using FFMPEG with silencedetect to remove audio silence

Question

I am trying to use the following command with the latest ffmpeg build to remove silence from my .mp3 files:

ffmpeg -i SILENCE.mp3 -af silencedetect=n=-50dB:d=1 -y -ab 192k  SILENCE_OUT.mp3

However, the following output is produced:

ffmpeg version N-66154-g1654ca7 Copyright (c) 2000-2014 the FFmpeg developers
  built on Sep  5 2014 22:10:38 with gcc 4.8.3 (GCC)
  configuration: --enable-gpl --enable-version3 --disable-w32threads --enable-av
isynth --enable-bzlib --enable-fontconfig --enable-frei0r --enable-gnutls --enab
le-iconv --enable-libass --enable-libbluray --enable-libbs2b --enable-libcaca --
enable-libfreetype --enable-libgme --enable-libgsm --enable-libilbc --enable-lib
modplug --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrw
b --enable-libopenjpeg --enable-libopus --enable-librtmp --enable-libschroedinge
r --enable-libsoxr --enable-libspeex --enable-libtheora --enable-libtwolame --en
able-libvidstab --enable-libvo-aacenc --enable-libvo-amrwbenc --enable-libvorbis
 --enable-libvpx --enable-libwavpack --enable-libwebp --enable-libx264 --enable-
libx265 --enable-libxavs --enable-libxvid --enable-decklink --enable-zlib
  libavutil      54.  7.100 / 54.  7.100
  libavcodec     56.  1.100 / 56.  1.100
  libavformat    56.  4.100 / 56.  4.100
  libavdevice    56.  0.100 / 56.  0.100
  libavfilter     5.  1.100 /  5.  1.100
  libswscale      3.  0.100 /  3.  0.100
  libswresample   1.  1.100 /  1.  1.100
  libpostproc    53.  0.100 / 53.  0.100
Input #0, mp3, from 'SILENCE.mp3':
  Metadata:
    title           : Snowblind (Featuring Tasha Baxter)
    artist          : Au5
    album           : Snowblind (Featuring Tasha Baxter)
    genre           : Electronica
    performer       : Au5
    track           : 1/1
    date            : 2014
    album_artist    : Au5,Tasha Baxter
    major_brand     : mp42
    minor_version   : 0
    compatible_brands: isommp42
    encoder         : Lavf55.42.100
  Duration: 00:05:50.80, start: 0.025057, bitrate: 192 kb/s
    Stream #0:0: Audio: mp3, 44100 Hz, stereo, s16p, 192 kb/s
Output #0, mp3, to 'SILENCE_OUT.mp3':
  Metadata:
    TIT2            : Snowblind (Featuring Tasha Baxter)
    TPE1            : Au5
    TALB            : Snowblind (Featuring Tasha Baxter)
    TCON            : Electronica
    TPE3            : Au5
    TRCK            : 1/1
    TDRL            : 2014
    TPE2            : Au5,Tasha Baxter
    major_brand     : mp42
    minor_version   : 0
    compatible_brands: isommp42
    TSSE            : Lavf56.4.100
    Stream #0:0: Audio: mp3 (libmp3lame), 44100 Hz, stereo, s16p, 192 kb/s
    Metadata:
      encoder         : Lavc56.1.100 libmp3lame
Stream mapping:
  Stream #0:0 -> #0:0 (mp3 (native) -> mp3 (libmp3lame))
Press [q] to stop, [?] for help
[silencedetect @ 0000000004398f40] silence_start: -0.00628118
[silencedetect @ 0000000004398f40] silence_end: 3.21413 | silence_duration: 3.22
041
[silencedetect @ 0000000004398f40] silence_start: 343.844
[libmp3lame @ 00000000043b2940] Trying to remove 1152 samples, but the queue is
empty
size=    8223kB time=00:05:50.79 bitrate= 192.0kbits/s
video:0kB audio:8222kB subtitle:0kB other streams:0kB global headers:0kB muxing
overhead: 0.011485%

The generated audio file however still has the original length without any silence removed. See the following images:

Any help is appreciated!

EDIT: Alright, silence detect is only DETECTING the silence. Not removing it. I will try to post a solution for this.

Solution 1

Use the silenceremove filter. This removes silence from the audio track only - it will leave the video unedited, i.e., things will go out of sync

Its arguments are a little cryptic.

An example

ffmpeg -i input.mp3 -af silenceremove=1:0:-50dB output.mp3

This removes silence

• at the beginning (indicated by the first argument 1)
• with minimum length zero (indicated by the second argument 0)
• silence is classified as anything under -50 decibels (indicated by -50dB).

Documentation: FFMPEG silence remove filter

Also anyone looking to find the right value to classify silence as may wish to look into normalising their input audio volume to 0dB first, to do this in ffmpeg see this answer.

Edit

As pointed out by @mems, to detect whether your version of ffmpeg has the filter run

ffmpeg -hide_banner -filters | grep silenceremove

if you have the filter it'll output something like

silenceremove A->A Remove silence

Solution 2

ffmpeg silence detect only detects the silence. One has to scan the ffmpeg output and cut the mp3 file.

In theory, this would be done as:

ffmpeg -i INPUT.mp3 -af silencedetect=n=-50dB:d=1

and monitoring for output in form of:

[silencedetect @ 0000000004970f80] silence_start: -0.00154195
[silencedetect @ 0000000004970f80] silence_end: 3.20435 | silence_duration: 3.2059
...
[silencedetect @ 0000000004970f80] silence_start: 343.84

And, cutting start and end silence:

ffmpeg -i INPUT.mp3 -ss 3.20435 -t (343.84-3.20435)

I ended up writing a small Java program which does it. