File size: 7,296 Bytes
604255a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 |
# import threading
# import pyaudio
# import wave
# import io
# import time
# from groq import Groq
# from pydub import AudioSegment
# import numpy as np
# class SpeechTotext:
# def __init__(self):
# self.client = Groq()
# self.is_recording = False
# self.frames = []
# self.chunk = 1024
# self.format = pyaudio.paInt16
# self.channels = 1
# self.rate = 44100
# self.p = pyaudio.PyAudio()
# # Silence detection parameters
# self.silence_threshold = -35.0 # Adjusted threshold to be more lenient
# self.silence_duration = 3.0 # seconds
# self.buffer_duration = 0.1 # seconds for each audio chunk analysis
# self.silent_chunks = 0
# self.chunks_per_second = int(1 / self.buffer_duration)
# def detect_silence(self, audio_segment):
# """Check if the audio chunk is silent using pydub"""
# return audio_segment.dBFS < self.silence_threshold
# def record_audio(self):
# stream = self.p.open(
# format=self.format,
# channels=self.channels,
# rate=self.rate,
# input=True,
# frames_per_buffer=self.chunk
# )
# self.frames = []
# buffer_samples = int(self.buffer_duration * self.rate)
# while self.is_recording:
# # Read enough chunks to fill our buffer duration
# buffer_data = b''
# chunks_needed = max(1, int(buffer_samples / self.chunk))
# for _ in range(chunks_needed):
# data = stream.read(self.chunk)
# buffer_data += data
# self.frames.append(data)
# # Convert the buffer to pydub AudioSegment
# audio_buffer = AudioSegment(
# data=buffer_data,
# sample_width=self.p.get_sample_size(self.format),
# frame_rate=self.rate,
# channels=self.channels
# )
# # Check for silence
# if self.detect_silence(audio_buffer):
# self.silent_chunks += 1
# if self.silent_chunks >= self.silence_duration * self.chunks_per_second:
# print(f"Silence detected for {self.silence_duration} seconds, stopping recording...")
# self.is_recording = False
# break
# else:
# self.silent_chunks = 0 # Reset silent chunk counter when sound is detected
# stream.stop_stream()
# stream.close()
# def start_recording(self):
# """Start recording audio"""
# self.is_recording = True
# self.silent_chunks = 0
# threading.Thread(target=self.record_audio).start()
# def stop_recording(self):
# """Stop recording audio and transcribe"""
# self.is_recording = False
# print("Recording stopped")
# # Save the recorded audio to a BytesIO object
# wav_buffer = io.BytesIO()
# with wave.open(wav_buffer, 'wb') as wf:
# wf.setnchannels(self.channels)
# wf.setsampwidth(self.p.get_sample_size(self.format))
# wf.setframerate(self.rate)
# wf.writeframes(b''.join(self.frames))
# # Rewind the buffer and transcribe
# wav_buffer.seek(0)
# try:
# transcription = self.client.audio.transcriptions.create(
# file=("audio.wav", wav_buffer),
# model="whisper-large-v3-turbo"
# )
# print(f"Transcript: {transcription.text}")
# except Exception as e:
# print(f"Error while transcribing audio: {str(e)}")
# finally:
# wav_buffer.close()
# def cleanup(self):
# """Cleanup PyAudio"""
# self.p.terminate()
# if __name__ == "__main__":
# recorder = SpeechTotext()
# try:
# print("Starting recording... (will stop after 3 seconds of silence)")
# recorder.start_recording()
# # Wait for recording to finish
# while recorder.is_recording:
# time.sleep(0.1)
# recorder.stop_recording()
# finally:
# recorder.cleanup()
# Upper one using pydub to detect silence if needed in future versions
import pyaudio
import wave
import io
from array import array
from groq import Groq
class SpeechToText:
def __init__(self):
self.client = Groq()
self.chunk = 4096
self.format = pyaudio.paInt16
self.channels = 1
self.rate = 16000
self.silence_threshold = 1000
self.silence_duration = 3.0
self.frames_per_chunk = self.chunk / self.rate
self.chunks_for_silence = int(self.silence_duration / self.frames_per_chunk)
def record_and_transcribe(self):
"""Records audio until 3 seconds of silence and returns the transcription."""
p = pyaudio.PyAudio() # Create a new PyAudio instance for each request (previously when declared on constructor cant execute more than once)
stream = p.open(
format=self.format,
channels=self.channels,
rate=self.rate,
input=True,
frames_per_buffer=self.chunk
)
frames = []
silent_chunk_counter = 0
print("Recording started... (will stop after 3 seconds of silence)")
while True:
try:
data = stream.read(self.chunk, exception_on_overflow=False)
frames.append(data)
# Detect silence
audio_data = array('h', data)
if max(abs(x) for x in audio_data) < self.silence_threshold:
silent_chunk_counter += 1
if silent_chunk_counter >= self.chunks_for_silence:
print(f"Detected {self.silence_duration} seconds of silence, stopping...")
break
else:
silent_chunk_counter = 0
except IOError as e:
print(f"Error recording: {e}")
break
stream.stop_stream()
stream.close()
p.terminate() # Ensure PyAudio is completely closed
wav_buffer = io.BytesIO()
try:
with wave.open(wav_buffer, 'wb') as wf:
wf.setnchannels(self.channels)
wf.setsampwidth(p.get_sample_size(self.format))
wf.setframerate(self.rate)
wf.writeframes(b''.join(frames))
wav_buffer.seek(0)
transcription = self.client.audio.transcriptions.create(
file=("audio.wav", wav_buffer),
model="whisper-large-v3-turbo"
)
return transcription.text
except Exception as e:
print(f"Error transcribing: {e}")
return str(e)
finally:
wav_buffer.close()
if __name__ == "__main__":
recorder = SpeechToText()
transcribed_text = recorder.record_and_transcribe()
if transcribed_text:
print(f"Transcription: {transcribed_text}") |