I occasionally do this to clean up unreadable sections of music.
In terms of software, it depends on the nature of your intended use, and the nature of the undesirable elements from scanning.
Low resolution scans can be difficult to fix.
Common artifacts are skewing of pages, and areas that are out of focus. Photo editing software makes easy work for correcting page skewing. There is also dedicated software such as scan tailor.
Bringing something into focus which was not, can be more of a challenge. You can try sharpening tools.
New ai based tools are being developed, but I do not know of any for music processing.
I'd suggest you try using https://www.musitek.com/ and then 'clean up' in musescore. I haven't personally tried this but I've been playing around with methods for doing optical character recognition (OCR_ on music.
William- Do you mean musescore pro? I have the free software and will look into the subscription based one.
Michael- I have used Paint, PDFescape, and will try out Photopea. Yes, I think I'm leaning towards sharpening tools. The pdfs I work with a lot are from IMSLP. My colleagues and I play the music for fun.
I have been also working with Readiris, but I haven't found it to be very useful with music.
Keep the ideas coming! Much thanks!
~Depends what you want but most likely you'll need Musescore Pro if you're trying to do anything complex.~ From the comment below - looks like just Musescore alone can do the job.
Photoshop is semi viable depending on how bad the scan is. https://www.youtube.com/watch?v=AdZ6G7XFgFU but this is non trivial.
@John I know playscore is semi-decent for consumers and I think you can import directly into Musescore. Honestly, most of my research on the topic is done for pracso.com which means that I'm looking at code solutions that don't work for a consumer.
I would be unenthusiastic about OCR. The worse the score the worse the quality of the OCR output. I think it is less tedious to enter the score manually into Musescore than having to proofread the score afterwards--very tedious work. I volunteered to proof read scans for the Gutenberg project for some months. Even "normal" OCR on printed books is quite unreliable. Proof reading for these errors is extremely tedious work.
The problem of bad scans is actually widespread on IMSLP--rescanning not being an easy option. And some of the errors are simply unfixable (e.g. missing pages, part of the music outside the page) with access to the originals.
By spending some time on the presentation of the score you can often significantly improve those features. Check out any Henle edition for "role models".
In terms of photoshop, it can automate tasks, so if the pages have a certain kind of skew, you can automate the task of correcting it.
Experiment with reducing the color depth. At a certain point you want it to end up in black and white. How early you reduce the depth depends on the color of the artifacts.
If correcting the originals is too time consuming, do consider entering the work in a notation package. Most can use midi or audio input, if working with a keyboard and mouse is not your style.
The supporting tools for music engraving continue to improve, and my programming students who are also musicians are quite enthused about interfaces like those presented by HackLily and Frescobaldi.
This discussion has been archived and is no longer accepting responses.
Violinist.com is made possible by...
Dimitri Musafia, Master Maker of Violin and Viola Cases
Violinist.com Business Directory
Violinist.com Guide to Online Learning
ARIA International Summer Academy
Johnson String Instrument and Carriage House Violins
Discover the best of Violinist.com in these collections of editor Laurie Niles' exclusive interviews.
Violinist.com Interviews Volume 1, with introduction by Hilary Hahn
Violinist.com Interviews Volume 2, with introduction by Rachel Barton Pine