使用MoviePy的cutout方法触发索引越界错误的解决咨询

问题 #

我尝试用MoviePy从音频片段中移除主题曲，代码如下：

audioclip = AudioFileClip(file)
audioclip.write_audiofile(audioFileName)
offset = findOffset(audioFileName, themeFile, 90)
print('\nFound offset: ', offset, ' - Writing audio file again...')
print('Cutting audio from', offset, ' to ', str(offset + themeFileDuration))
cutout = audioclip.cutout(offset, offset + themeFileDuration)
cutout.write_audiofile(cutoutAudioFileName)

代码逻辑为先创建AudioFileClip实例，写入文件后定位主题曲起始偏移量，再调用cutout方法移除对应片段并写入新文件。但执行到写入新文件步骤时，进度到3%就抛出错误：

Found offset:  49.1  - Writing audio file again...
Cutting audio from 49.1  to  93.1
MoviePy - Writing audio in /Users/chud37/Desktop/Friends/output/120 - The Evil Orthodontist_cutout.wav
chunk:   3%|████▋                                                                                                                                                                       | 755/27881 [00:00<00:06, 3914.41it/s, now=None]Traceback (most recent call last):
  File "/Users/chud37/opt/anaconda3/envs/themeTuneRemover/lib/python3.6/site-packages/moviepy/audio/io/readers.py", line 193, in get_frame
    result[in_time] = self.buffer[indices]
IndexError: index -1842399 is out of bounds for axis 0 with size 200000

During handling of the above exception, another exception occurred:

File "ttr.py", line 94, in <module>
        main()
      File "ttr.py", line 91, in main
        convertToAudioFiles(files, directory)
      File "ttr.py", line 60, in convertToAudioFiles
        cutout.write_audiofile(cutoutAudioFileName)
      File "<decorator-gen-47>", line 2, in write_audiofile
      File "/Users/chud37/opt/anaconda3/envs/themeTuneRemover/lib/python3.6/site-packages/moviepy/decorators.py", line 54, in requires_duration
        return f(clip, *a, **k)
      File "/Users/chud37/opt/anaconda3/envs/themeTuneRemover/lib/python3.6/site-packages/moviepy/audio/AudioClip.py", line 210, in write_audiofile
        logger=logger)
      File "<decorator-gen-11>", line 2, in ffmpeg_audiowrite
      File "/Users/chud37/opt/anaconda3/envs/themeTuneRemover/lib/python3.6/site-packages/moviepy/decorators.py", line 54, in requires_duration
        return f(clip, *a, **k)
      File "/Users/chud37/opt/anaconda3/envs/themeTuneRemover/lib/python3.6/site-packages/moviepy/audio/io/ffmpeg_audiowriter.py", line 169, in ffmpeg_audiowrite
        logger=logger):
      File "/Users/chud37/opt/anaconda3/envs/themeTuneRemover/lib/python3.6/site-packages/moviepy/audio/AudioClip.py", line 86, in iter_chunks
        fps=fps, buffersize=chunksize)
      File "<decorator-gen-46>", line 2, in to_soundarray
      File "/Users/chud37/opt/anaconda3/envs/themeTuneRemover/lib/python3.6/site-packages/moviepy/decorators.py", line 54, in requires_duration
        return f(clip, *a, **k)
      File "/Users/chud37/opt/anaconda3/envs/themeTuneRemover/lib/python3.6/site-packages/moviepy/audio/AudioClip.py", line 127, in to_soundarray
        snd_array = self.get_frame(tt)
      File "<decorator-gen-13>", line 2, in get_frame
      File "/Users/chud37/opt/anaconda3/envs/themeTuneRemover/lib/python3.6/site-packages/moviepy/decorators.py", line 89, in wrapper
        return f(*new_a, **new_kw)
      File "/Users/chud37/opt/anaconda3/envs/themeTuneRemover/lib/python3.6/site-packages/moviepy/Clip.py", line 93, in get_frame
        return self.make_frame(t)
      File "/Users/chud37/opt/anaconda3/envs/themeTuneRemover/lib/python3.6/site-packages/moviepy/Clip.py", line 136, in <lambda>
        newclip = self.set_make_frame(lambda t: fun(self.get_frame, t))
      File "/Users/chud37/opt/anaconda3/envs/themeTuneRemover/lib/python3.6/site-packages/moviepy/Clip.py", line 187, in <lambda>
        return self.fl(lambda gf, t: gf(t_func(t)), apply_to,
      File "<decorator-gen-13>", line 2, in get_frame
      File "/Users/chud37/opt/anaconda3/envs/themeTuneRemover/lib/python3.6/site-packages/moviepy/decorators.py", line 89, in wrapper
        return f(*new_a, **new_kw)
      File "/Users/chud37/opt/anaconda3/envs/themeTuneRemover/lib/python3.6/site-packages/moviepy/Clip.py", line 93, in get_frame
        return self.make_frame(t)
      File "/Users/chud37/opt/anaconda3/envs/themeTuneRemover/lib/python3.6/site-packages/moviepy/audio/io/AudioFileClip.py", line 77, in <lambda>
        self.make_frame = lambda t: self.reader.get_frame(t)
      File "/Users/chud37/opt/anaconda3/envs/themeTuneRemover/lib/python3.6/site-packages/moviepy/audio/io/readers.py", line 205, in get_frame
        result[in_time] = self.buffer[indices]
    IndexError: index -1842399 is out of bounds for axis 0 with size 200000

我确认cutout方法的使用符合示例规范，但错误似乎源于MoviePy内部，该如何解决？

解决方案 #

• 避免复用原始AudioFileClip实例：原始实例写入文件后，内部缓存或读取指针可能出现异常。获取偏移量后，重新从写入的音频文件创建新实例再剪切：

audioclip = AudioFileClip(file)
audioclip.write_audiofile(audioFileName)
offset = findOffset(audioFileName, themeFile, 90)
# 重新加载音频文件
new_audioclip = AudioFileClip(audioFileName)
print('\nFound offset: ', offset, ' - Writing audio file again...')
print('Cutting audio from', offset, ' to ', str(offset + themeFileDuration))
cutout = new_audioclip.cutout(offset, offset + themeFileDuration)
cutout.write_audiofile(cutoutAudioFileName)
# 手动关闭实例释放资源
audioclip.close()
new_audioclip.close()
cutout.close()

• 手动拼接音频片段替代cutout方法：绕过cutout的内部bug，直接拆分音频为两段后拼接：

# 导入拼接工具
from moviepy.editor import concatenate_audioclips

# 替换原cutout代码段
part1 = new_audioclip.subclip(0, offset)
part2 = new_audioclip.subclip(offset + themeFileDuration)
cutout = concatenate_audioclips([part1, part2])
cutout.write_audiofile(cutoutAudioFileName)

• 升级MoviePy及Python版本：你当前使用的Python3.6搭配旧版MoviePy，这类索引越界bug可能已在新版本修复。执行命令升级：

pip install --upgrade moviepy

若Python3.6兼容性存在问题，建议升级Python到3.8及以上版本后再更新MoviePy。

• 修复音频文件编码异常：原始音频可能存在损坏或编码问题，先用ffmpeg重新编码：

ffmpeg -i 原始音频文件 -acodec pcm_s16le 重编码后的文件.wav

使用重编码后的文件进行后续处理。

内容的提问来源于stack exchange，提问作者Chud37