Sign up ×
Music: Practice & Theory Stack Exchange is a question and answer site for musicians, students, and enthusiasts. It's 100% free, no registration required.

I conduct research on automatic transcription of polyphonic music. Currently, I use MAPS dataset.

In particular, I need monophonic recordings to be used as training data and polyphonic ones to be used during test. And I need the transcription of the test recordings so as to report how well the system performs.

Do you know any publicly available database that can be used for this purpose? It can involve solo instrument as well as many instruments.

share|improve this question
+1 for a very interesting field of research. – Panagiotis Palladinos Oct 31 '12 at 13:51
For your training and test data, do you just need recordings, or do you also need correct transcriptions? – evanrmurphy Nov 6 '12 at 19:23
I also need ground truth data to test the performance of the algorithm. – petrichor Nov 7 '12 at 8:00

1 Answer 1

This isn't exactly what you were thinking of, but one way you could do it would be to get a lot of free midi files and render them into audio. Then, check the resulting score against the original midi. I'm not sure about the methods you are using to do this, but you could automate the entire process. For example, FluidSynth can render audio files from the command line, so you would only have to feed a midi file into your program and it would do the rest. This would be especially helpful if you are using any AI learning for this program.

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.