NYPO score extraction

[geniusboy98]

I hope I haven't trumped anybody of their job, but I took it upon myself to download the score to Respighi's Sinfonia Drammatica, which is PD (I think, published by UE in 1922) and not on IMSLP. I got it from the NYPO Archives site here: http://archives.nyphil.org/index.php/ar ... intedMusic

using the Java tool documented here: http://imslp.org/wiki/IMSLP:Obtaining_H ... ion_Images

After the long download, I now have 214 JPG files (not TIFF files) that must be put together. It is my understanding that I need to convert these to monochrome and then clean them of conductor marks, then deskew, crop, and compress with CCITT4, but I am stuck on the converting to monochrome part. I have tried various convert/ImageMagick commands that I found both on IMSLP's pages and elsewhere, but I have had no success. I would appreciate some advice on how to fix this and eventually produce a score nice enough to be uploaded.

If it helps, I'm using a Mac, and I have access to ImageMagick and Adobe Acrobat.

Thanks!

[cypressdome]

First, I salute you for tackling the clean up on that particular score. There seem to be markings on every other page! Interestingly, the parts look cleaner than the score. The copyright situation with that score is somewhat interesting. It appears the only person to have conducted it with the NYPO was Willem van Hoogstraten (1884-1965). His "editing" of that score won't be in the public domain in Canada until Jan. 1, 2016 (add another twenty years for the EU). Also, the cover page of that scores lists a copyright date of 1923 which in the U.S. would usually mean the work would still be under copyright (provided the copyright was renewed 28 years after publication). However, the first page of the score has a copyright date of 1922 and I believe that in the U.S. having a score published with two conflicting copyright dates would have invalidated the copyright claim when the work was first published. Nevertheless, the 1922 date places the score in the public domain in the U.S. (without any of conductor's marking).

As for converting the files to black and white I won't be of any help with anything that is Mac-specific but I'll give you the script I use for Imagemagick that I run under Cygwin on Windows 7. I downloaded a test image from the score and it's 2274 by 3000 pixels. NYPO lists the score's dimensions as 11" by 15" so the images are about 200dpi. To convert a single image for testing the black and white threshold I ran the following script:

Code: Select all

convert filename.jpg -resize 300% -filter lanczos -threshold 50% -density 600x600 -compress Group4 filename.jpg.tif

50 seems a bit high but you can test to get what you think looks best. To batch convert all the jpgs I run the following script:

Code: Select all

for f in *.jpg
do convert $f -filter lanczos -resize 300% -threshold 50% -density 600x600 -compress Group4 600bwtif/$f.tif
done

600bwtif is a sub-directory of the directory in which the jpg files reside so all the output tif files will be placed there. I hope this proves to be of some help.

[geniusboy98]

Thank you, this seems to work well! I also grabbed all the parts, so I'll go through those in due time. This will take a few weeks to clean all of it and upload.

[coulonnus]

Beware: if you write a batch AKA script you have to put double % signs because '%' means something in Windows command language. e.g. this batch: @FOR /L %%A IN (%2,1,%3) DO (convert -threshold %1%% p%%A.jpg p%%A.tif) converts a set of jpg's to tif. To convert p1.jpg...p214.jpg with a 60% threshold, type mybatch 60 1 214.

If you always use a 60% threshold say @FOR /L %%A IN (%1,1,%2) DO (convert -threshold 60%% p%%A.jpg p%%A.tif) and type mybatch 1 214.