feat: copy and refacotor the code from DDSP_SVC #1017
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Outdated suggestions cannot be applied.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.
I see the feature request from issue #762. Then, I referenced the newest code from DDSP-SVC 4.0 and refactored the DDSP-SVC code. Here are several changes and the reasons:
According to the recommendation of DDSP-SVC, there's no need to pass the DDSP model; only the Diff model is required. So, I changed the frontend code to allow the model to be empty.
Referring to gui_diff.py in DDSP-SVC's repository, I modified the settings of the DDSP-SVC voice changer.
Known bugs:
When the chunk size is too small, it seems that DDSP-SVC itself will encounter an error related to f0_extraction.
This is my first time doing a PR. If @w-okada can help me merge, I would be really apppreciated for that.