Added waveform & spectrogram verification to integration tests #209
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
…mparison in integration tests, except for demucs which is pretty inconsistent with itself
…n integration test
…is for waveform images, but spectrograms are still all over the place
…es work as expected
… artifacts for CI
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
At the moment, the integration tests only check that the output files exist, not that the audio separation actually separated the stems correctly.
This PR makes the integration tests way more meaningful and valuable, by adding expected waveform & spectrogram images for the separated output audio stems, for the test track, for every model included in the integration test.
The test then compares the similarity of the expected vs. actual waveform/spectrogram images using the SSIM image similarity index, and only passes if the similarity is > 90%.
As such, if the actual audio separation implementation gets broken, or one of these model files gets corrupted or accidentally swapped etc. this test should fail, as the expected output audio stems will no longer match.