:50:05,560 --> 01:50:11,890
which is creating a system of cross polarisation filters that could be adapted to the Selene scanner.

947
01:50:12,310 --> 01:50:20,770
We are talking about installing a special set of filters, both on the light units, on the flash units and on the lens on the objective,

948
01:50:21,310 --> 01:50:27,610
so that any excess of reflection can be balanced, can be somehow masked, let's say

949
01:50:27,820 --> 01:50:33,880
and then we would obtain more accurate results, not in the depiction of the texture,

950
01:50:34,120 --> 01:50:42,490
but especially in avoiding, um, an artificial recreation of fake shape.

951
01:50:44,000 --> 01:50:47,090
A few images to explain this improvement,

952
01:50:47,090 --> 01:50:57,550
so if we are recording a very flat metallic plate thanks to the cross polarisation filters, we are getting a very accurate and precise and

953
01:50:58,520 --> 01:51:05,780
the best correspondence between the data we obtain in digital and the actual shape of the plate.

954
01:51:06,710 --> 01:51:12,890
If we are not applying this system, of course polarising, we might get, let's say,

955
01:51:12,890 --> 01:51:18,230
fake data because of the reflections, because of this shiny material qualities,

956
01:51:18,230 --> 01:51:23,000
and maybe extract a shape that does not correspond to reality.

957
01:51:24,120 --> 01:51:29,820
So this is part of the ongoing development that we are carrying out in order to make

958
01:51:30,180 --> 01:51:36,390
the Selene capable of getting to record more and more diverse types of origins.

959
01:51:37,580 --> 01:51:47,810
Another way of presenting this same object without or with this accessories would provide an even better and even more accurate depiction of it.

960
01:51:49,470 --> 01:51:53,490
Once we developed this new improvement, it was possible, for example,

961
01:51:53,490 --> 01:52:00,899
to go to the Museum of Antiquities in Leiden and then record a series of highly reflective objects like this,

962
01:52:00,900 --> 01:52:05,520
series of originals from the Middle Eastern department, like this metallic shield,

963
01:52:06,090 --> 01:52:11,489
and then obtain information that it's accurate both in texture and in the shape,

964
01:52:11,490 --> 01:52:20,350
in the overall shape. Also, as part of this project, we have the opportunity of using the Selene for capturing a series of cylinders.

965
01:52:20,380 --> 01:52:22,870
Not too different to the ones that John was showing earlier.

966
01:52:23,140 --> 01:52:33,540
In this case, by rotating the cylinder in a few, um, few sections of angles so that we can complete the entire area,

967
01:52:33,550 --> 01:52:38,890
external area with the Selene, and then maybe make it as an

968
01:52:38,890 --> 01:52:43,840
educational tool, something that could be used for museums to have tactile versions of them,

969
01:52:44,200 --> 01:52:51,250
reproductions for, um, new engagement with the original without touching, um, the real object.

970
01:52:52,810 --> 01:52:58,330
So the Selene Circle is the name we have been given to this community.

971
01:52:58,450 --> 01:53:02,440
We are somehow nurturing with institutions all over the world.

972
01:53:03,400 --> 01:53:08,440
It's about creating a network of users of the Selene scanner.

973
01:53:09,100 --> 01:53:16,749
We have the opportunity of sharing with what started as a direct dialogue between Madrid and Oxford,

974
01:53:16,750 --> 01:53:22,059
between Factum and the Bodleian Library is now including new actors,

975
01:53:22,060 --> 01:53:27,490
such as Princeton University Library, that since the summer of 2023,

976
01:53:27,490 --> 01:53:31,840
they also own a Selene scanner that they are using it for recording their collection,

977
01:53:32,740 --> 01:53:40,890
and it's also very interesting to see how other institutions are applying their systems for their own means.

978
01:53:40,900 --> 01:53:47,500
So for instance, in the actual platform, the website or the digital platform of the Princeton University Library,

979
01:53:47,500 --> 01:53:53,739
they are presenting recordings like this, like these are the Joseph's screenshots from their platforms,

980
01:53:53,740 --> 01:54:04,330
and we can see, um, different layers that you can navigate both to highlight the colour on one hand and the 3D without the colour on the other,

981
01:54:04,630 --> 01:54:06,250
or the composite of both.

982
01:54:06,580 --> 01:54:14,920
So again, this is giving access to libraries to understand books and flat objects as three dimensional objects for the first time.

983
01:54:20,760 --> 01:54:25,830
We are also very proud of having British Museum as one of our partners in this Selene circle.

984
01:54:25,890 --> 01:54:32,700
So also since last summer we were installing the Selene scanner and

985
01:54:33,180 --> 01:54:38,130
making it available for them to embark in a very challenging project of recording,

986
01:54:38,610 --> 01:54:43,290
hundreds of thousands of clay tablets with cuneiform texts on them.

987
01:54:44,490 --> 01:54:51,360
In addition to installing the system, we are, of course, providing the training, the necessary technical support and

988
01:54:51,570 --> 01:55:00,600
let's say conceptual support that that's what we are trying to provide in this network of users so that every new requirement,

989
01:55:00,600 --> 01:55:07,320
any need or problem or new challenge that present in the process of recording their collections,

990
01:55:07,620 --> 01:55:12,150
we can help develop the necessary tools to make it more accessible.

991
01:55:13,020 --> 01:55:21,749
So this is a project that is in a way starting, and it's already required for us to go back and apply certain modifications to the hardware,

992
01:55:21,750 --> 01:55:29,730
to the structure of the settings, because, for example, they were telling us that they rather be working, standing instead of sitting on a desk,

993
01:55:29,880 --> 01:55:37,620
so things as simple as that might be the difference between a proper systematic recording or having it less useful.

994
01:55:37,620 --> 01:55:42,330
When we install a system like this in an institution, we want to make sure it's being used.

995
01:55:42,540 --> 01:55:46,560
We don't want it to be, you know under used for any reason.

996
01:55:48,990 --> 01:55:52,920
The Institute of Valencia de Don Juan is a unique institution in Madrid.

997
01:55:53,160 --> 01:56:03,510
It is a place that is actually not open to the public, but it's open for researchers who want to go and inspect the their incredible collection,

998
01:56:03,510 --> 01:56:11,100
mostly of medieval originals in Spain, including ceramics, for example,

999
01:56:11,640 --> 01:56:20,760
that for a period of about two months, we have a team working with their stuff for capturing as many objects as possible,

1000
01:56:20,790 --> 01:56:26,700
so that, again, researchers can access this information without having to get to the originals,

1001
01:56:27,330 --> 01:56:39,790
therefore making it possible to protect them better. Ceramics, textiles, etc are just some of the unique objects in this collection.

1002
01:56:39,830 --> 01:56:45,290
This will be this collaboration will stay for the next two weeks as well,

1003
01:56:45,590 --> 01:56:53,000
and we will be able to provide a proper overview of the different capabilities of this system for different objects,

1004
01:56:55,260 --> 01:57:00,680
and then as part of our work is the dissemination of the system.

1005
01:57:00,690 --> 01:57:10,740
We have been discussing the technology in different locations in the Getty Museum in Los Angeles, in Madrid, La Biblioteca Nacional.

1006
01:57:11,370 --> 01:57:18,510
John was with Jorge Cano presenting, for example, in the 2 and 3D photography conference in the Rijks museum in Amsterdam.

1007
01:57:19,080 --> 01:57:25,740
All these events are important to gather and to capture the interest and attention of potential users

1008
01:57:26,130 --> 01:57:35,070
and even more important to that is the workshops is the training initiatives that can be shaped around the Selene scanner.

1009
01:57:35,310 --> 01:57:40,350
So it is incredible to see how not only students but also young professionals,

1010
01:57:40,650 --> 01:57:48,139
technicians in church in charge of imaging in different institutions are engaged with

1011
01:57:48,140 --> 01:57:52,790
this system because it's allowing to see things that it was not possible before.

1012
01:57:53,510 --> 01:57:59,450
Workshops with the University of Complutense in Madrid, with Granada University,

1013
01:57:59,810 --> 01:58:07,550
also with our partner Initiative Archive in Venice, in which we were also capable of presenting the system and

1014
01:58:08,560 --> 01:58:11,830
this will for sure will have a continuation in the next months.

1015
01:58:13,150 --> 01:58:17,170
So at the end, the Selene project, 

1016
01:58:17,320 --> 01:58:27,309
it has been a great opportunity for getting to work with some of the most exciting institutions dedicated to to be custodians of this unique

1017
01:58:27,310 --> 01:58:37,120
objects and get new access and shed a new light on these unique artefacts and allowing researchers to engage them in a different way,

1018
01:58:37,780 --> 01:58:44,890
but especially, it's been a reason for getting in touch with incredible professionals and the best

1019
01:58:45,340 --> 01:58:55,360
the best group that we are, we are creating the best network, the best community and being able to share all this excitement with all of them.

1020
01:58:57,480 --> 01:59:05,670
This is an ongoing project, and of course we will keep working and developing the tool over the next months, over the next years,

1021
01:59:06,150 --> 01:59:17,030
and two of the possible ways in which this system will be going is, for example, trying to get even closer to the objects, even deeper.

1022
01:59:17,660 --> 01:59:27,020
When we are changing the lens with, with a different for example, in this case from 50 to 100 or one 80mm,

1023
01:59:27,230 --> 01:59:36,020
it's possible to get to see micro mosaics, even with higher detail, with higher resolution, but also getting higher.

1024
01:59:36,230 --> 01:59:42,740
So one of the latest improvements is being able to put the Selene not working horizontally,

1025
01:59:42,740 --> 01:59:51,650
but working vertically on a kind of telescopic mast so that it's possible to record wall paintings,

1026
01:59:52,010 --> 01:59:56,060
easel paintings, elements that are actually on a vertical position,

1027
01:59:56,570 --> 02:00:08,060
and then extending the possibilities of discovering new stories in texture of things beyond, um, the walls of a library, for example.

1028
02:00:09,940 --> 02:00:19,850
Thank you very much.

1029
02:00:24,060 --> 02:00:27,960
Okay. Um, third and final speaker for this session,

1030
02:00:28,470 --> 02:00:33,420
welcome, Richard Allen, who is a software engineer in Bodleian Digital Library Systems and Services,

1031
02:00:33,720 --> 02:00:38,610
where he plays a role in developing the digital volume platform and a variety of internal tools.

1032
02:00:39,240 --> 02:00:44,910
He's the main developer for the Bodleian libraries' contribution to ARCHiOx, and in addition to his work in the university,

1033
02:00:44,910 --> 02:00:48,420
Richard is the founder of a university spin-out company called PalaeoPi,

1034
02:00:48,780 --> 02:00:54,240
which you heard John referenced earlier that specialises in cutting-edge photogrammetry hardware.

1035
02:00:54,570 --> 02:00:56,830
And so Richard is going to talk to us about

1036
02:00:56,850 --> 02:01:02,760
the work that we've been doing to integrate all of this work with our IIIF and image delivery workflows in the Bodleian.

1037
02:01:06,010 --> 02:01:09,350
introduce, so I begin. Today,

1038
02:01:09,350 --> 02:01:13,729
I'm going to guide you through the work that my team has done over the last two

1039
02:01:13,730 --> 02:01:17,900
years to make sure that we could beat the amazing work we did in part one.

1040
02:01:19,100 --> 02:01:24,710
the focus this time around was more on developing ARCHiOx as a service and getting more items online,

1041
02:01:25,340 --> 02:01:29,570
and my hope is that by the end of this talk, it'll be clear to you that we've done a lot of work.

1042
02:01:29,840 --> 02:01:31,280
A hero's journey, if you will,

1043
02:01:31,850 --> 02:01:38,660
and the ARCHiOx part two is up there with the sequels better than the originals, in my opinion, like the movie Aliens.

1044
02:01:39,350 --> 02:01:45,800
Um, so without further ado, I, I believe the last time I was here, I left you on a cliff-hanger.

1045
02:01:46,430 --> 02:01:50,840
Um, would Richard get the tiles loading seamlessly or wouldn't he?

1046
02:01:50,870 --> 02:01:54,200
Well, I did, but it was surprisingly difficult.

1047
02:01:54,320 --> 02:01:58,580
Um, but first, before I talk about a bunch of technical stuff, we need a quick recap.

1048
02:01:59,390 --> 02:02:03,680
Um, you've seen some of these slides earlier, in the other talks, but,

1049
02:02:04,920 --> 02:02:08,249
for the benefit of people who don't have much experience with this,

1050
02:02:08,250 --> 02:02:18,479
I'll go over some of them again. For the last two years, we've been working very closely with Factum to develop, the photometric stereo,

1051
02:02:18,480 --> 02:02:23,219
to capture the 3Dness of shallow relief objects and record them in a 2D format,

1052
02:02:23,220 --> 02:02:30,480
and here you can see Selene again, which is, probably the best machine on the market for what it does.

1053
02:02:32,140 --> 02:02:41,770
Basic principle is a series of four lights, you eliminate the subject whilst a high resolution camera takes a photo of four photos.

1054
02:02:42,340 --> 02:02:48,790
Um, these photos are combined in a process processing processing application written by Jorge Cano,

1055
02:02:48,790 --> 02:02:53,770
to output a set of derivatives and normal map, and albedo map and a depth map.

1056
02:02:54,520 --> 02:02:59,470
again, for the benefit of those who are new to this project, I hear you ask, what is a normal map?

1057
02:03:00,040 --> 02:03:07,950
A normal map makes up makes use of a two dimensional image file to encode the direction that a normal would face at each pixel,

1058
02:03:07,960 --> 02:03:12,310
using the RBG values to plot a 3D vector.

1059
02:03:12,320 --> 02:03:18,520
In other words, a normal map stores the orientation of a surface at each pixel location inside an RGB image.

1060
02:03:19,090 --> 02:03:23,080
here you can see a normal map of one of our wax seals,

1061
02:03:24,480 --> 02:03:33,690
and once we have this information encoded in a 2D format, we can blend it with a colour map of the same image, also known as an albedo map,

1062
02:03:33,690 --> 02:03:41,879
in a special graphical program called a Shader, and render and relight the object in real time via the browser like,

1063
02:03:41,880 --> 02:03:46,080
Google Chrome or Firefox, using something called WebGL.

1064
02:03:47,590 --> 02:03:54,220
The photo on the right is a scene of one of our copper plates from an earlier test object on our QA system.

1065
02:03:56,560 --> 02:04:00,730
And we achieve this by, my line manager, Andy,

1066
02:04:00,910 --> 02:04:01,990
you might meet him later,

1067
02:04:02,290 --> 02:04:10,390
writing a small extension to the IIIF presentation API to feed into our manifest labels that a computer would understand,

1068
02:04:10,840 --> 02:04:16,210
so that our Shader program could get the, um, correct image to process in real time.

1069
02:04:16,840 --> 02:04:20,700
and here you can see that it's all about context.

1070
02:04:20,790 --> 02:04:29,370
we use a context and it allows us to label an image as a depth map, normal map and an albedo map, a composite render and a shaded render,

1071
02:04:29,880 --> 02:04:37,410
and we are actually thinking of getting rid of the last two, since they aren't really maps in in the sense of the technology.

1072
02:04:37,770 --> 02:04:41,370
But early on, being able to label all of them was really useful.

1073
02:04:41,730 --> 02:04:44,850
skipping ahead to the end of the first part of the project,

1074
02:04:44,850 --> 02:04:51,150
and here you can see the demo video I did on the 5th of December 2022 displaying the fruits of our labour,

1075
02:04:51,480 --> 02:04:58,680
a plug-in for the IIIF called Mirador, rendering a normal map and albedo a map combined in real time with interactive lighting.

1076
02:05:00,590 --> 02:05:08,960
so back to the cliff-hanger. The first thing we aimed to solve when I last gave the talk was the efficient loading of images.

1077
02:05:09,290 --> 02:05:16,850
You may remember that to deliver gigapixel images efficiently over the internet, you need to have a dividing conquer strategy,

1078
02:05:17,240 --> 02:05:21,080
by chopping up the images into tiny squares called tiles, 

1079
02:05:21,530 --> 02:05:28,370
and by doing this, you can deliver them in parallel and get around the texture size limits of this thing called WebGL,

1080
02:05:28,370 --> 02:05:33,260
which does the rendering in the browser, well, we had to achieve this

1081
02:05:33,320 --> 02:05:40,730
but we were showing people a loading screen and we were also replicating the work viewer was doing already,

1082
02:05:41,360 --> 02:05:48,320
so essentially duplicating the gigapixel amount of data twice and that needed to be gotten rid of

1083
02:05:48,410 --> 02:05:54,160
then there was also the goal of trying out different shaders and perhaps write our own.

1084
02:05:54,170 --> 02:06:00,650
And so,  , readers could experience a higher level of, of authenticity when engaging with the objects.

1085
02:06:01,220 --> 02:06:07,430
those were my aims at the end of the last project. However, it wouldn't have been a very good project if it was just about me,

1086
02:06:07,850 --> 02:06:14,180
so the main focus of part two was to get ready for transition, for ARCHiOx to become a service.

1087
02:06:14,690 --> 02:06:20,749
And to do so, we needed to also work out the back end tools for delivering content and ironed

1088
02:06:20,750 --> 02:06:24,680
out any issues that got in the way of quality of media and user experience.

1089
02:06:26,040 --> 02:06:30,540
Anyway, regardless of my lofty goals from the last time

1090
02:06:30,990 --> 02:06:40,310
What happened? Well, a few weeks after the December 5th talk, I got back to work and quickly came up with a method to load in the images,  ,

1091
02:06:40,320 --> 02:06:48,990
asynchronously piggybacking off the, um, the viewer that we were plugging into which is called OpenSeadragon.

1092
02:06:49,410 --> 02:06:54,330
and then we were now managing to load the data once.

1093
02:06:54,360 --> 02:06:57,870
and Tim and Andy were now very happy in my work was done.

1094
02:06:58,200 --> 02:07:05,430
Except it wasn't and I'll tell you why later. My next wanted to work on making our own shaders with Jorge at Factum.

1095
02:07:05,490 --> 02:07:12,270
we began having weekly meetings together, working on the shader code in a test application he wrote called Scan and Shade.

1096
02:07:12,840 --> 02:07:17,370
However, once we started meeting and got talking, it turned out we were both talking about the same thing.

1097
02:07:17,820 --> 02:07:20,520
PBR shading or physical based rendering

1098
02:07:21,870 --> 02:07:28,650
So instead of rewriting our own shaders, we realised very quickly that these shaders were already available to us in the tools that we were using.

1099
02:07:29,130 --> 02:07:32,220
let's pause for a moment. What is PBR?

1100
02:07:32,460 --> 02:07:39,510
Well, it's a shading technique that aims to recreate realistic renders of everyday objects using a set of simplified principles

1101
02:07:39,990 --> 02:07:43,980
basically dividing materials into rough or smooth metal or non-metal.

1102
02:07:44,550 --> 02:07:50,070
and you saw an example of that on the Bible that Dylan imaged using photogrammetry.

1103
02:07:50,610 --> 02:07:57,030
now the shaders we were already using and just kind of shade had some of these properties available to us

1104
02:07:57,030 --> 02:08:03,540
so we started playing around and I worked with Jorge getting it's kind of hard to work on multiple image tiles,

1105
02:08:03,540 --> 02:08:08,880
and soon discovered a problem with renders coming from the Bodleian imaging server looked terrible.

1106
02:08:10,440 --> 02:08:22,019
Why did Jorge look so good but ours look so bad?

1107
02:08:22,020 --> 02:08:27,110
I sat and looked at the screen a while, perplexed. The render

1108
02:08:27,290 --> 02:08:32,030
whenever I changed, the shader parameters looked very grainy, very blocky, very ugly,

1109
02:08:32,480 --> 02:08:36,250
and then it dawned on me I was seeing pixels lump together with their neighbours.

1110
02:08:36,320 --> 02:08:43,830
This was lossless compression of images, and you can see it clearly in this image with the depth exaggerated.

1111
02:08:43,850 --> 02:08:47,660
You can see it's very blocky. That is a JPEG image.

1112
02:08:48,650 --> 02:08:54,000
now what is lossless compression, I hear you ask? This picture demonstrates it quite well

1113
02:08:54,030 --> 02:09:00,020
from left to right, the image is compressed progressively and you can immediately see the effects.

1114
02:09:00,030 --> 02:09:04,530
It's basically data loss and something like compressing our normal maps.

1115
02:09:05,990 --> 02:09:12,980
Well, so what? We use JPEGs all the time for other digital modelling items, and they look great, I hear you say.

1116
02:09:13,610 --> 02:09:16,990
However, you may remember back to one of our previous slides.

1117
02:09:17,000 --> 02:09:20,660
A normal map stores information on the way a surface interacts with light.

1118
02:09:20,960 --> 02:09:25,340
So data loss is not what you want to see, when you want to relight your scene.

1119
02:09:27,940 --> 02:09:32,940
However, this is a well-established problem in internet technology and photo and photography,

1120
02:09:32,950 --> 02:09:36,640
so there is a solution. Don't use a lossless image format

1121
02:09:36,850 --> 02:09:42,670
and to solve this, we need to figure out how we could get Mirador to feed lossless images into our Shader program.

1122
02:09:43,330 --> 02:09:44,940
So we started with IIIF,

1123
02:09:44,950 --> 02:09:53,440
and the way to do this already existed to in the form of something called preferred formats via the IIIF Image API version three.

1124
02:09:53,680 --> 02:10:03,670
Easy, right? Well, not exactly. We use IIPImage to deliver our images in tiles, and therefore there were some constraints.

1125
02:10:03,890 --> 02:10:07,090
We could only choose a format that IPPImage supported.

1126
02:10:07,630 --> 02:10:11,410
The format had to be lossless, and we had to be able to deliver them via the web.

1127
02:10:11,980 --> 02:10:16,090
So final file, final tile size was also an issue.

1128
02:10:16,870 --> 02:10:20,470
Here you can see a PNG is the best in terms of quality,

1129
02:10:21,400 --> 02:10:24,440
but look at the file size compared to JPEG or webP,

1130
02:10:24,580 --> 02:10:31,630
it's an order of magnitude greater. We tried PING first, an easy change because the stack already supported the delivery of them.

1131
02:10:31,930 --> 02:10:33,370
The images looked amazing,

1132
02:10:33,760 --> 02:10:39,940
but looking at the network traffic for the file streaming, it quickly became apparent we couldn't do this in a production environment.

1133
02:10:40,510 --> 02:10:49,870
So I looked into expanding our stack to serve, um, the webP format, um, a format that afforded lossless compression for a similar size to JPEG.

1134
02:10:51,390 --> 02:10:57,900
which turned out to be a full stack change, meaning some part of every part of our system architecture needed to be changed

1135
02:10:58,380 --> 02:11:07,050
and everybody loves those. I had to update the plugin several times, including messing around with something called Content Negotiation,

1136
02:11:07,440 --> 02:11:10,950
which I will note now came back to haunt us later.

1137
02:11:11,880 --> 02:11:18,450
I had to build our own version of OpenSeaDragon to add in a one word change to get it to support the new format

1138
02:11:19,080 --> 02:11:24,630
we were not that, that far at the cutting edge that they hadn't gotten around to updating it yet.

1139
02:11:25,320 --> 02:11:30,840
I then had to install and compile the dependency for IIPImage that makes webP tiles,

1140
02:11:31,110 --> 02:11:34,980
and ask for the developer of it to make some changes to support this

1141
02:11:35,520 --> 02:11:40,139
after weeks of work, basically, um, the quality

1142
02:11:40,140 --> 02:11:45,390
and then you'll say that the quality problem was solved. No more data loss in our normal maps.

1143
02:11:45,420 --> 02:11:50,190
Huzzah! However, my celebration was short lived

1144
02:11:50,250 --> 02:11:53,670
there was now another issue. We noticed the relighting.

1145
02:11:53,680 --> 02:11:56,760
when relighting the images, something weird was going on with the lights

1146
02:11:57,840 --> 02:12:03,120
they were inverted on the y axis. In other words, the light moved in the opposite direction.

1147
02:12:03,150 --> 02:12:10,560
Up was down, and down was up. After some research, it turned out that normal maps have two formats direct X and OpenGL.

1148
02:12:11,310 --> 02:12:16,830
Our renderer uses OpenGL, and the Selene processing app was producing normal maps in direct X.

1149
02:12:17,220 --> 02:12:21,900
A classic problem that continues to haunt newbies in the field every day

1150
02:12:22,350 --> 02:12:29,470
and this is one reason we need standards. Um, here you can see the impact of using one format versus the other.

1151
02:12:29,490 --> 02:12:36,080
You need to get this right because an inward facing facet becomes an outward facing facet, and vice versa,

1152
02:12:36,090 --> 02:12:40,080
when using a direct x normal map in an OpenGL renderer.

1153
02:12:40,770 --> 02:12:45,480
The only difference between the two is the green channel. On the other, BGR is inverted.

1154
02:12:47,260 --> 02:12:55,900
The problem we were having is that unless you have a visual inspection process, there is no way to know which format your normal map is in

1155
02:12:56,530 --> 02:13:00,550
until you until you can actually interact with it.

1156
02:13:01,090 --> 02:13:07,090
There was also, the possibility of one image being recorded in one orientation having the other axis inverted

1157
02:13:07,090 --> 02:13:07,510
too

1158
02:13:07,990 --> 02:13:16,330
we decided that the best thing to do was to correct the rotation and invert the y channel of the normal maps somewhere in the processing pipeline.

1159
02:13:17,170 --> 02:13:19,060
We came up with multiple ways to do this

1160
02:13:19,090 --> 02:13:26,380
first we expanded our lighting map extension, and then this allowed other users to use whatever format they want.

1161
02:13:27,010 --> 02:13:31,620
And then John was able to correct the rotation based on the advice of Jorge.

1162
02:13:32,200 --> 02:13:36,279
And since then,

1163
02:13:36,280 --> 02:13:42,129
he's updated the processing application for the solution that allows users to choose which format

1164
02:13:42,130 --> 02:13:47,770
they want to and export normal maps in and embedded so that a machine can read what format it's in.

1165
02:13:47,950 --> 02:13:51,310
Getting around the problem of having to do visual inspections.

1166
02:13:56,850 --> 02:14:04,799
Weeks of work are relighting our works as expected, and now lighting is the right way round, and although the process is still manual today,

1167
02:14:04,800 --> 02:14:10,200
there's still a little bit of this going on with Tim, Andy and I staring at images as they come online.

1168
02:14:10,530 --> 02:14:15,209
We, in theory, know how to automate this now, so this is going to be.

1169
02:14:15,210 --> 02:14:22,380
In parallel to these issues, I also spent days looking at how we could feed in start-up parameters to our shader so that each object,

1170
02:14:23,160 --> 02:14:27,180
had the properties of the shaders encoded in the manifest.

1171
02:14:27,180 --> 02:14:28,740
And this is still a work in progress.

1172
02:14:29,130 --> 02:14:34,709
And but a lot of the homework has already been done in order to make something that people in libraries and,

1173
02:14:34,710 --> 02:14:38,760
archives can find useful, we need to use something called controlled vocabulary.

1174
02:14:39,000 --> 02:14:43,440
So we need to figure out library ways of describing these technical things

1175
02:14:43,500 --> 02:14:51,570
and you can see in the next slide, we've also expanded our lighting map extension of IIIF to begin to

1176
02:14:52,520 --> 02:14:53,300
talk about these

1177
02:14:53,810 --> 02:15:02,480
we have a conductor, which means an object that's 100% metal, dielectric, which means something which is should be set at zero metal.

1178
02:15:02,930 --> 02:15:06,880
then there are words that comment on how shiny an object is

1179
02:15:06,890 --> 02:15:14,090
and the idea being that we can create a set of shininess depending on the level that an object needs when it first loads in the viewer.

1180
02:15:14,480 --> 02:15:18,740
none of this is in place in our infrastructure yet, but

1181
02:15:19,100 --> 02:15:25,250
because I think we should seek the advice and feedback of the IIIF community before deciding on the controlled vocabulary we use.

1182
02:15:27,170 --> 02:15:31,100
And the next major part of the project this time around was UX and user testing.

1183
02:15:31,210 --> 02:15:39,770
Tim and Ruth made a lot of, um, putting a lot of effort in arranging for a plugin to be tested by experts on the objects we were digitising.

1184
02:15:40,490 --> 02:15:44,840
It was compared against similar tools, and through the process we got a lot of positive feedback,

1185
02:15:44,840 --> 02:15:50,510
but also some critical feedback to help really make our tool something useful to them as researchers or curators.

1186
02:15:50,990 --> 02:15:55,190
And I had already anticipated some of these issues arising from this exercise.

1187
02:15:55,430 --> 02:16:02,210
And so I was working on this when disaster struck. Out of nowhere, a bug in Google Chrome,

1188
02:16:02,600 --> 02:16:06,940
propped up with an update that added a black screen of death,

1189
02:16:06,950 --> 02:16:16,910
basically that required a user to hard reload the browser plug in and something that became was it was happening during testing.

1190
02:16:17,300 --> 02:16:21,379
Now, although this didn't break the plugin because you could technically reload it from the cache,

1191
02:16:21,380 --> 02:16:24,980
is unacceptable for the user to have to know that.

1192
02:16:25,310 --> 02:16:29,900
So we spent a huge amount of time trying to figure this out, and we finally did.

1193
02:16:30,290 --> 02:16:33,530
And here's a joke. How many software engineers did it take to find a bug?

1194
02:16:33,560 --> 02:16:36,049
the answer is three days,

1195
02:16:36,050 --> 02:16:42,710
one engineer, one senior systems architect, one digitisation and digital engagement manager, one project manager, and one Judith Siefring.

1196
02:16:44,540 --> 02:16:49,159
It turns out that Chromium browser doesn't respect content negotiation like normal.

1197
02:16:49,160 --> 02:16:57,210
And, if you first nav it, if you navigate from something like Digital Bodleian and using version two, it remembers that it's used version two.

1198
02:16:57,230 --> 02:17:04,880
So when you try and load version three, it it doesn't know the the, um, lighting map extension doesn't appear.

1199
02:17:05,090 --> 02:17:08,390
So the machine doesn't know what it's looking at, and you get the black screen of death.

1200
02:17:09,640 --> 02:17:15,549
But we managed to fix that. Um, the next issue that, um, has been pointed out is

1201
02:17:15,550 --> 02:17:22,540
we needed to do mobile optimisation because we had some people in IIIF community trying to use their phones on trains to look at ARCHiOx objects,

1202
02:17:23,080 --> 02:17:29,560
and, um, it became apparent that the um, user interface was too big for,  , tiny, um, iPhones.

1203
02:17:30,070 --> 02:17:37,510
So we got a lot of useful feedback on that, and we actually managed to put in some of it in place, ready for today.

1204
02:17:38,110 --> 02:17:45,310
and these are different shots of mobile views that you can activate in Google Chrome to check it's working.

1205
02:17:45,940 --> 02:17:54,260
See on the left there is something called Specular enhancement mode, which has less sliders available,

1206
02:17:54,290 --> 02:18:01,040
and then in the middle you've got PBR shading and then you've got the tool menu,

1207
02:18:01,040 --> 02:18:05,780
which can be basically put away into a drawer so that you can see more of the screen on a mobile.

1208
02:18:06,350 --> 02:18:11,989
Um, another issue was that people weren't finding the image layers in Mirador, because there's all these.

1209
02:18:11,990 --> 02:18:16,340
There's this stack of wonderful renders that we have available for people to, um, look at.

1210
02:18:16,640 --> 02:18:22,190
So we've now got a button that allows you to cycle through them when you're not looking, when you're not relighting it.

1211
02:18:23,100 --> 02:18:27,120
I have the ability to toggle between the shading techniques,

1212
02:18:27,120 --> 02:18:29,609
as I previously mentioned, we've got specular enhancement mode,

1213
02:18:29,610 --> 02:18:34,019
which is basically allowing the user to remove that colour like it's been mentioned before,

1214
02:18:34,020 --> 02:18:39,690
so that you can focus on the details and bring out the shadows and highlights of an object in real time.

1215
02:18:40,530 --> 02:18:46,800
I made improvements to the way we lit by adding in

1216
02:18:48,220 --> 02:18:52,870
the ability to choose a particular angle that you using a light,

1217
02:18:53,080 --> 02:18:58,960
you can see on the sphere, there are little grid gradations of

1218
02:18:59,290 --> 02:19:07,720
the angle that you can put the lighter so you can go from a really exaggerated, light to almost

1219
02:19:08,080 --> 02:19:17,030
zenithal overhead. And, um, we've even added in a little helper so that people can see where the direction of the light is coming from,

1220
02:19:17,050 --> 02:19:24,090
it's a little blue arrow. Um, and there's also a number of sliders that we've added,

1221
02:19:24,120 --> 02:19:27,239
one of them is normal depth that really makes the details pop out.

1222
02:19:27,240 --> 02:19:36,819
And you can see here. Um, that the people who created these Japanese woodblocks pressed really hard on the paper because this is really shallow,

1223
02:19:36,820 --> 02:19:46,630
but in the viewer, we can we can exaggerate the depth and the, um, when you put raking light over it, you get real detail popping out now.

1224
02:19:47,170 --> 02:19:52,720
To wrap up quickly, I'm just going to go over some of the impact that this projects, spawned,

1225
02:19:53,350 --> 02:20:03,880
this is a shot of a chess set that, one of the convenors of the IIIF 3D community did.

1226
02:20:03,910 --> 02:20:07,660
I got invited to help draft a

1227
02:20:08,600 --> 02:20:13,459
the draft, the new IIIF 3D specification in Washington,

1228
02:20:13,460 --> 02:20:20,450
and that wouldn't have happened if I hadn't become an expert in Mirador and creating these 3D tools.

1229
02:20:20,660 --> 02:20:26,120
So if it wasn't for ARCHiOx, we wouldn't have, um, been invited to be part of that community.

1230
02:20:26,570 --> 02:20:33,290
You've already seen in the last presentation that John and Jorge were given the opportunity to go to the conference in Amsterdam,

1231
02:20:33,290 --> 02:20:36,800
and they received a lot of positive feedback on their work because I was there to witness it.

1232
02:20:37,100 --> 02:20:41,830
I was able to improve OpenSeadragon several times,

1233
02:20:41,890 --> 02:20:47,830
version five of OpenSeadragon now in which is used by thousands of people around the world,

1234
02:20:47,830 --> 02:20:53,380
now includes fixes that were needed for our project, but they benefit the whole community basically.

1235
02:20:53,830 --> 02:20:57,190
So that now supports WebP as an image format,

1236
02:20:57,190 --> 02:21:02,260
and also we fixed rotation and flipping of overlays, which were broken.

1237
02:21:03,740 --> 02:21:09,590
I've also written a static website generator in Python so people can host the photometrics,

1238
02:21:09,590 --> 02:21:13,760
stereo images or RTI images without having a complicated infrastructure.

1239
02:21:14,090 --> 02:21:21,319
and this is the project here. Um, it wasn't working for Jorge and but I'm confident that I've now fix the black screen of death,

1240
02:21:21,320 --> 02:21:26,030
it should work, now. It's Google Chrome fault. Um, and here is our

1241
02:21:26,060 --> 02:21:29,510
ARCHiOx plug in using an alternative. This is

1242
02:21:30,500 --> 02:21:35,299
just basically to point out that, um, the plug in works with anything that produces a normal map.

1243
02:21:35,300 --> 02:21:42,140
So this is just a little RTI I did of a fossil, and it works just the same as a photometric stereo.

1244
02:21:44,100 --> 02:21:50,040
I've gone almost over everything, but I'll now quickly go over what we're going to do next in part three.

1245
02:21:50,490 --> 02:21:53,160
We're going to focus on getting more of these UX changes out

1246
02:21:53,160 --> 02:21:59,550
and we have this huge, um, board that we prepared after user testing, and I'm going to go through it and,

1247
02:22:00,210 --> 02:22:09,360
and follow the advice that we were given and make our tool like the best in terms of web browser tools,

1248
02:22:09,450 --> 02:22:16,230
we're going to focus on maybe publishing our IIIF line extension after involving the community.

1249
02:22:16,710 --> 02:22:24,090
We're going to be we're part of the 2.5 D, group, and we're going to I'm going to ask for that help in deciding,

1250
02:22:24,540 --> 02:22:27,570
some of the vocabulary that we're going to use, and then we're going,

1251
02:22:27,810 --> 02:22:30,960
I think we should try and publish this so that people can,

1252
02:22:31,590 --> 02:22:38,149
because it's one of the requirements to being able to host the stuff yourself is being able to understand the lighting map extension.

1253
02:22:38,150 --> 02:22:39,690
And so, to do that.

1254
02:22:40,170 --> 02:22:46,560
but mainly we're going to focus on getting our pipeline completely smooth and allow us to deliver more ARCHIOx as of course service.

1255
02:22:46,830 --> 02:22:50,489
And you'll see here we have the pleasure of

1256
02:22:50,490 --> 02:22:55,530
we have a dedicated collection of ARCHiOx objects on our Digital Bodleian website that you can go to and on,

1257
02:22:55,530 --> 02:23:00,630
when you click on each object on the right, there is a button that says view in 2.5 D,

1258
02:23:01,020 --> 02:23:08,100
and when you click that, it will take you to Mirador, and you'll be able to interact with these images yourself using our improved plugin.

1259
02:23:08,550 --> 02:23:11,970
And there is a QR code if you want a shortcut to the collection.

1260
02:23:12,420 --> 02:23:14,430
And I'd like to thank you all for listening.

1261
02:23:14,430 --> 02:23:23,580
And thank, thank my colleagues, um, in the Bodleian and our colleagues at Factum and more and most importantly, our kind benefactors

1262
02:23:23,610 --> 02:23:34,819
Helen Hanlyn Trust. Thank you. Wonderful.

1263
02:23:34,820 --> 02:23:37,940
Thank you Richard. Carlos, do you want to come down to the front?

1264
02:23:37,940 --> 02:23:42,950
Do we have another chair? Yes we do, so we have around 20 minutes for questions.

1265
02:23:43,010 --> 02:23:46,880
I'm hoping that you've all been scribbling them down during those talks.

1266
02:23:47,300 --> 02:23:50,540
Do we have, we do have a microphone. Excellent.

1267
02:23:51,020 --> 02:23:54,090
Any questions? Don't be shy. There we go.

1268
02:23:56,680 --> 02:24:00,400
Hello, how are you defining 2.5D?

1269
02:24:02,950 --> 02:24:09,850
So 2.5D I mean, I've referred to all of the recordings as 3D because they they all have a Z axis

1270
02:24:10,030 --> 02:24:16,180
but two and a half D is, is the texture recorded from a static position above

1271
02:24:16,480 --> 02:24:22,930
So with photogrammetry, you recording an original from every conceivable angle, evenly illuminated.

1272
02:24:23,170 --> 02:24:31,420
The opposite is true from   with the Selene, which uses photometric stereo that is a two and a half the recording system.

1273
02:24:32,400 --> 02:24:36,540
Does that make sense? Yeah.

1274
02:24:37,260 --> 02:24:42,059
So 2.5D, for instance, if you had a triangle that you were recording,

1275
02:24:42,060 --> 02:24:46,050
but there was nothing underneath the triangle, given that it's been recorded from above,

1276
02:24:46,410 --> 02:24:52,680
it would assume that that triangle is solid, but it might just be something which was sticking out, which had a void underneath it.

1277
02:24:52,980 --> 02:24:56,760
The camera doesn't know that there's something, whether there's something there or not.

1278
02:24:57,330 --> 02:25:00,690
so that's the really the limitation with, with photometric stereo.

1279
02:25:01,470 --> 02:25:09,480
So you're am I correct in thinking that you, you use the term two and a half deep for just a surface?

1280
02:25:10,050 --> 02:25:12,120
Yeah. Okay. I know that's another definition

1281
02:25:12,130 --> 02:25:20,190
perhaps it can also be understood as any three dimensional object which has three axis of direction x, y,

1282
02:25:20,190 --> 02:25:29,550
z, in which one of the axes is much, much smaller than the other two, meaning flat shape with some texture.

1283
02:25:30,120 --> 02:25:35,900
Yeah. And in our plugin, the geometry is just a flat plane, completely flat,

1284
02:25:36,470 --> 02:25:44,150
it's a bunch of tiles linked together and, um, the only thing I'm feeding in 2D images

1285
02:25:44,150 --> 02:25:50,000
and the normal map, is basically telling the shader to where the light direction is coming from,

1286
02:25:50,000 --> 02:25:55,890
so it gives the illusion of 3D. So it's not I'm using a 3D technology, but it's really 2D

1287
02:25:55,910 --> 02:26:02,520
so that's why I prefer the term 2.5D for this. This, for instance, is a 2.5D recording

1288
02:26:02,520 --> 02:26:10,799
so this was recorded with the Selene. Um, the skeleton of this um, recording is just simply a mesh of, of squares,

1289
02:26:10,800 --> 02:26:20,910
in fact, and those squares have been displaced with a texture map, um, something called a height map or depth map, and

1290
02:26:23,570 --> 02:26:30,500
right, there we go, so as you can see, the sides have not been recorded and we can't turn the seal over.

1291
02:26:30,520 --> 02:26:38,960
So this is a this is a 2.5 D recording, whereas this Magna Carta seal we could see that we could turn it completely over.

1292
02:26:39,020 --> 02:26:42,130
We've recorded the sides. So this is a photogrammetry recording.

1293
02:26:42,610 --> 02:26:49,360
and as I said, there are advantages and benefits to both those, um,  , both those procedures.

1294
02:26:49,870 --> 02:26:53,349
This takes a very, very long time to, to record.

1295
02:26:53,350 --> 02:26:59,320
So this, this probably took a day to, to make this recording of the, of the, of the seal using photogrammetry.

1296
02:26:59,770 --> 02:27:05,800
The previous seal probably took two seconds to, to record and 40s to process.

1297
02:27:06,790 --> 02:27:19,450
That is so incredibly efficient. What do you think of the potential consumer uses of this system outside of research?

1298
02:27:24,330 --> 02:27:30,520
You can teach. You can use them. I'll plug in works in a browser and it's delivered as software, as a service.

1299
02:27:31,210 --> 02:27:38,170
and it's all open source as well. So, the front end is open source, so anyone can deploy it for a collection that they have.

1300
02:27:38,620 --> 02:27:45,820
and it, it just means it's a nice way of,getting students and things to engage with it.

1301
02:27:46,180 --> 02:27:52,540
given that the viewers that we're using also support technology such as annotating and things like that.

1302
02:27:52,870 --> 02:27:59,199
you could set up a, um, you could set up a system where

1303
02:27:59,200 --> 02:28:03,370
students could make their own annotations and keep them, keep their own notes on objects.

1304
02:28:03,400 --> 02:28:08,690
so I think... And how small are you planning on going with the images?

1305
02:28:08,760 --> 02:28:12,640
Do you recon you could get microscopic with the software or even cellular level.

1306
02:28:12,850 --> 02:28:17,370
Already it already does deliver so millimetre details.

1307
02:28:17,380 --> 02:28:20,280
You're kind of only limited by the lens which you put into the system.

1308
02:28:20,290 --> 02:28:24,580
So typically we use a 50 millimetre lens and that gives you a million pixels per square inch.

1309
02:28:25,000 --> 02:28:29,590
If you tried that for 100 um, millimetre lens, you get 4 million pixels per square inch.

1310
02:28:30,130 --> 02:28:35,920
So I think it's something that I'd like to explore putting on really, really high magnification lenses.

1311
02:28:35,920 --> 02:28:42,850
And I'm sure it'll work for far too I, I've used a dioptre on a camera and I got 30 micron details.

1312
02:28:42,850 --> 02:28:46,810
So it's it's really you can get really high high magnification

1313
02:28:46,900 --> 02:28:49,420
well, relatively high magnification just using a camera.

1314
02:28:49,990 --> 02:28:55,840
So, but the principle I reckon you could use it on a microscope if you had four tiny lights.

1315
02:28:58,480 --> 02:29:02,510
So challenge, I think that the limiting factor might be depth of field.

1316
02:29:02,530 --> 02:29:07,210
And you do need good sharpness,so yes, it must be sharp.

1317
02:29:11,730 --> 02:29:18,330
question for anyone who wants to, you talking about dealing with highly reflective surfaces such as metal?

1318
02:29:18,510 --> 02:29:21,990
How far do you think you could push that? Could you do glass or jewellery?

1319
02:29:25,490 --> 02:29:31,280
I mean, there's some limitation with, translucent or transparent objects at the end.

1320
02:29:31,910 --> 02:29:39,090
We are talking about optical systems that depend on getting up

1321
02:29:40,010 --> 02:29:46,790
something that can be properly measured once you have images that are yeah, fine in detail.

1322
02:29:46,790 --> 02:29:50,600
So the moment you are, you have confusing information.

1323
02:29:51,080 --> 02:29:55,400
It's not possible to compute the 4 images into creating relief data.

1324
02:29:56,630 --> 02:30:08,930
said that there's ways of, you know, in a way helping to avoid the negative effects of highly varnished surface, for example.

1325
02:30:10,040 --> 02:30:16,759
Or, Saddam was showing this morning, perhaps combining Selene with other technologies like Gaussian splat.

1326
02:30:16,760 --> 02:30:22,220
It's possible to, even if you are not recording 3D, but effectively being able to visualise

1327
02:30:23,150 --> 02:30:28,820
3D. So yeah, again, it's about pushing the limits of the technology

1328
02:30:29,180 --> 02:30:40,040
one step at a time. You might recall the, um, small portrait which was, um, under glass in the first session.

1329
02:30:40,430 --> 02:30:45,190
So that was highly reflective, like recording through that glass is is possible because the,

1330
02:30:45,190 --> 02:30:50,270
the lights of the Selene are in position such that they're not really reflecting onto the surface.

1331
02:30:50,450 --> 02:30:54,170
So really the central portion of each recording is the most accurate part.

1332
02:30:54,490 --> 02:30:59,690
And especially if you're making a multi tiled image, you're combining the best part of each image.

1333
02:31:00,050 --> 02:31:02,690
so actually reflections is not something that I struggle with.

1334
02:31:02,690 --> 02:31:10,400
And a lot of our copper plates, I mean, some examples that I've showed you are very dull and highly corroded, but others are extremely reflective.

1335
02:31:10,700 --> 02:31:14,150
And actually Selene is very, very good at recording these surfaces.

1336
02:31:16,220 --> 02:31:20,390
do you have any plans of incorporating that spectral domain?

1337
02:31:20,420 --> 02:31:23,630
you mentioned pushing through,the technology.

1338
02:31:24,770 --> 02:31:27,770
you can incorporate multispectral imaging, for example.

1339
02:31:27,770 --> 02:31:33,780
Is that a plan that you have? I mean, you can layer multi-band images with the Selene recordings.

1340
02:31:33,780 --> 02:31:39,840
So that's something that I showed in the in the first section with a I don't know, Adam.

1341
02:31:40,320 --> 02:31:46,950
Is that something that you've considered thinking to the question? So, so it's combining multi-band imaging into the Selene.

1342
02:31:50,910 --> 02:32:03,260
I mean. I think to get the best results, probably it's best to use a very high end multi-band system and the 2.5 to the Selene,

1343
02:32:03,740 --> 02:32:12,080
because I think it's better to to get really good quality recording systems rather than trying to make one machine to do everything.

1344
02:32:12,080 --> 02:32:21,050
not as well, if you see what I mean. I was very interested to see your mobile unit

1345
02:32:21,770 --> 02:32:23,240
because obviously,

1346
02:32:23,260 --> 02:32:32,000
the full scale Selene thing is a thing that's static in a few museums or galleries that are lucky enough to have one or the means to buy one.

1347
02:32:32,360 --> 02:32:38,269
but there's huge potential for this type of technology to go out to

1348
02:32:38,270 --> 02:32:44,180
other places who might need it for an object or two objects or, you know, for three weeks or something of that sort.

1349
02:32:44,540 --> 02:32:54,410
so the machine is one thing, but the expert to take the images and process the images is another aspect of that.

1350
02:32:54,740 --> 02:33:02,299
I was wondering if you could say a little bit more about how, whether it's through ARCHiOx or through the Factum Foundation,

1351
02:33:02,300 --> 02:33:08,420
or how really you're going to be able to, operationalise this for the broader,

1352
02:33:08,450 --> 02:33:16,400
heritage community, which may not have the means or the need for one of these

1353
02:33:16,700 --> 02:33:24,710
piece of kit all the time, but could, generate immense value from it for particular projects.

1354
02:33:25,760 --> 02:33:32,840
Thank you, so I would say the goal of all this project is about making sure that

1355
02:33:34,660 --> 02:33:39,610
you want to know about making sure that it's possible to record as many original objects as possible,

1356
02:33:39,610 --> 02:33:45,280
as many collections as possible, because that's essential for the conservation, the long term preservation of the objects.

1357
02:33:45,760 --> 02:33:51,700
So it's part of the work of the Factum Foundation to provide the technology and the know

1358
02:33:51,700 --> 02:34:00,400
how, about how to do that. And I think the focus in this case is about making it very specialised to the needs of the institution.

1359
02:34:00,730 --> 02:34:05,890
In some cases, they will need a full working system like what they have here at the Bodleian.

1360
02:34:06,190 --> 02:34:14,800
In other cases, it's about one particular object or one specific type of object that perhaps doesn't need the complete system.

1361
02:34:15,070 --> 02:34:25,570
So we could train the, technicians into only one aspect out of the complexity of this project so that they don't really need to understand.

1362
02:34:26,290 --> 02:34:30,190
I mean, there's so many variables, both in the capturing and in the processing,

1363
02:34:31,000 --> 02:34:36,040
but sometimes it's just about focusing on one specific, set of skills.

1364
02:34:36,340 --> 02:34:44,750
And yes, answering to your question, the idea is, any possible way of making it as accessible as possible.

1365
02:34:46,020 --> 02:34:46,120
Jo

1366
02:34:46,210 --> 02:34:57,940
I mean, I think in what we're doing with the archive in Venice is running the Archive Online Academy, where, um, there's learning by doing courses.

1367
02:34:59,860 --> 02:35:09,460
So that both professionals within institutions and interested third parties can understand how to do the work.

1368
02:35:09,830 --> 02:35:15,840
then the development of the portable Selene means it can go very easily.

1369
02:35:15,850 --> 02:35:18,430
I mean, it's tripod mounted. It fits in a box.

1370
02:35:18,850 --> 02:35:26,740
So, what we're proposing really to put into Saint Catherine's Monastery is effectively a slightly glorified,

1371
02:35:27,070 --> 02:35:34,150
um, portable system because it needs to tie in with what, Justin's doing with the multi-spectral recording.

