Real-Time Transcription Intermediate

import pyaudio
from google.cloud import speech

# Audio recording parameters
RATE = 16000
CHUNK = 1600  # 100ms chunks

client = speech.SpeechClient()
config = speech.RecognitionConfig(
    encoding=speech.RecognitionConfig.AudioEncoding.LINEAR16,
    sample_rate_hertz=RATE,
    language_code="en-US",
    enable_automatic_punctuation=True,
)
streaming_config = speech.StreamingRecognitionConfig(
    config=config,
    interim_results=True,  # Get partial results
)

# Open microphone stream
audio = pyaudio.PyAudio()
stream = audio.open(
    format=pyaudio.paInt16, channels=1,
    rate=RATE, input=True, frames_per_buffer=CHUNK
)

def audio_generator():
    while True:
        data = stream.read(CHUNK, exception_on_overflow=False)
        yield speech.StreamingRecognizeRequest(audio_content=data)

requests = audio_generator()
responses = client.streaming_recognize(streaming_config, requests)

for response in responses:
    for result in response.results:
        if result.is_final:
            print(f"Final: {result.alternatives[0].transcript}")
        else:
            print(f"Partial: {result.alternatives[0].transcript}", end="\r")

import azure.cognitiveservices.speech as speechsdk

speech_config = speechsdk.SpeechConfig(
    subscription="YOUR_KEY", region="eastus"
)

# Use default microphone
audio_config = speechsdk.AudioConfig(use_default_microphone=True)
recognizer = speechsdk.SpeechRecognizer(
    speech_config=speech_config, audio_config=audio_config
)

# Event handlers for real-time results
def on_recognizing(evt):
    print(f"Partial: {evt.result.text}", end="\r")

def on_recognized(evt):
    print(f"Final: {evt.result.text}")

recognizer.recognizing.connect(on_recognizing)
recognizer.recognized.connect(on_recognized)

# Start continuous recognition
recognizer.start_continuous_recognition()
input("Press Enter to stop...\n")
recognizer.stop_continuous_recognition()

from faster_whisper import WhisperModel
import numpy as np
import pyaudio
import webrtcvad

model = WhisperModel("base", compute_type="int8")
vad = webrtcvad.Vad(3)  # Aggressiveness 0-3

RATE = 16000
CHUNK_DURATION = 0.03  # 30ms for VAD
CHUNK = int(RATE * CHUNK_DURATION)

audio = pyaudio.PyAudio()
stream = audio.open(format=pyaudio.paInt16, channels=1,
                    rate=RATE, input=True, frames_per_buffer=CHUNK)

buffer = []
silence_count = 0

print("Listening... (Ctrl+C to stop)")
while True:
    data = stream.read(CHUNK, exception_on_overflow=False)
    is_speech = vad.is_speech(data, RATE)

    if is_speech:
        buffer.append(data)
        silence_count = 0
    elif buffer:
        silence_count += 1
        if silence_count > 30:  # ~1 second of silence
            audio_data = np.frombuffer(b"".join(buffer), dtype=np.int16)
            audio_float = audio_data.astype(np.float32) / 32768.0
            segments, _ = model.transcribe(audio_float, beam_size=3)
            for seg in segments:
                print(seg.text, flush=True)
            buffer = []
            silence_count = 0