有什么方法可以在不先将文件写入内存的情况下执行此操作吗？

Question

我有这段代码，可以根据

link

转录一些 YouTube 视频的音频流， 现在你可能会发现这很慢，因为我必须首先将视频流下载为

.mp4

，然后使用

moviepy

将其转换为

.wav

，然后录制音频，然后转录它。 我想要拥有完全相同的功能，但不需要首先通过将流的数据写入某个缓冲区来下载流。

from pytube import YouTube
import speech_recognition as sr
from moviepy.editor import *

video = YouTube(url=link)
audio_stream = video.streams.get_by_itag(140)

recognizer = sr.Recognizer()

audio_stream.download(filename="mp4_output.mp4")
audio = AudioFileClip("mp4_output.mp4")
audio.write_audiofile("wav_output.wav")

                    
with sr.AudioFile("./wav_output.wav") as audio_file:
    audio_data = recognizer.record(audio_file, duration=100)
    transcript = recognizer.recognize_sphinx(audio_data=audio_data)

我尝试过以下方法

from pytube import YouTube
import speech_recognition as sr

video = YouTube(url=link)
audio_stream = video.streams.get_by_itag(140)

buffer = io.BytesIO()
audio_stream.stream_to_buffer(buffer)
                    
recognizer = sr.Recognizer()

with sr.AudioFile(buffer) as audio_file:
    audio_data = recognizer.record(audio_file, duration=100)
    transcript = recognizer.recognize_sphinx(audio_data=audio_data) 

我收到以下错误

audio file could not be read as PCM WAV, AIFF/AIFF-C, or Native FLAC; check if file is corrupted or in another format

有没有办法将

buffer

转换为其中一种格式？

Answer 1

您可以使用

pydub

库，它可以帮助进行音频格式转换：

pip install pydub ffmpeg-python

然后，您可以修改脚本以将音频流从缓冲区转换为 WAV 格式，然后再与语音识别库一起使用：

from pytube import YouTube
import speech_recognition as sr
import io
from pydub import AudioSegment

# Your YouTube link
video = YouTube(url=link)
audio_stream = video.streams.get_by_itag(140)

buffer = io.BytesIO()
audio_stream.stream_to_buffer(buffer)

# Move back to the start of the BytesIO object
buffer.seek(0)

# Convert the audio stream to WAV format using pydub
audio_segment = AudioSegment.from_file(buffer, format="mp4")
wav_buffer = io.BytesIO()
audio_segment.export(wav_buffer, format="wav")

# Reset the buffer position to the start
wav_buffer.seek(0)

recognizer = sr.Recognizer()

with sr.AudioFile(wav_buffer) as audio_file:
    audio_data = recognizer.record(audio_file)
    transcript = recognizer.recognize_sphinx(audio_data=audio_data)