Python中WAV录音与语音转文本异常求助：录制文件无内容

Hey there! Let's troubleshoot why your output.wav is coming up empty and get that speech recognition working smoothly. Here are the most common fixes and a revised version of your code:

Common Causes & Solutions #

1. Microphone Permissions #

First off, make sure your Python environment has access to your microphone—this is a super common oversight:

• Windows: Head to Settings > Privacy & Security > Microphone, and check that the terminal/code editor running your script has permission.
• macOS: Open System Settings > Privacy & Security > Microphone, and tick the box next to your terminal or IDE.
• Linux: Add your user to the audio group with sudo usermod -aG audio $USER, then log out and back in to apply the change.

2. Incorrect Channel Configuration #

Most built-in microphones are mono (1 channel), but your code uses channels=2 (stereo). If your mic only outputs one channel, recording in stereo can result in silent files. Switch to mono:

myrecording = sd.rec(int(seconds * fs), samplerate=fs, channels=1)

3. Pick the Right Microphone Device #

If you have multiple audio devices (like a webcam mic + headset), sounddevice might be picking the wrong one. List all available devices first to find your microphone's index:

print(sd.query_devices())

Then specify the device in your recording line (replace 0 with your mic's index):

myrecording = sd.rec(int(seconds * fs), samplerate=fs, channels=1, device=0)

4. Verify Recording Before Saving #

Add a quick playback step right after recording to confirm audio is actually being captured:

sd.wait()
print("Playing back your recording...")
sd.play(myrecording, fs)
sd.wait()

5. Tweak Speech Recognition Settings #

Sometimes ambient noise calibration is too aggressive, or the language isn't specified properly. Adjust these parts of your code:

with sr.AudioFile(sound) as source:
    # Shorten ambient noise adjustment (default is 1 second) to avoid cutting your speech
    recognizer.adjust_for_ambient_noise(source, duration=0.5)
    print("Converting the answer to text...")
    audio = recognizer.listen(source)

try:
    # Specify your language if it's not English (e.g., 'en-US', 'fr-FR')
    text = recognizer.recognize_google(audio, language='en-US')
    print("The converted text:", text)
except sr.UnknownValueError:
    print("Oops! Couldn't understand the audio—try speaking more clearly.")
except sr.RequestError as e:
    print(f"Failed to connect to Google's service: {e}")
except Exception as e:
    print(f'An unexpected error occurred: {e}')

Revised Full Code #

import speech_recognition as sr
import sounddevice as sd
import numpy as np
from scipy.io.wavfile import write

fs = 44100  # Sample rate
seconds = 15  # Duration of recording

# List available devices to find your microphone index
print("Available audio devices:")
print(sd.query_devices())
device_index = int(input("Enter your microphone device index: "))

print("Start recording the answer.....")
# Record with mono channel and specified device
myrecording = sd.rec(int(seconds * fs), samplerate=fs, channels=1, device=device_index)
sd.wait()  # Wait until recording is finished

# Play back to verify recording
print("Playing back your recording...")
sd.play(myrecording, fs)
sd.wait()

# Save as WAV file
write('output.wav', fs, myrecording.astype(np.int16))

recognizer = sr.Recognizer()
sound = "output.wav"

with sr.AudioFile(sound) as source:
    recognizer.adjust_for_ambient_noise(source, duration=0.5)
    print("Converting the answer to text...")
    audio = recognizer.listen(source)

try:
    text = recognizer.recognize_google(audio, language='en-US')
    print("The converted text:", text)
except sr.UnknownValueError:
    print("Oops! Couldn't understand the audio. Make sure you're speaking clearly into the mic.")
except sr.RequestError as e:
    print(f"Failed to connect to Google Speech Recognition: {e}")
except Exception as e:
    print(f'An error occurred: {e}')

Quick Extra Tips #

• Speak close to the microphone and minimize background noise while recording.
• If issues persist, try updating your libraries: pip install --upgrade speechrecognition sounddevice scipy numpy
• Test with a different microphone (like a headset) to rule out hardware issues.

内容的提问来源于stack exchange，提问作者Hirushi Ekanayake