1372
02:35:34,570 --> 02:35:39,880
So I think the idea is, rather than just thinking there's this system,

1373
02:35:40,830 --> 02:35:49,740
desk and studio, if you're thinking it's really a camera with a different lighting system and some software,

1374
02:35:50,040 --> 02:35:54,620
and I mean, from the event to years ago, what I think,

1375
02:35:54,630 --> 02:36:00,000
Richard was just talking about what's been clear from John and, and Carlos's

1376
02:36:00,030 --> 02:36:06,300
presentations is it's becoming easier and easier as the software solves the problems.

1377
02:36:06,450 --> 02:36:09,719
I mean, you still have to know how to take good photographs.

1378
02:36:09,720 --> 02:36:12,990
You have to understand it. But the principle of it... has to be

1379
02:36:15,730 --> 02:36:22,690
I mean, if you have a single object, does it make, you know, how do you get that piece of kit to and the person who knows how to do it?

1380
02:36:23,200 --> 02:36:32,680
Well, I mean, one way is you get on an airoplane and you take the scanner, I mean, so a scanner's going to Naples today working on a special project.

1381
02:36:32,980 --> 02:36:34,959
I mean, what

1382
02:36:34,960 --> 02:36:44,440
Helen Hamlyn's Trust enables us to do by having two scanners that could be lent out is to do single use or relatively focussed projects.

1383
02:36:44,740 --> 02:36:49,479
But but basically, it's growing very fast. So it's becoming easier.

1384
02:36:49,480 --> 02:36:54,240
And the systems, I mean, the systems cost money, but they're not horribly expensive.

1385
02:36:54,280 --> 02:37:02,860
I mean, you know, when you've got institutions who spend, you know, a large amount on buying a phase one medium format camera

1386
02:37:03,430 --> 02:37:09,160
this is a 35 millimetre with four lights with very good software.

1387
02:37:09,430 --> 02:37:15,490
So it's becoming affordable in different ways and producing very, very good results.

1388
02:37:15,820 --> 02:37:26,050
I mean, comparative comparable to I mean, is that Selene circle something else that could enable as a sharing system to.

1389
02:37:27,270 --> 02:37:36,340
The Selene circle is a sharing system, a community of practice amongst all the people who are working with the equipment and that spreading.

1390
02:37:36,360 --> 02:37:40,379
I mean, the more people who have them, then hopefully there'll be one near you.

1391
02:37:40,380 --> 02:37:49,330
Jo, I think that's part of our vision with ARCHiOx

1392
02:37:49,780 --> 02:37:53,040
John's already been out and taken the mobile Selene

1393
02:37:53,050 --> 02:37:58,900
yeah, so, yes, that is part of our vision.

1394
02:37:58,900 --> 02:38:04,330
But we need to make sure that we get it right. And, um, and, you know...

1395
02:38:06,160 --> 02:38:11,530
well, no, we just need to make sure that we've got enough people within the Studio downstairs,

1396
02:38:11,530 --> 02:38:18,580
which is also part of our vision to to get that training and become part of integrated into how we work here.

1397
02:38:20,170 --> 02:38:29,380
And the community of practice will just help that. And I think in the longer term, it would be nice for the community of practice to to have

1398
02:38:30,590 --> 02:38:34,580
software people excuse my lack of technical knowledge, everybody.

1399
02:38:34,910 --> 02:38:42,020
photographers and other people, not necessarily Selene users, but people who are really,

1400
02:38:42,020 --> 02:38:46,349
really interested in what we're doing so that we can just share that knowledge.

1401
02:38:46,350 --> 02:38:55,670
So I think it's there is a lot to work forward to, but we've come a long way in the last nearly three years from um,

1402
02:38:55,910 --> 02:39:04,460
Factum and Jorge and Adam and idea all the work they'd done leading up to that because there's three years, 3 or 4 years before that.

1403
02:39:05,080 --> 02:39:09,800
And we're getting there, but we've got a long way to go and a lot more to do.

1404
02:39:13,490 --> 02:39:17,930
Will this developing technology speed up the

1405
02:39:17,930 --> 02:39:23,960
repatriation of artefacts back to their original locations?

1406
02:39:26,440 --> 02:39:30,930
I think that might be above the pay grade of everyone in the room. Adam, do you want to take that one?

1407
02:39:32,850 --> 02:39:37,860
I think probably it's always good to answer a question like that with an example.

1408
02:39:38,490 --> 02:39:47,180
So, Factum has been working for the last 7 or 8 years in Nigeria, in southeast Nigeria,

1409
02:39:47,190 --> 02:39:53,310
on the recording of the Bakor monoliths and the end last year,

1410
02:39:53,680 --> 02:39:59,759
beginning of last year, there was an exhibition at the British Museum of the Bakor

1411
02:39:59,760 --> 02:40:04,050
a Bakor monolith that's in the collection of the British Museum.

1412
02:40:04,710 --> 02:40:13,230
And Factum put out a rather, I suppose, cheeky call saying that any museum that had a Bakor,

1413
02:40:13,230 --> 02:40:17,820
monolith that thought it shouldn't have if 

1414
02:40:17,820 --> 02:40:24,870
they gave us permission to make an exact facsimile, which they kept and returned the original object to the,

1415
02:40:25,320 --> 02:40:30,210
local site Factum Foundation would fund the project.

1416
02:40:30,720 --> 02:40:38,280
And slightly to our surprise the Chrysler Museum in North Carolina

1417
02:40:38,280 --> 02:40:43,980
approached us and said they had a Bakor monolith that they'd been donated that they felt they shouldn't have.

1418
02:40:44,630 --> 02:40:48,270
we went in and recorded it. We made a facsimile.

1419
02:40:48,780 --> 02:40:54,089
the facsimile is now on show in the Chrysler Museum

1420
02:40:54,090 --> 02:40:58,830
as part of a display focussed on the problems of preserving cultural heritage

1421
02:40:59,040 --> 02:41:02,250
in parts of the world where there isn't really a culture of preservation.

1422
02:41:03,000 --> 02:41:11,490
the original monolith was handed over to the Nigerian ambassador in Washington, and it's now made its way back.

1423
02:41:11,760 --> 02:41:19,470
It's not quite where it needs to be. It's about 50 miles away, but it's in Calabar and it's on show in the museum in Calabar.

1424
02:41:19,830 --> 02:41:26,340
While the site in the village is being prepared for it to be looked after.

1425
02:41:26,610 --> 02:41:32,970
So I think there are examples where I think

1426
02:41:33,000 --> 02:41:42,389
attitudes towards ownership are shifting and I think there's a lot more to be done in terms of how you can share objects.

1427
02:41:42,390 --> 02:41:48,719
But really, for me, the excitement that I tried to show this morning is once you've got a digital recording,

1428
02:41:48,720 --> 02:41:56,620
that's not just an image, not just a photograph extracted and put online, but is an object that you can handle

1429
02:41:56,880 --> 02:42:07,110
that you could look at that you couldn't in a virtual space engage with, then I think you're changing the whole relationship between,

1430
02:42:07,530 --> 02:42:13,260
objects, how they communicate, what they communicate, and what people can get from them.

1431
02:42:13,440 --> 02:42:22,739
And I think the whole question of the relationship between originality and authenticity is really opening up a lot.

1432
02:42:22,740 --> 02:42:31,230
And I'm sure the debate will run for a long time. Um, but hopefully in some cases things can migrate to where

1433
02:42:32,410 --> 02:42:36,850
they're more purposeful or where they can fulfil a greater need.

1434
02:42:37,690 --> 02:42:41,680
So I think we're getting past a minor yours battle.

1435
02:42:43,040 --> 02:42:51,760
I hope, I think that's taken us to the end of our Q&A time.

1436
02:42:51,790 --> 02:42:55,780
so I'm going to invite Adam up to wrap up for us.

1437
02:42:56,230 --> 02:43:04,300
before the all important drinks reception in the Blackwell Hall, but before Adam, wraps up for us, just join me in thanking, Carlos and John and Richard.

1438
02:43:15,470 --> 02:43:22,629
I think you've all had enough of me for today, but I mean, I actually feel quite emotional.

1439
02:43:22,630 --> 02:43:27,740
I feel the, the level of conversation and focus has been really extraordinary.

1440
02:43:27,760 --> 02:43:40,030
So from 1:00 till now, there seems to have been a really intense, concentration on the materiality of things in museum library collections.

1441
02:43:41,410 --> 02:43:49,600
You know, I loved seeing the Raphael drawings being kind of unpacked, unpacked, looked at as physical objects.

1442
02:43:49,960 --> 02:43:56,160
And I had the great privilege when I was a student at the Ruskin, to have

1443
02:43:56,210 --> 02:44:02,020
a studio which was called the Ruskin Masters Drawing Studio, which was within the Ashmolean.

