Transcribing Audio Files using Python speech_recognition and ffmpeg, by ChatGPT

transcribe-audio.py

import speech_recognition as sr
import sys
import os
import subprocess
import tempfile

def convert_to_wav(input_file):
    """Convert input audio to a temporary WAV file using ffmpeg"""
    tmp_wav = tempfile.NamedTemporaryFile(delete=False, suffix=".wav")
    tmp_wav.close()

    command = [
        "ffmpeg", "-y",  # overwrite if needed
        "-i", input_file,
        "-ac", "1",      # mono
        "-ar", "16000",  # 16kHz sample rate
        tmp_wav.name
    ]

    try:
        subprocess.run(command, check=True, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
    except subprocess.CalledProcessError:
        print("❌ ffmpeg failed to convert the audio file.")
        sys.exit(1)

    return tmp_wav.name

if len(sys.argv) < 2:
    print("Usage: python3 speech1.py <audio_file>")
    sys.exit(1)

input_path = sys.argv[1]
file_ext = os.path.splitext(input_path)[1].lower()

if file_ext != ".wav":
    print(f"🔄 Converting {file_ext} to .wav...")
    input_path = convert_to_wav(input_path)
    cleanup_needed = True
else:
    cleanup_needed = False

recognizer = sr.Recognizer()
with sr.AudioFile(input_path) as source:
    audio = recognizer.record(source)
    print("📝 Transcribing...")
    try:
        print(recognizer.recognize_google(audio))
    except sr.UnknownValueError:
        print("😕 Could not understand the audio.")
    except sr.RequestError as e:
        print(f"🔌 API error: {e}")

# Clean up temp file
if cleanup_needed:
    os.remove(input_path)

It works with many languages, eg:

recognizer.recognize_google(audio, language="zh-CN")

https://cloud.google.com/speech-to-text/docs/speech-to-text-supported-languages