diff options
Diffstat (limited to 'transcribe.py')
| -rw-r--r-- | transcribe.py | 41 |
1 files changed, 27 insertions, 14 deletions
diff --git a/transcribe.py b/transcribe.py index 206dc22..1aabf6f 100644 --- a/transcribe.py +++ b/transcribe.py @@ -12,6 +12,7 @@ from pydub import effects as pydub_effects # python3 -m pip install pyaudio # License: MIT. import pyaudio +import numpy as np import steamvr import string_matcher import sys @@ -32,7 +33,7 @@ class AudioState: # The maximum length that recordAudio() will put into frames before it # starts dropping from the start. - MAX_LENGTH_S = 25 + MAX_LENGTH_S = 30 # The minimum length that recordAudio() will wait for before saving audio. MIN_LENGTH_S = 1 @@ -192,22 +193,33 @@ def transcribe(audio_state, model, filename): audio = whisper.load_audio(filename) audio_state.lock.release() - audio = whisper.pad_or_trim(audio) + audio = whisper.pad_or_trim(audio, length = audio_state.RATE * + audio_state.MAX_LENGTH_S) + mel = whisper.log_mel_spectrogram(audio).to(model.device) - result = whisper.transcribe(model, audio, language=audio_state.language) + result = None + #for temp in (0.00, 0.05, 0.10, 0.15, 0.20): + for temp in (0.00, 0.05): + print("temp: {}".format(temp)) + options = whisper.DecodingOptions(language = audio_state.language, + beam_size = 5, temperature = temp) + result = whisper.decode(model, mel, options) - for segment in result["segments"]: - if segment["no_speech_prob"] > 0.60: - print("no speech prob: {}".format(segment["no_speech_prob"])) - return None - if segment["avg_logprob"] < -1.0: - print("avg logprob: {}".format(segment["avg_logprob"])) - return None - if segment["compression_ratio"] > 2.4: - print("compression ratio: {}".format(segment["compression_ratio"])) + if result.avg_logprob < -1.0: + print("avg logprob: {}".format(result.avg_logprob)) + continue + + if result.compression_ratio > 2.4: + print("compression ratio: {}".format(result.compression_ratio)) + continue + + if result.no_speech_prob > 0.60: + print("no speech prob: {}".format(result.no_speech_prob)) return None - return result["text"] + return result.text + + return None def transcribeAudio(audio_state, model): while audio_state.run_app == True: @@ -256,7 +268,8 @@ def transcribeAudio(audio_state, model): #new_words = text.split() audio_state.text = string_matcher.matchStrings(audio_state.text, - text, window_size = 5) + text, window_size = 30) + #audio_state.text = text if old_text != audio_state.text: # We think the user said something, so reset the amount of # time we sleep between transcriptions to the minimum. |