You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I would like to have your permission to use your dataset to create "piano, guitar, back/lead vocals (and may be wind etc)" quality checker. Can I use it for this?
For this I will cut some parts of your dataset and create mixtures of 1 minute length. And I'll post them as archive on site. So developers of MSS models will download this archive process it with their model and upload result back to check how good their models are based on SDR metrics.
I will give you credits on page with links on your paper and repository.
I have other ideas for collaborations. For example may be you will be interested to create fully hidden quality checker datasets and put them on your site with checking system (I can help you with code). We need these leaderboards to create better MSS models and compare which one are the best.
The text was updated successfully, but these errors were encountered:
Hey @ZFTurbo, thanks for your interest in MoisesDB. Unfortunately, due to our agreement with our provider, the MoisesDB and other datasets created using MoisesDB can only be distributed by Moises.ai. We would happily host the processed mixtures you created on our website. Does this work for you?
Hello. I have quality checker for vocal/bass/drums/other/instrumental music source separation (MSS) here:
https://mvsep.com/en/quality_checker
I would like to have your permission to use your dataset to create "piano, guitar, back/lead vocals (and may be wind etc)" quality checker. Can I use it for this?
For this I will cut some parts of your dataset and create mixtures of 1 minute length. And I'll post them as archive on site. So developers of MSS models will download this archive process it with their model and upload result back to check how good their models are based on SDR metrics.
I will give you credits on page with links on your paper and repository.
I have other ideas for collaborations. For example may be you will be interested to create fully hidden quality checker datasets and put them on your site with checking system (I can help you with code). We need these leaderboards to create better MSS models and compare which one are the best.
The text was updated successfully, but these errors were encountered: