pytorch implementation of MEL spectrograms (no librosa req'd) #23
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
29 tasks
mthrok
pushed a commit
to mthrok/audio
that referenced
this pull request
Feb 26, 2021
More clear and intuitive math explanation
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Using the
stftfunction in themasterbranch, this create spectrograms in the power scale, converts those spectrograms to the MEL scale, and also converts the power scale to decibels. This is similar to the basic functionality of the librosa mel spectrograms that is already in the library.Notably, I have not done any checks to see if the pytorch version supports
stft. The current 0.3 branch does not support this function. Additionally, the dimension of the input and the output differ from the original MEL transform. This is because thestftfunction expects (c, n) rather than (n, c). Finally, the input variable names don't correspond to librosa, because librosa has a lot more options than these functions. This is all to say, one couldn't just changeMELtoMEL2.I split this into several transforms so one could create non-MEL or non-db spectrograms if one so desired.