Score splitter program

lailyx · Post by **lailyx** » Sun Oct 05, 2008 2:05 am

Before I go and spend time programming my own, is anyone aware of a free program that can take images from a score (say from an extracted score pdf on this site), and attempt to split it into the different parts?

I don't necessarily need to digitize it, though that would also lead to a solution too of course. It seems something a good quality composition program would be able to do (scan a score and digitize it), though I wouldn't know if they could do it from image (no technical reason why not), but I'm not very familiar with any of those tools, or their costs.

Many thanks.

Post by **Yagan Kiely** » Sun Oct 05, 2008 2:28 am

I don't believe either Finale or Sibelius can. Both can scan it however (I assume with Sibelius anyway).

Vivaldi · Post by **Vivaldi** » Sun Oct 05, 2008 12:16 pm

The old way is to use programs such as Photoshop to manually crop out the intended orchestral parts, page by page, and arrange them in order in a new pdf file. This, however, is both tedious and time consuming.

Website · Post by **vinteuil** » Sun Oct 05, 2008 3:19 pm

On a related subject, are there any pdf-splitting (I.E. individual pages) free programs for linux? It would seem that they all cost a lot, and I just dont like to pay for something that should be free.
BTW, I tried PDFsam, and it didn't work - I don't think it's intended for large files like the one I need to split (BGA 44)

lailyx · Post by **lailyx** » Sun Oct 05, 2008 4:39 pm

Re: Vivaldi -- Yeah, that was what I wanted to avoid. Or rather, automate. It doesn't seem like it would be too hard to write something that would do that process to close enough accuracy (or also to reduce hint time for misses). Just wanted to see if someone has already written it before I re-engineer the wheel.

horndude77 · Post by **horndude77** » Sun Oct 05, 2008 10:56 pm

On a related subject, are there any pdf-splitting (I.E. individual pages) free programs for linux?

Look up pdftk. It's based on the Java pdf library iText. So you could write a program using this library to do it also. The clunkiest option is using ghostscript. In environments that don't allow 3rd party tools (i.e. my work) I've had to make a script which splits the pdf using ghostscript. It's not pretty, but it'd doable.

Before I go and spend time programming my own, is anyone aware of a free program that can take images from a score (say from an extracted score pdf on this site), and attempt to split it into the different parts?

This isn't an easy problem. Many scores omit parts in systems where they just rest. The same part will jump around from system to system. I've also seen horn parts switch the pairing. For one system it will be I+II, III+IV then for the next it will be I, II+III+IV. Many times full scores omit the short instrument names at the beginning of the system. The instruments are determined from context.

I don't think there are any programs out there currently which try to solve this problem. I say go for it. Let us know if you come up with anything. Even if the input was the score and a list of staves to pick out of each system it would still be immensely easier than using an image editor.

lailyx · Post by **lailyx** » Mon Oct 06, 2008 3:58 am

I will guiltily admit that I am a string player, and will use this mostly for splitting quartet scores where most of aforementioned problems are rare. That being said, odds are a lot of good groundwork for handling more complex problems will fall out. Human in the loop capability needs to be there anyway for correction to the algorithm. Time will tell. If anyone else knows anything, I'll still watch this thread, but I'll start playing around this week if I find time. If/when I come up with anything useful, I'll post a new thread.

Website · Post by **vinteuil** » Sun Oct 12, 2008 2:19 am

Look up pdftk. It's based on the Java pdf library iText. So you could write a program using this library to do it also. The clunkiest option is using ghostscript. In environments that don't allow 3rd party tools (i.e. my work) I've had to make a script which splits the pdf using ghostscript. It's not pretty, but it'd doable.

I suck at ghostscript. Period. And Pdftk doesn't install (I'm missing some packages that need some packages that need the first packages)
Thanks anyways!
I'll Try just selecting and saving manually... That might work....[/quote]

horndude77 · Post by **horndude77** » Sun Oct 12, 2008 3:24 am

I suck at ghostscript. Period. And Pdftk doesn't install (I'm missing some packages that need some packages that need the first packages)
Thanks anyways!
I'll Try just selecting and saving manually... That might work....

Here's another solution (really pdftk is the best option however if you can make it work): pdfimages. It spits out a pnm for each page so you lose page size and dpi information. If you know that (or don't care to lose it) you can then use imagemagick (optionally with tiffcp + tiff2pdf) to get a pdf again.

Website · Post by **vinteuil** » Sun Oct 12, 2008 6:51 pm

Thank you horndude! I used pdfimages then printed to pdf using cups-pdf, then pdfsam merged them (it works really well with smaller files, just not the 169 page one I was using).
Thanks again!