1444
02:44:02,410 --> 02:44:04,569
And so a big part

1445
02:44:04,570 --> 02:44:14,580
of my education as an artist was actually going to the prints and drawings and handling many of these things and thinking drawings,

1446
02:44:14,590 --> 02:44:21,040
they're not images. They're not just images. Drawings are objects absolutely packed with information.

1447
02:44:21,400 --> 02:44:30,490
And one of the things that we didn't see is we also recorded one of the extraordinary Michelangelo, um, late crucifixion drawings.

1448
02:44:31,090 --> 02:44:37,660
There, you could start to almost see how the laid lines in the paper were influencing the drawing.

1449
02:44:38,840 --> 02:44:42,090
How the the watermark suddenly becomes a part of it.

1450
02:44:42,090 --> 02:44:45,899
So the drawings immediately become physical objects.

1451
02:44:45,900 --> 02:44:50,400
And I think that's been visible all day today.

1452
02:44:51,030 --> 02:44:56,010
And in the sort of three years that ARCHiOx has been running

1453
02:44:56,610 --> 02:44:59,830
I think we've moved a very long way.

1454
02:44:59,880 --> 02:45:02,730
I mean, I think what you saw in Carlos,

1455
02:45:02,730 --> 02:45:09,780
talk from the software development that's going on in Richard's talk the whole way that you put it online.

1456
02:45:10,200 --> 02:45:17,639
And for me, getting this true three dimensional handling of an object in a,

1457
02:45:17,640 --> 02:45:23,550
in a virtual space, in an online space is something that's going to be a game changer.

1458
02:45:23,940 --> 02:45:32,849
And I'm very, very happy that in a way, it's been the collaboration here at the Bodleian between, you know, the studio in Madrid.

1459
02:45:32,850 --> 02:45:38,430
And it should be said that, you know, Factum Foundation, in a way, is is an idea.

1460
02:45:38,610 --> 02:45:43,860
I mean, Factum is a group of people who are focussed

1461
02:45:44,020 --> 02:45:49,799
it's almost like a playground where everyone's focussed on trying to work as a team,

1462
02:45:49,800 --> 02:45:56,160
trying to be cross-disciplinary and stand outside of normal professional positions

1463
02:45:56,610 --> 02:46:02,130
and focus on shared problems and find solutions that have a real application.

1464
02:46:02,610 --> 02:46:09,749
So, I mean, what's become very clear is that more and more people are wanting to visit Factum in Madrid,

1465
02:46:09,750 --> 02:46:18,120
which is a bit of a nightmare because actually it's a chaotic, highly functioning, hard working studio and more and more people want to come and talk.

1466
02:46:18,450 --> 02:46:23,069
But something's happening. And I think the collaboration here and, you know,

1467
02:46:23,070 --> 02:46:28,350
it's it's very true that Elaine and John and the team in ARCHiOx has

1468
02:46:28,350 --> 02:46:33,900
really worked as a collaboration because they've responded in amazing ways.

1469
02:46:34,140 --> 02:46:37,740
Um, and you know what, what Richard was talking about these weekly meetings with

1470
02:46:37,740 --> 02:46:41,880
Jorge, so I should say it's very sad that Jorge is not here today.

1471
02:46:42,330 --> 02:46:48,270
Um, he's not been so well recently. And I hope he will be back to full health very soon.

1472
02:46:48,690 --> 02:46:54,209
Um, but, um, you know, he's missed. I mean, he's one of a team

1473
02:46:54,210 --> 02:47:03,750
probably of about 8 or 10 people who over 7 or 8 years have been focussed on trying to get this photometric system to work.

1474
02:47:03,750 --> 02:47:09,989
So in a way, Jo to answer to your question, I mean, the portable Selene is exactly what you ask about.

1475
02:47:09,990 --> 02:47:14,070
I mean, we need to be someone needs to be able to say, can you bring it here and do this?

1476
02:47:14,400 --> 02:47:20,850
We can say, yes, it's going to cost you X to have it in there for one day, for two days with an operator.

1477
02:47:21,390 --> 02:47:24,480
And, you know, this is not prohibitively expensive.

1478
02:47:25,050 --> 02:47:33,990
obviously, the more the systems get spread around the world, hopefully the different institutions will be able to detach.

1479
02:47:33,990 --> 02:47:40,980
I mean, like John does a lot on his detach the head and put it on a tripod and use it as a portable system.

1480
02:47:41,250 --> 02:47:44,250
So, um, it's actually functioning in different ways.

1481
02:47:44,880 --> 02:47:56,730
I on a personal level, I'm incredibly happy to see, um, the focus not just on cultural objects, but also on natural objects.

1482
02:47:56,730 --> 02:48:04,320
So I think where the recording of fossils, um, you know, I mean, John knows I get very overexcited.

1483
02:48:04,680 --> 02:48:11,250
But when I saw these images coming through of what you could see on the fossil with and without colour,

1484
02:48:11,970 --> 02:48:18,299
and how suddenly the whole detail of the wasp's legs and everything else comes up

1485
02:48:18,300 --> 02:48:27,840
and now we're working in Madrid with the Museo Nacional de Ciencias Naturales and with Natural History Museum here

1486
02:48:27,840 --> 02:48:33,600
I hope we have, I mean Factum put in the Diplodocus, with Keith who is here too

1487
02:48:33,600 --> 02:48:41,940
and, the big bronze one, and I really hope we can use some of these technologies to show when you get

1488
02:48:41,940 --> 02:48:48,300
groups of people who are really communicating what can be achieved and how.

1489
02:48:49,770 --> 02:48:54,450
Culture and nature technology suddenly build a polytechnic world

1490
02:48:54,600 --> 02:49:03,180
that's incredibly exciting. And I hope in a way, the things that I feel in the room today is there's a real interest.

1491
02:49:03,810 --> 02:49:08,970
And this isn't just a dry conversation about some academic footnote

1492
02:49:09,630 --> 02:49:16,920
this feels much more like a revolution in thinking of objects as material evidence.

1493
02:49:17,310 --> 02:49:21,330
And so the question this morning about forensics, you know,

1494
02:49:21,330 --> 02:49:30,180
can we do this in more and more forensically accurate ways is obviously something we're all trying to do.

1495
02:49:30,690 --> 02:49:39,300
And can we go smaller and smaller? And the insect I showed you is about a centimetre, maybe one and a half

1496
02:49:39,870 --> 02:49:46,470
but we've got a whole case of mosquitoes from the Natural History Museum in Madrid.

1497
02:49:46,980 --> 02:49:52,770
And we're desperately trying to do those with magnifying lenses

1498
02:49:53,340 --> 02:50:03,329
and it would be wonderful to go microscopic. So I think for the next phase, which I think I'm allowed to, to say now is has again,

1499
02:50:03,330 --> 02:50:07,800
very generously been supported by the Helen Hamlyn Trust

1500
02:50:08,160 --> 02:50:17,550
we're moving into investigations, both in the software, both in the software to access the data

1501
02:50:17,670 --> 02:50:26,070
and in different hardware developments, which will hopefully result in very fast systems for recording insects.

1502
02:50:26,730 --> 02:50:29,850
I mean, many, as John pointed out a couple of times,

1503
02:50:30,300 --> 02:50:40,470
the photogrammetry can be quite time consuming both in the recording and the processing, whereas the Selene has got it down to a few seconds.

1504
02:50:41,220 --> 02:50:46,580
But if we can get the recording of insects, tiny insects down

1505
02:50:48,750 --> 02:50:52,830
probably 20 minutes an insect and there's 2 billion of them,

1506
02:50:53,490 --> 02:51:02,000
then you just have to do the sums; how many systems can each system, each operator, operate 3 or 4 systems.

1507
02:51:02,000 --> 02:51:06,640
So they put an insect in place, press the button, go to the next one, put it in place

1508
02:51:06,660 --> 02:51:10,770
but you get multiple recordings at high resolution.

1509
02:51:10,770 --> 02:51:15,040
And then of course, it's all about machine learning.

1510
02:51:15,370 --> 02:51:23,680
it's all about how you analyse the data, what you can do with it, how you process it, how you cross-linked it,

1511
02:51:23,920 --> 02:51:37,030
how you basically find connections that make that data mean different things in different contexts, how new meanings get generated by doing this.

1512
02:51:37,820 --> 02:51:41,560
So I mean, really, thank you all very much for coming.

1513
02:51:42,070 --> 02:51:47,110
thank you all for the feeling that I can pick up in this room.

1514
02:51:47,530 --> 02:51:55,810
And I just like, I think if I got a; I'd just like to say a big thank you to the Helen Hamlyn Trust.