通过 Sphinx 引擎进行语音识别会产生 AttributeError: 'NoneType' object has no attribute 'close'

Question

I am trying to transcribe a video using the code below, but I received the following error:

ERROR : transcribe_video_to_text
    audio.close()

AttributeError: 'NoneType' object has no attribute 'close'

Could you please share your ideas to solve this problem, or suggest alternative libraries or AI that could perform the transcription?

Code:

import moviepy.editor as mp
import speech_recognition as sr

def transcribe_video_to_text(video_path):
    # Extract audio from the video
    video = mp.VideoFileClip(video_path)
    audio = video.audio.write_audiofile("temp_audio.wav")

    # Initialize the speech recognition object
    recognizer = sr.Recognizer()

    # Read the audio file
    with sr.AudioFile("temp_audio.wav") as source:
        audio_data = recognizer.record(source)

    # Perform speech recognition on the audio
    text = recognizer.recognize_sphinx(audio_data, keyword_entries=[('whisper', 1.0)])

    # Remove the temporary audio file
    video.close()
    audio.close()
    os.remove("temp_audio.wav")

    return text

# Provide the path to your video file
video_path = "Gods Architect Antoni Gaudis glorious vision.mp4"

# Transcribe the video to text
transcription = transcribe_video_to_text(video_path)

# Print the resulting transcription
print("Video transcription:")
print(transcription)

Answer 1

You're receiving the error because the call to write_audiofile() returns None , and you are trying to call close() on it.

You don't need to store the returned value in audio in the first place. The updated code below will transcribe your video:

import os

import moviepy.editor as mp
import speech_recognition as sr

def transcribe_video_to_text(video_path):
    # Extract audio from the video
    video = mp.VideoFileClip(video_path)
    video.audio.write_audiofile("temp_audio.wav")

    # Initialize the speech recognition object
    recognizer = sr.Recognizer()

    # Read the audio file
    with sr.AudioFile("temp_audio.wav") as source:
        audio_data = recognizer.record(source)

    # Perform speech recognition on the audio
    text = recognizer.recognize_sphinx(audio_data)

    video.close()
    # Remove the temporary audio file
    os.remove("temp_audio.wav")

    return text

# Provide the path to your video file
video_path = "Gods Architect Antoni Gaudis glorious vision.mp4"

# Transcribe the video to text
transcription = transcribe_video_to_text(video_path)

# Print the resulting transcription
print("Video transcription:")
print(transcription)