At the International Association for Neo-Latin Studies Vacation School in Bonn, on the theme of Digital Neo-Latin Studies, Antonia Karaisl introduced us to Rescribe, “desktop tool for OCR, optimised for historical printed works”. We were impressed by Rescribe’s results. These days there are a lot of digital facsimiles of old books in Latin and I could hardly wait to test the software myself.
The test was a book by a Jesuit Ioannes Luccarus (Ivan Lukarević, Dubrovnik, 19 Sept 1622 – Rome, 2 March 1709), a posthumous publication of Luccarus’ Latin play on Stanislaus Kostka: Stanislaus Kostka. Drama sacrum authore patre Ioanne Luccaro, Soc. Jesu. Romae: Ex officina Cajetani Zenobii, 1709. A facsimile of the copy from the National Central Library of Rome is freely available at Archive.org (with the persistent identifier ark:/13960/t1jh6kj53), as digitized by Google.
The results of my OCR adventure with Rescribe are indeed amazing. Here is the last page of the book (page 42, scan 52) and below the image you will find OCR as produced by Rescribe, without any changes or corrections. It looks very, very usable:
42 A CT US.
Pa. Nunc omnis in me vindicis Olympi ruat
Fragor. Tonantis ira quid ſegnis jaces 2
Quin vile caput hoc turbine in feſto petis :
Preme, percute, Feri; Non poteſt in me tut
Eirare fulmen. Fratris hic peſtem vides.
Egomet pudicum, ſobrium, mitem, pium
Superis amicum, luce perfuſum ſacra,
ILntaminatum labe, prodigiis gravem
Immitis, excors, durus, implacabilis
PFratrem Tyrannus expuli. Aſſultus meos,
Meos furores fugit. Heu quæ nunc vagus
Deſerta luſtras loca peregrino pede ?
Ubi frater erras ? forſan immanis fera
Tuo cruore luſit. O luctus patris l
Ocivium ſpes irrita O matris dolor
O lux abacta patriæ o fratris ſcelus!
Crat. Fletus inanes amove, ac potius tuo
Gratare fratri, vota qui cordi tenet-
Non ille te, ſed gravibus effugit malis
Mundum ſcatentem. Roma terrarum caput
Juveni viarum eſt meta; nec honorum fame
Stimulatus arces ille Romuleas petit;
Sed amoris actus ſlamine. Ibi ſacros colet
Jeſu penates Cantabri miles Ducis.
.a. Si vota cordis obtines frater, licet
Tibi gratulari jure. Tu Cœlum rapis,
Terram reliquis mini; Sed affectus meos
Ipſe quoque ad Aſtra transferam. Tecum juva
Sociare corda Tendere in Cęlum juvat.
Gor. Portenta rerum mira Perfidiam in fidem,
Odium in amorem Numinis mutat manus-
Hoſpitia ſubeunt noſtra concordi gradu
Religio, Pietes, Pax, Fides, Pudor, Salus.
Nos ergo tot cęleſtibus ovantes bonis
Raũt colamus proſperum plauſu Diem,
-ÆN-
ſõ f I N I S.
N,
b .
.&5 dm ︎4 h
&α
︎n
& . ]
-
-
-
For comparison, here is the OCR available from the Archive.org (done with ABBYY FineReader 11.0):
42 • A C! T U S .
ra. Nunc omnls in mc vindicis Olympi ruat
Fraqor . Tonaniis iraquid fcgnis jaccs/
Quin vilc caput hoc turbinc in feflo pctis /
Pfcmc , pcrcuic, Fcri ; Nonpotcft in me tuui
Eirarc fulmc n . Fratris hic pcftem vidcs .
ligomct pudicum « fobrium t mitem t pium t
Supcris amicum , lucc pcrfufum facra ,
Intaminatum labc i prodigiisgravcm
^ Immitis , cxccrs , durus , implacabilis
Fraircm Tyrannuscxpuli . Affultus meo6 1
Mco^ furorcs fugit . Hcu qux nunc vagus
Dcfcrta iuftras loca pcrcgrino pcdc /
Ubi fraccrerras / forfan immanis fcra
'Tuocruorc lufit • O luftus patris I
OcivJum fpes irrita J O matris dolor !
O lux abafta patria! ! o fratris fcclus !
Crat. Flctus inancs amovc , ac potius tuo
Gratarc fratri t vota qui cordij tcnct •
Non illc tc 1 fcd gravibus cfFugit malis
Mundum fcatentcm . Roma terrarum caput
Juvcni viarum cft meta ; ncc honorum famc
Stimulatus arccs illc Romulcas pctit ; <
Scd amoris aftus flamine . ibi facros colct
Jcfu penaitsCantabri miles Ducis .
3P<j. Si vota cordis obtincs fratcr % licct
Tibi gratulari jurc . Tu Coelum rapis ♦
Tcrram rcliquismmi ;
Ipfc quoque ad Aftra iransferam.Tecum juvat
Sociare corda . Tendere in Cq\um j«va« •
^ 'Cot ' Portenta rerum mira ! Pcrfidiam in fidem t
Odium in amorem Numinis murat manus.
Hofpitia fubeunt noftra concordi gradu
Rcligio , Pietes , Pax , Fides , Pudor , Salus .
>v. Nos ergo tot c^lcftibws ovantes bonis
/^^^fto colamus profperum plaufu Dicm ,
Googlj
On my long-serving Thinkpad X220 (running Manjaro Linux) Rescribe worked without any hiccups. The only stumbling point with Rescribe was exactly four-digit ending of the png filename (I did not read closely, and had three digits in the name), which the instructions very explicitly warn you about.
Rescribe is a great tool, many thanks to people who are developing it!
Technical details
The following steps were done on Linux command line:
- Export images for pages 3-52 from the archive.org DJVU file (all images in ppm format, black and white), add page numbering (four digits):
ddjvu -format=pnm -page=3-52 -eachpage lukarevic-kostka-bub_gb_ldx6wPj7IMMC.djvu lucc-kostka%04d.ppm - Convert PPM images to PNG:
mogrify -format png *.ppm - Run Rescribe on all PNG files in the
lucc/pngdirectory:/rescribe lucc/png/
And that is all!
(Doors photo by Monstera from Pexels.)