every_holiday/VRCT
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

🐛[bugfix] Model : speaker channel audio_sources["channels"] >= 2 の時sample_rateを調整するように修正

Browse Source
This commit is contained in:
misyaguziya
2024-06-20 00:56:12 +09:00
parent 593eed7988
commit 44b36801ed
1 changed files with 10 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
models/transcription/transcription_transcriber.py
View File

@@ -97,11 +97,19 @@ class AudioTranscriber:
return audio_data return audio_data
def processSpeakerData(self): def processSpeakerData(self):
original_channels = self.audio_sources["channels"]
if original_channels <= 2:
channels = original_channels
sample_rate = self.audio_sources["sample_rate"]
else:
channels = 2
sample_rate = self.audio_sources["sample_rate"]*original_channels/2
temp_file = BytesIO() temp_file = BytesIO()
with wave.open(temp_file, 'wb') as wf: with wave.open(temp_file, 'wb') as wf:
wf.setnchannels(2) wf.setnchannels(channels)
wf.setsampwidth(get_sample_size(paInt16)) wf.setsampwidth(get_sample_size(paInt16))
wf.setframerate(self.audio_sources["sample_rate"]) wf.setframerate(sample_rate)
wf.writeframes(self.audio_sources["last_sample"]) wf.writeframes(self.audio_sources["last_sample"])
temp_file.seek(0) temp_file.seek(0)
with AudioFile(temp_file) as source: with AudioFile(temp_file) as source: