magenta / mt3

MT3: Multi-Task Multitrack Music Transcription
Apache License 2.0
1.41k stars 186 forks source link

Training with Vocals #7

Open yacaeh opened 2 years ago

yacaeh commented 2 years ago

Hi, this is by far the most accurate transcription!

I know the project is for piano transcription, and as you mentioned in caveat it's not trained on singing vocals. Would it be possible to transcribe as good as piano if I train this with vocals? What should I consider if I want to use this for vocal transcriptions?

AgentHitmanFaris commented 2 years ago

For vocal the transcription not works well. However to simple adjustment can improved te vocal recognition by create the melody by humming or something similar...the programs have difficulties to recognize words