We closed this forum 18 June 2010. It has served us well since 2005 as the ALPHA forum did before it from 2002 to 2005. New discussions are ongoing at the new URL http://forum.processing.org. You'll need to sign up and get a new user account. We're sorry about that inconvenience, but we think it's better in the long run. The content on this forum will remain online.
Index › Programming Questions & Help › Sound,  Music Libraries › Manuscript recognition
‹ Previous Topic | Next Topic ›
Page Index Toggle Pages: 1
Manuscript recognition (Read 1254 times)
mark hill
Manuscript recognition
Feb 18th, 2006, 8:14am  

Does anyone know anything about the kinds of algorithms used in score recognition software, such as photoScore etc?

Or, how one might go about doing some image-to-notation stuff?

Bit of a vague one this, but I'd appreciate any pointers.
Back to top
 
 
uoou
Re: Manuscript recognition
Reply #1 - Feb 21st, 2006, 8:31pm  
Did you get anywhere with this? It sounds interesting.

I had a think about how I'd do it so this is probably all stuff you've already thought of. And I get stuck pretty quick, but:

First take vertical slices of pixels at equal distances across the image. Comparing these slices for dark bits should give you the lines (forget what they're called) (i.e. dark bits at roughly equal y positions).

Then take horiontal slices just above, below and midway between those lines to look for notes (when you find what you think is a note you'd obviously have to take more slices to check).

And that's where I get stuck. No idea how you'd tell notes appart. A bit of conditional cropping and some blob recognition stuff I guess.

Anyway, just expressing my interest really.
Back to top
 
WWW  
mark hill
Re: Manuscript recognition
Reply #2 - Feb 21st, 2006, 11:52pm  
Well, the plan is:

musical notation is purely symbolic, and therefore one could explore new methods of musical creativity through encoding a drawing as musical notation, that a musician could then attempt to perform.

It would obviously be highly subjective, as I suspect a great deal of musical performance is, but revising a drawing after listening to it performed in a music app might be an interesting way of learning about music.

I agree about scanning the image density for creating staves etc., but it's quite difficult when you begin to think of notes, slurs and the like.

hmmm.


I'm sort of reminded of John Cage's 1960's collection of musical manuscripts, where circuit diagrams and all sorts of things were included as being interpretable by a musician.
Back to top
 
 
uoou
Re: Manuscript recognition
Reply #3 - Feb 22nd, 2006, 2:39am  
Sounds like the sort of thing that's really hard to keep cohesive. But worth the effort.

The way you described it, it's drawing>notation. Or is the intention that a notation could be converted into a drawing too? (which could then be revised and converted to notation again, like you say)

Thinking of the john cage stuff, there's a synthesiser emulator I used a bit ago that provides various different abstractions for what you're doing (including a circuit diagrammy kinda one) which might be interesting to look at if you've not already. I'll find it.

Ere
http://web.hibo.no/~mva/getbuzz.php
It works like this:
http://pasokon-yugi.cool.ne.jp/image/Buzz.gif
Back to top
 
WWW  
Ale_k
Re: Manuscript recognition
Reply #4 - Feb 23rd, 2006, 7:10pm  
Hi Mark,

maybe this could be a good starting point:
https://audiveris.dev.java.net/

Alessandro
Back to top
 
WWW  
Page Index Toggle Pages: 1
‹ Previous Topic | Next Topic ›