Ambrosiana_C73inf (Latin Moses and Latin Commentary on Luke) Processing Guide


  1. Focused attention on captures from both sides of folio aligned (LINK)
  2. Ordered in sequence of folio number (LINK)
  3. Focused attention on lines unread in critical edition, Ceriani 1861 (LINK)

See also:

Focused attention on captures from both sides of folio aligned


10/17/2017 TRH
I identified a good folio to focus efforts on after 112-111.
124-123 has everything we want: some parts Ceriani couldn’t read, others he claimed to read but we would like to check, the Ruby does not work very well, it is an important passage in which even a few letters could be important. Also, there seems to be more in the margin than just the modern penciled line numbers and scribal doodle.
It’s all important and hard to read, but one particular region Ceriani could not read is x,y,w,h=3064,4303,2004,484.
illegible line

10/20/2017 RLE
First tries on 124 with registered reversed 123
so far, just used transmissive + transmissive and reflective + reflective, nothing with fluorescence yet.
registration looks "pretty good" -- easy to find the holes along the RHS of 124, much more difficult along the LHS
uploaded a few images to loading dock, RLE, folder Amb_C73inf_124+123
"least worst" in my eye is Amb_C73inf_123_flipH_warp_124_cal_b01-b14_28bands_RF_med3_yellow_underink_ICA_R-17_G+23_B+27_imga.tif
"second least worst" is Amb_C73inf_123_flipH_warp_124_cal_b39-b42_TX_8bands_med3+box71_bad_yellow_underink_ICA_R-17_G+23_B+27_iml2.tif

10/23/2017 TRH
I agree that the best image is Amb_C73inf_flipH_warp_124_cal_b01-b14_28bands_RF_med3_yellow_underink_ICA_R-17_G+23_B+27_imga.tif. It is noticeably better than Accurate Color for the lines I tried to read, but still short of arguing with Ceriani.
I'm also seeing things in Amb_C73inf_123_flipH_warp124_cal_b01-b14_28bands_RF_med3+box71_bad_blue_textblock_PCA_R+6_G-8_B-11_imga.tif. My guess is that it is from the other side.

10/31/2017 RLE

  1. Amb_C73inf_123_flipH_warp_124_cal_365_NBGORBP_14bands_FL_yellow_mask_med3+box51_bad_magenta_IC6.tif
    This is one of the "cropped" files I made last night. Some text (not a lot) is showing up near the left and bottom center (as presented ... perhaps upside down?).
  2. Amb_C73inf_123_flipH_warp_124_cal_365_NBGORBP_14bands_FL_yellow_mask_med3+box51_bad_magenta_PCA5.tif
    (again, the cropped version, but I can figure out how to implement on the full image if this is helpful)
    The undertext appears to me to show up in spots as an "embossing" effect due to the misregistration of the fluorescence bands. Guessing it likely that this only helps where you can see the text otherwise, but I'm not sure.
  3. Amb_C73inf_123_flipH_warp_124_cal_365_NBGORBP_14bands_FL_yellow_mask_med3+box51_bad_magenta_PCA_R+5_G-8_B-10_imga.tif
    perhaps the "best" so far ...

11/1/2017 TRH
I can't get oriented. Normally I rely on the ability to flicker between accurate color and processed color images to find where I am and identify the layer I am trying to read. How hard would be it be to either avoid cropping or refit the images to a canvas of the original pixel dimensions?
On a related note, do we need to worry about the baseline for Keith's registration correction? If the "normative" image to which other images are corrected does not align with the hemisphere captures we will have some strange effects in the Spectral RTI processing.

