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'm trying to diagnose the problem, and one thing I can't compare is the style and quality of my audio vs. the audio used to train the demo.
Since they're public domain anyway (see below, at least two of them), could you provide the links to these, so that users can get a sense of what constitutes useful training data?
Specifically, it would be helpful to hear what range of dynamics, articulations, etc. are used. Was the training done on "musical" examples, or on systematic patterns, like one note at a time through every note and articulation?
Training Data citation on bottom right corner of Demo Page
Violin: Musopen (Creative Commons)
Flute: Musopen (Creative Commons)
Trumpet: University of Rochester
Saxophone: Kevin Malta
The text was updated successfully, but these errors were encountered:
I've been working with the Tone Transfer colab, training my own models, but I'm unable to achieve results as convincing as on the demo webpage.
I'm trying to diagnose the problem, and one thing I can't compare is the style and quality of my audio vs. the audio used to train the demo.
Since they're public domain anyway (see below, at least two of them), could you provide the links to these, so that users can get a sense of what constitutes useful training data?
Specifically, it would be helpful to hear what range of dynamics, articulations, etc. are used. Was the training done on "musical" examples, or on systematic patterns, like one note at a time through every note and articulation?
Training Data citation on bottom right corner of Demo Page
Violin: Musopen (Creative Commons)
Flute: Musopen (Creative Commons)
Trumpet: University of Rochester
Saxophone: Kevin Malta
The text was updated successfully, but these errors were encountered: