在 ffmpeg 中采样准确的音频切片? [英] Sample accurate audio slicing in ffmpeg?

问题描述

我需要将 .wav 格式的音频文件切成 10 秒的块.这些块需要正好是 10 秒，而不是 10.04799988232 秒.

I need to slice an audio file in .wav format into 10 second chunks. These chunks need to be exactly 10 seconds, not 10.04799988232 seconds.

我当前使用的代码是

ffmpeg -i test.wav -ss 0 -to 10 -c:a libfdk_aac -b:a 80k aac/test.aac

ffmpeg version 3.2.2 Copyright (c) 2000-2016 the FFmpeg developers
  built with Apple LLVM version 8.0.0 (clang-800.0.42.1)
  configuration: --prefix=/usr/local/Cellar/ffmpeg/3.2.2 --enable-shared --enable-pthreads --enable-gpl --enable-version3 --enable-hardcoded-tables --enable-avresample --cc=clang --host-cflags= --host-ldflags= --enable-ffplay --enable-libass --enable-libfdk-aac --enable-libfreetype --enable-libmp3lame --enable-libopus --enable-libvorbis --enable-libvpx --enable-libx264 --enable-libx265 --enable-libxvid --enable-opencl --disable-lzma --enable-nonfree --enable-vda
  libavutil      55. 34.100 / 55. 34.100
  libavcodec     57. 64.101 / 57. 64.101
  libavformat    57. 56.100 / 57. 56.100
  libavdevice    57.  1.100 / 57.  1.100
  libavfilter     6. 65.100 /  6. 65.100
  libavresample   3.  1.  0 /  3.  1.  0
  libswscale      4.  2.100 /  4.  2.100
  libswresample   2.  3.100 /  2.  3.100
  libpostproc    54.  1.100 / 54.  1.100
Guessed Channel Layout for Input Stream #0.0 : stereo
Input #0, wav, from '/Users/chris/Repos/mithc/client/assets/audio/wav/test.wav':
  Duration: 00:04:37.62, bitrate: 2307 kb/s
    Stream #0:0: Audio: pcm_s24le ([1][0][0][0] / 0x0001), 48000 Hz, stereo, s32 (24 bit), 2304 kb/s
Output #0, adts, to '/Users/chris/Repos/mithc/client/assets/audio/aac/test.aac':
  Metadata:
    encoder         : Lavf57.56.100
    Stream #0:0: Audio: aac (libfdk_aac), 48000 Hz, stereo, s16, 80 kb/s
    Metadata:
      encoder         : Lavc57.64.101 libfdk_aac
Stream mapping:
  Stream #0:0 -> #0:0 (pcm_s24le (native) -> aac (libfdk_aac))
Press [q] to stop, [?] for help
size=     148kB time=00:00:15.01 bitrate=  80.6kbits/s speed=40.9x    
video:0kB audio:148kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.000000%

此代码不会产生精确的切片，任何想法如何实现?

This code does not produce exact slices, any ideas how can this be accomplished?

推荐答案

不可能*.AAC 音频存储在解码为 1024 个样本的帧中.因此，对于 48000 Hz 的馈送，每帧的持续时间为 0.02133 秒.

Not possible*. AAC audio is stored in frames which decode to 1024 samples. So, for a 48000 Hz feed, each frame has a duration of 0.02133 seconds.

如果您将音频存储在 M4A 等容器中，该容器指示每个数据包的持续时间，则调整最后一帧的持续时间以满足指定的 t/ss-to.但是最后一帧仍然包含完整的 1024 个样本.请参阅下面对 M4A 中指定为 10 秒的无声流的最后 3 帧的读数.比较数据包大小与持续时间.

If you store the audio in a container like M4A which indicates duration per-packet, the duration of the last frame is adjusted to satisfy the specified t/ss-to. But the last frame still contains the full 1024 samples. See the readout below of the last 3 frames of a silent stream specified to be 10 seconds in a M4A. Compare the packet size(s) vis-a-vis the duration.

stream #0:
  keyframe=1
  duration=0.021
  dts=9.941  pts=9.941
  size=213
stream #0:
  keyframe=1
  duration=0.021
  dts=9.963  pts=9.963
  size=213
stream #0:
  keyframe=1
  duration=0.016
  dts=9.984  pts=9.984
  size=214

如果此流最初存储在 .aac 中，则总持续时间不会是 10.00 秒.现在，M4A 是否适合您将取决于您的播放器.

If this stream were originally stored in .aac, total duration would not be 10.00 seconds. Now whether M4A does the trick for you will depend on your player.

*有一种 AAC 变体，可解码为 960 个样本.因此，可以将 48 kHz 音频编码为恰好 10 秒长的流.FFmpeg 不支持这样的 AAC 编码器.AFAIK，包括 iTunes 在内的许多应用程序都无法正确播放此类文件.如果您想按照此规范进行编码，可以在 https://github.com/Opendigitalradio/ODR 上找到一个编码器-AudioEnc

*there is a variant of AAC which decodes to 960 samples. So, a 48 kHz audio could be encoded to a stream exactly 10 seconds long. FFmpeg does not sport such an AAC encoder. AFAIK, many apps including itunes will not play such a file correctly. If you want to encode to this spec, there's an encoder available at https://github.com/Opendigitalradio/ODR-AudioEnc

这篇关于在 ffmpeg 中采样准确的音频切片?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！