admin-valentin/WhisperLiveKit
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

small code review

Browse source
This commit is contained in:
Dominik Macháček 2024-08-19 10:43:29 +02:00
parent 4a89935ee5
commit 827425bb91
1 changed files with 3 additions and 5 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

8
whisper_online.py
View file

@ -558,7 +558,6 @@ class VACOnlineASRProcessor(OnlineASRProcessor):
def insert_audio_chunk(self, audio): def insert_audio_chunk(self, audio):
res = self.vac(audio) res = self.vac(audio)
print(res)
self.audio_buffer = np.append(self.audio_buffer, audio) self.audio_buffer = np.append(self.audio_buffer, audio)
if res is not None: if res is not None:
@ -584,11 +583,10 @@ class VACOnlineASRProcessor(OnlineASRProcessor):
if self.status == 'voice': if self.status == 'voice':
self.online.insert_audio_chunk(self.audio_buffer) self.online.insert_audio_chunk(self.audio_buffer)
self.current_online_chunk_buffer_size += len(self.audio_buffer) self.current_online_chunk_buffer_size += len(self.audio_buffer)
if self.status is not None:
self.clear_buffer() self.clear_buffer()
else: # we are at the beginning of process, no voice has ever been detected else:
# We keep the 1s because VAD may later find start of voice in it. # We keep 1 second because VAD may later find start of voice in it.
# But trimming it to prevent OOM. # But we trim it to prevent OOM.
self.buffer_offset += max(0,len(self.audio_buffer)-self.SAMPLING_RATE) self.buffer_offset += max(0,len(self.audio_buffer)-self.SAMPLING_RATE)
self.audio_buffer = self.audio_buffer[-self.SAMPLING_RATE:] self.audio_buffer = self.audio_buffer[-self.SAMPLING_RATE:]