Big SBB files

[WilliamBunting]

SBB has some treasures, but they seem to be 1MB per page. Is there a way to shrink them ? (I use Gimp, PDF edit and Windows 7).

Regards WB

[daphnis]

Can you point us to a sample we might consider? I don't normally work with those files.

[coulonnus]

I guess you mean Staatsbibliothek zu Berlin http://digital.staatsbibliothek-berlin.de/dms/suche/. It seems that in Germany the primary concern is ... a faithful reproduction of the colour of the paper! e.g. see http://www.beethoven-haus-bonn.de/sixcm ... Digitising Did you try to convert to monochrome?

[kalliwoda]

These are 400 dpi color scans (4500 pixel), and apparently already heavily compressed using something like jpg-compression to reach 1-1.5 mb per page of the pdf downloads. Decent prints can be converted to 600 dpi monochrome - see here http://imslp.org/wiki/Variations_sur_un ... rol_Józef)
for a side-by-side comparison, but some of the less well preserved prints and the manuscripts are not that readable in monochrome.

Maybe some reduction of resolution to 2000 pixel (200 dpi) would still preserve enough readability, but please don't reduce resolution to less than 1000 pixel to get small file size!

Most of the SBB offerings seem to be special orders but they digitized some of their collection focused on Beethoven, Boccherini, Mozart, Quantz, Weber etc.

[coulonnus]

SBB has some treasures, but they seem to be 1MB per page. Is there a way to shrink them ? (I use Gimp, PDF edit and Windows 7).

kalliwoda made an excellent work on IMSLP, but do you mean there is already a PDF present on SBB? Then please provide the link for this Lipiński-Cenerentola. I couldn't find it. Is it colour or monochrome?

Here the few jpg single pages I have downloaded from SBB are about .4MB big (about 1400x1800 pixels, colour).

[kalliwoda]

If you have a page display on the SBB site open (like the one I linked to from the Lipinski), you can find a "Werkzeugkasten" button. In the Werkzeugkasten you find a pdf icon, and it will allow you download of a multipage high-res pdf file. This was my starting material for the uploads to imslp. And they actually encourage mirroring of their content in their "Nutzungsbedingungen", if it is non-commercial.
So these scans are far better than what is offered for free by Beethoven-Haus.

For the larger works pdf-files may be several hundred MB(!) but become somewhat more manageable if you split into individual parts, or movements.

Unfortunately, monochrome pdf is not on offer, you would have to do the conversion yourself before uploading to imslp.

[coulonnus]

If you have a page display on the SBB site open (like the one I linked to from the Lipinski), you can find a "Werkzeugkasten" button. In the Werkzeugkasten you find a pdf icon

Ouch, isn't it small!

and it will allow you download of a multipage high-res pdf file.

Ouch again, 40 MB for 26 pages!

Piupianissimo has explained the right method viewtopic.php?f=12&t=2163
starting from the png's (resize by at least 300% before you convert to monochrome)

[kalliwoda]

Keep in mind that piupianissimo was trying to get the maximum out of 120 dpi scans. With the 400 dpi SBB scans you can just change directly to 600 dpi, convert to greyscale, choose a splitpoint, apply the dust filter - no need to fill up your harddisk with these huge images enlarged by another 300% ...

Different question: what is the most agreeable way to upload the manuscripts that cannot be converted to readable monochrome??

[coulonnus]

Different question: what is the most agreeable way to upload the manuscripts that cannot be converted to readable monochrome??

Please provide an example. I'll see what I can do.

[kalliwoda]

You could have a look at the original scans from SBB for Krommer Harmonie op.71 - I basically gave up on achieving a usable solution and uploaded my typeset and the originals purely for reference.
But I should ask this question directly: The Zauberflöte autograph as uploaded by Generoso is clearly much higher resolution than the 1050x1400 pixel jpg displayed at SBB, but at about 300k per page much smaller than the pdf available from SBB - how?

[coulonnus]

You could have a look at the original scans from SBB for Krommer Harmonie op.71 - I basically gave up on achieving a usable solution and uploaded my typeset and the originals purely for reference.

I'll have a look tomorrow.

YBut I should ask this question directly: The Zauberflöte autograph as uploaded by Generoso is clearly much higher resolution than the 1050x1400 pixel jpg displayed at SBB, but at about 300k per page much smaller than the pdf available from SBB - how?

I downloaded the pdf's of the Ouverture of both versions. (let's forget about jpg's). That's 38MB vs. 9MB. I magnified the same portion of each version a lot and I compared them. In the SBB version a staff line is about 12-pixel thick and in Generoso's version it is only about 4-pixel thick. I think this is the answer

[WilliamBunting]

Dear colleagues,

Thank you for your advice.

It was my uploading of Zum goldenen Löwen (Seyfried, Ignaz) overture for p duet which prompted my question.
The Werkzeugkasten was easily found.

Regards WB

[coulonnus]

It was my uploading of Zum goldenen Löwen (Seyfried, Ignaz) overture for p duet which prompted my question.

Ah, next time, please provide the site URL, composer and piece. I thought you meant Soprano-Bass-Bass

I have uploaded my version along with yours. My method is:
pdftk to get rid of SBB's "introduction", "appendix" and "courtesy blank pages" and "burst" this big pdf into single-page pdf's.
With Imagemagick: convert -density 400 page.pdf page.jpg
With Irfanview: rotation, crop and sometimes unskew
With Imagemagick: convert -threshold 50% page.jpg page.eps
Then - perhaps some will dislike - I made a LaTeX file with {\epsfig{figure=mypage.eps, height=20.9cm, width=29.7cm} and I made the pdf the usual way.

I concede that an autograph or Krommer's Harmonie don't give a very good pdf because of the lesser contrast in the source pdf. I hope my version of Seyfried will be usable for 4 pianist's hands.

[Mazin]

I have uploaded my version along with yours. My method is:
pdftk to get rid of SBB's "introduction", "appendix" and "courtesy blank pages" and "burst" this big pdf into single-page pdf's.
With Imagemagick: convert -density 400 page.pdf page.jpg
With Irfanview: rotation, crop and sometimes unskew
With Imagemagick: convert -threshold 50% page.jpg page.eps
Then - perhaps some will dislike - I made a LaTeX file with {\epsfig{figure=mypage.eps, height=20.9cm, width=29.7cm} and I made the pdf the usual way.
.

Why not pdfimages to extract the images from the PDF rather than re-rendering the PDF to a JPG? It would probably also skip the intro and blank pages if there are no images on those.