Ordered in sequence of folio number

Ambrosiana_C73inf_001 Navigate

8/3/2017, nine images uploaded by RLE and interns:
Amb_C73inf_001_SAMoverlay331.tif (RLE01)

8/22/2017 TRH: SAMoverlay331.tif is nice because the overtext and undertext appear in slightly different colors. Note that on this page and others like it there is some erased red ink (rubric) in addition to the erased black ink. It is less legible in the natural and enhanced images. For example, x,y,w,h=1350,970,2050,230 should read "et relictis omnibus."
Ac Accurate Color
Ac RLE01

Ambrosiana_C73inf_002 Navigate

8/3/2017, five images uploaded by RLE and interns:
Amb_C73_002_SAM_overlay331withinteractivestretchig.tif (RLE01)

8/22/2017 TRH: The best overall is Amb_C73_002_SAM_overlay331withinteractivestretchig.tif. Also useful, but not as consistently, is Amb_C73inf_002_SAM_background_2.tif.

The erased rubric (red ink) is especially tough. At x,y,w,h=775,6134,1934,260 we expect to read "sequere me."
Ac Accurate Color

Ambrosiana_C73inf_003 Navigate

8/3/2017, nine images uploaded by RLE and interns:
Amb_C73inf_003_PCA_SAM_band-1-1undertext.tif (RLE02)
Amb_C73inf_003_SAM_background_2.tif (RLE01)

8/22/2017 TRH: Amb_C73inf_003_SAM_background_2.tif has nice separation between overtext and undertext. There are some interesting markings visible on Amb_C73inf_003_PCA_SAM_band-1-1undertext.tif, but I believe it is show-through. The unreadable rubric on this page is at x,y,w,h=1519,5322,1358,312. We would expect it to read, "et sedens docebat de nauicula turbas."
Ac Accurate Color

Ambrosiana_C73inf_004 Navigate

8/3/2017, six images uploaded by RLE and interns:
Amb_C73inf_004_PCA_SAM_undertextovertextband1.tif (RLE01)

8/22/2017 TRH: Amb_C73inf_004_PCA_SAM_undertextovertextband1.tif is the best of these. No luck on the rubric. We expect to read "Simonis" right before "illis" at x,y,w,h=946,3213,1390,164. We can't blame overtext obfuscation on this one. Also the entire line above it probably reads "dens autem in unam nauem quae erat."
Ac Accurate Color

Ambrosiana_C73inf_005 Navigate

8/11/2017, ten images uploaded by RLE and interns:
PCA/Color/Amb_C73inf_PCA_SAM_band1_undertextovertext.jpg (RLE01)

8/22/2017 TRH: The PCA is similar but just a little clearer. Could you upload an uncompressed TIFF to maintain our quality standards?

Ambrosiana_C73inf_006 Navigate

8/11/2017, eight images uploaded by RLE and interns:
PCA/Color/Amb_C73inf_006_PCA_SAM_band1undertextovertext.jpg (RLE01)

8/22/2017 TRH: The color image is nice, but can I have the uncompressed tiff?
The region x,y,w,h=690,4370,1541,321 is interesting in two ways. First, there is an illegible rubric, probably "duc in altum." Second, there is an example of an interlinear addition, "ubi copiosae."
Ac Accurate Color

Ambrosiana_C73inf_007 Navigate

8/11/2017, fourteen images uploaded by RLE and interns:
ICA/Color/Amb_C73inf_007_ICA_PCA_SAM_undertextPCAband1ICAband2.jpg (RLE01)

8/22/2017, TRH: Amb_C73inf_007_ICA_PCA_SAM_undertextPCAband1ICAband2.jpg looks very nice. Can we have it as an uncompressed tiff for the archive?
If you're up for an extra challenge, there is an area we can't read with room for about eight letters at x,y,w,h=3636,6195,1200,266.
Ac Accurate Color

Ambrosiana_C73inf_008 Navigate

Ambrosiana_C73inf_009 Navigate

Ambrosiana_C73inf_010 Navigate

Text in Latin Moses Unread by Ceriani (1861)

Ordered in sequence of decreasing number of lines unread on a given page

Page 113: col a, ll 4-10; col b, ll 1-16

Navigator Link:

column a
1) beatus sum si adpro
2) ximauerunt dies mei
3) in diebus uitae tuae
4) ---
5) ---
6) in me --- simi
7) l --- est lo
8) cuta es me cum de
9) morte tua. Et intro
10) iuit --- ad isac

column b
1) ---m meo
2) --- fratrem
3) suum ---
4) omnia quaecumque
5) fecit ---
6) die
7) ---
8) ---  u --- in diem
9) hunc quo ---
10) ---
11) ---
12) ---
13) ---
14) rapuit ---
15) petentes eramus ---
16) eo de nostris ---

Initial round of advanced processing


TRH notes 3/2/2017

Mostly able to read much less than Ceriani, except one region of optimism


4 lines unread by Ceriani, I can make out several letters, but not so many that improvement wouldn’t be appreciated.

Page 68: col a, ll 1-13; col b, ll 7-15

column a
1) s ... n ... ca ..…
2) .... nus diis .…
3) .... omnis ....…
4) ........…
5) ..... u .... o ....
6) rue ..... elen
7) tes ..... rum bo
8) norum comesto
9) res dicentes se haec
10) facere propter mi
11) sericordiam qu ...
12) se et extermina
13) tores quaeru ...
14) fallaces celantes se

column b
1) tamquam principes
2) erimus et manus
3) eorum et mentes
4) inmunda tractantes
5) et os eorum loque
6) tur ingentia et su
7) per dicent noli ....
8) tange ne inquines
9) me loco in quo ... s .…
10) .. is d ........
11) su ..... us .....
12) in .........
13) re .... raui ....
14) in plebem quae s .. a ..
15) illis et .. ta .. ue
16) niet in eos ultio et

Initial round of advanced processing


TRH notes 3/2/2017

Mostly frustration here but one letter that looks very clear AND contradicts Ceriani. He reads an N where I see an S rather clearly. If it’s not the other side we have an example of disproving Ceriani. Or if it’s a G he asserts more letters than could possibly fit.

The coordinates are: x,y,w,h = 3153,2824,170,205

Page 93: col a, ll 1-11; col b, ll 11-12

column a
1) mulieri mercedem
2) l--- ar ---
3)  ---
4) ---
5) et docuit ---
6) et quando comple ---
7) --- septima
8) na --- erunt te
9) in atrium regale et eras
10) ex --- trium septi
11) manarum annorum

column b
9) me uis quemadmodum
10) occidisti hesterna die
11) aegyptium. --- u ---
12) tia propter --- o --- huius ---
13) et quadragesimo no

Initial round of advanced processing


TRH notes 3/2/2017

This page is really badly damaged at the parchment level… let’s not spend more time here

Page 77: col b, ll 3-11

column b
1) ex quo facto finien
2) tur tempora momen
3) to .... etur cursus
4) a .... horae iiii ue
5) niant coguntur secun
6) ..... ae .... pos ....
7) … initiis tribus ad
8) exitus viiii propter
9) initium tres sep
10) timae secunda tria
11) in tertia duae h .. ra ..
12) tae et regnarunt
13) de his homines pes

Initial round of advanced processing


TRH notes 3/3/2017

I can read considerably less than Ceriani. However, the image above is the best yet for this page. The suppression of the overtext is impressive. It’s not terribly promising for now, but if we find a good technique this would be a good one to come back to.

Page 117: col b, ll 4-6, 22-24

col. b
1) nunc amplificauit
2) nos dominus deus et multi
3) plicauit nos super
4) terram. et ---
5) ---
6) et --- in anno
7) primo septimana
8) rum primi uisus est
9) dominus ad illum in noc
10) te ipsa in prima die
11) mensis primi et di
12) xit ei ego sum deus abra
13) ham patris tui noli
14) timere quoniam te
15) cum sum ego et be
16) nedicam te et mul
17) tiplicabo semen
18) tuum sicut hare
19) nam terrae prop
20) ter abraham pue
21) rum meum et aedif
22) icauit .......
23) .........
24) .....…

Initial round of advanced processing


TRH notes 3/1/2017

No luck on the upper right but very promising in the lower right. I think I can read several letters and would like to read more. The region is x,y,w,h=3476,5586,2013,629.

Page 112: col a, ll 1-3, 10-11

Work has already been done on this page, at least the first three lines.

column a
1) .........…
2) ..........…
3) .................
4) qui est bis millesi
5) mus et quingente
6) simus annus a crea
7) tura orbis terrae
8) nam secus qui in ori
9) ente sunt numerus
10)  ... mus et .. mus
11) et .... mus profec
12) tionis fynicis cum
13) exiuit plebs post

Initial Round of Advanced Processing


TRH notes 2/27/2017

This is the page with the super-important first three lines, but I’m still not seeing anything other than the other side. I also looked carefully at the beginning of line 10 (x,y,w,h=585,3115,2079,444). I thought I saw something, but now believe it is from the other side. I don’t want to forget this page because it is so important, but maybe table it for now until we think of a different approach.

Page 114: col b, ll 5-6, 11-12

This is the other side of page 113 (above)

4) in uita mea ego mo
5) riar et non uiuam
6) amplius …..
7) …. quinquagin
8) ta annorum sum
9) conplens in uita
10) mea et risit iacob
11) de sermonibus ma
12) tris suae qu….
13) .. ri … ram
14) et ipsa sedebat in
15) conspectu eius et

Initial round of advanced processing


TRH notes 3/3/2017

I’m not seeing much in the images above. The KTK Ruby is the best so far, but I don’t think that will be enough to go beyond or correct Ceriani.

Page 128: col b, ll 12-14

This is the other side of page 127.

10) nequam ex ore eius
11) quoniam quieta et
12) pacifica erat et …
13) ……… hono
14) …….. et erat
15) memor omnium ope

Initial advanced processing


TRH notes 3/3/2017

Not quite readable yet, but cause for optimism in the two missing lines.

The coordinates are x,y,w,h = 2940,3883,2197,368

The challenge is that the overtext is directly on top of the undertext

Page 143: col b, ll 2-4

1) ta sua in manibus
2) meis qui …..
3) ram ex imo ei uesti
4) menta sua …..
5) Et uidet aegyptius ues

Initial advanced processing


TRH notes 3/3/2017

Well that didn’t work.

Is this the same process as the others?

The text that is visible is overtext (the suppression didn’t work).

The undertext is not visible, even for letters that are visible in the accurate color.

Page 97: col a, ll 9, 10, 13

8) uidit et ecce aufer
9) tur ……
10) …s me consti
11) tuit pro eis et pro pec
12) catis eorum …
13) … et in .. ccare..
14) pro eis non enim

Initial advanced processing


TRH 3/3/2017

Sorry… the red box in the initial document is wrong… it should be in the left hand column.

Tough getting separation here… I don’t see much hope in the current approach

Page 127: col a, ll 1-2

Other side of page 128 (above)

1) isaac maiorem …..
2) honorif……
3) portionem jacob

Initial advanced processing


TRH notes

The only thing I’m seeing here is backwards (i.e., from the other side)

I don’t know why Ceriani put five dots in the first line. There is not room for that many letters.

Not a priority to continue here…

Page 129: col b, ll 3-5

1) uinculatorio et nar
2) rauit in conspectu
3) eius duo somnia ….
4) interpraetatus est
5) …………..
6) duo somnia unum
7) sunt septem autem

Initial round of advanced processing


TRH notes 2/27/2017

Nothing encouraging here… had trouble even getting oriented to what Ceriani could read

Page 100: col b, ll 1-2

Other side of page 99 (next)

1) plebe ist ……
2) quis locus recipit …
3) te aut quod erit mo

Initial round of advanced processing


TRH notes 2/27/2017

The above images were not very readable, but interestingly I was able to add to Ceriani’s reading using the royal blue fluorescence. The region is x,y,w,h=3112,1214,2247,623 and where Ceriani read PLEBE IST I could read PLEBE IST RAM. ET

Page 99: col a, ll 18, 24

Other side of page 100 (previous)

17) rendum supra ter
18) ram ..... de uo
19) luntatem eorum
20) praestabo illis ci
21) borum et potui se
22) cus uoluntatem
23) uoluntatis eorum
24) ..... enim illorum

Initial advanced processing


TRH notes 3/4/2017

I was very excited about the lower left until I realized it was from the other side.

Line 18 on the left is hard to tell. There is a lot going on in the pseudo listed above, but the overtext and undertext on both sides are overlapping. The region is x,y,w,h=1653,4829,778,176. Not optimistic…

Page 53: col a, ln 13

11) eorum erat incen
12) dens super altari
13) ……………..
14) ostiam in odorem
15) suauitatis et mane

Initial round of advanced processing


TRH Notes 3/1/2017

The line that Ceriani could not read looks very close to readable in the image above. Some improvement may be enough to make it readable. The coordinates are x,y,w,h=1114,3762,2068,208.

RLE 3/8/2017

these both combine the 940 band after "blur and divide" with the best ICA band to make a pseudocolor (different renderings)

TRH 3/16/2017

I wouldn't necessarily say the two new ones are intrinsically better, but flashing between many images seems to leave an impression. Transmissive seems to be key on the letters I'm seeing.
. . . . . . TRUSTU AS

Page 121: col b, ln 11

column b
10) sit eos ab isac filio suo
11) ---
12) et filii eius et filii

Initial advanced processing


TRH notes 3/4/2017

The pseudo above is the best image of line 11, which isn’t saying a whole lot. Causes for optimism are that the line is between lines of overtext, and I think I see an “N” (not a backwards N). The region is x,y,w,h = 3292,3146,2191,225.

Page 76: col a, ln 1

1) ipsius ---r
2) et angustiae et non
3) est pax propter quod

Initial advanced processing


TRH notes 3/4/2017

I can’t read much now but it looks like it might be close. We should be mostly between lines of overtext. The region is x,y,w,h = 837,1515,2212,252.

Page 69: col b, ln 4

column b
1) princeps exercitus
2) ipsius et cognouit
3) isac in die illa quo
4) niam --- n --- iu
5) rauit ille facere

Initial advanced processing


TRH notes 3/4/2017

That’s not page 069, it’s  075. Not sure what went wrong…

RLE 3/8/2017


TRH 3/16/2017

The second image is the clearer of the two. It's tough getting oriented, but this might be a candidate for contradicting Ceriani. The RAVIT in line 5 and CUM in line 6 are clear enough, but what appears in line 4 can't be NIAM, as Ceriani claims. It would be nice to see better what it could be.

RLE 3/24/2017

Trying some "different" variants. Uploaded one image just now
I think that the "dark text" is what you're after. Doubting that you'll see anything "new", but just perhaps a bit more visible.

TRH 3/29/2017

I could confirm that the desired text is black, but at least one other text is also black. I would bet that's an “A” that Ceriani couldn’t read, but overall I can read much less than Ceriani. The letters are all on top of each other. Maybe something less dependent on transmissive? For now I suggest we work on an easier page (such as 67, next).

Page 67

This page is fully read by Ceriani but one word is worth checking on because the majority of scholarship on the whole Testament of Moses has been concerned with this word. Ceriani reads “TAXO”, and that looks likely enough based on the KTK Ruby, but still, worth checking. An “X” usually crosses higher, so maybe it is not an X.

The coordinates are x,y,w,h = 3438,3233,640,180

April 6, 2017, RLE

New images uploaded to LoadingDock:

May 10, 2017, TRH

I examined all of these with special attention to the reading of the name “Taxo” which is crucial to scholars. My favorite is Amb_C73inf_067_b15-b37+b99-b108_FL_cal_med3+box51_bad_BLUE_W365BGOR+W455BGOR_8bands_ICA_R1G2B3.tif. I put this in the archive and on the IIIF repository, where it is known as Here is how it compares to some of the other images:

Accurate Color Accurate Color
Extended Spectrum Extended Spectrum
KTK Ruby KTK Ruby
Toolkit Pseudo Pseudocolor built into SpectralRTI_Toolkit
RLE 01 RLE01

This was supposed to be “low hanging fruit” in that a reading was proposed before. I’d still like to find something that grabs the attention of scholars a little more quickly. I also wonder if we should step back and try a different approach to the reagent issue.