fix(separation): prevent leakage removal collar from being applied to…

… diarization
pyannote · Nov 26, 2024 · d043a06 · d043a06
1 parent 4254401
commit d043a06
Show file tree

Hide file tree

Showing 2 changed files with 5 additions and 3 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -36,6 +36,7 @@ Clipping and speaker/source alignment issues in speech separation pipeline have
 
 - fix(separation): fix clipping issue in speech separation pipeline ([@joonaskalda](https://github.com/joonaskalda/))
 - fix(separation): fix alignment between separated sources and diarization ([@Lebourdais](https://github.com/Lebourdais/) and [@clement-pages](https://github.com/clement-pages/))
+- fix(separation): prevent leakage removal collar from being applied to diarization ([@clement-pages](https://github.com/clement-pages/))
 - fix(doc): fix link to pytorch ([@emmanuel-ferdman](https://github.com/emmanuel-ferdman/))
 - fix(task): fix corner case with small (<9) number of validation samples ([@antoinelaurent](https://github.com/antoinelaurent/))
 

diff --git a/pyannote/audio/pipelines/speech_separation.py b/pyannote/audio/pipelines/speech_separation.py
@@ -632,15 +632,16 @@ def apply(
                 )
             )
             if asr_collar_frames > 0:
+                dilated_speaker_activations = np.zeros_like(discrete_diarization.data)
                 for i in range(num_speakers):
                     speaker_activation = discrete_diarization.data.T[i]
                     non_silent = speaker_activation != 0
                     dilated_non_silent = binary_dilation(non_silent, [True] * (2 * asr_collar_frames))
-                    speaker_activation_with_context = dilated_non_silent.astype(np.int8)
-                    discrete_diarization.data.T[i] = speaker_activation_with_context
+                    dilated_speaker_activations.T[i] = dilated_non_silent.astype(np.int8)
 
+            dilated_speaker_activations = SlidingWindowFeature(dilated_speaker_activations, discrete_diarization.sliding_window)
             sources.data = (
-                sources.data * discrete_diarization.align(sources).data
+                sources.data * dilated_speaker_activations.align(sources).data
             )
 
         # separated sources might be scaled up/down due to SI-SDR loss used when training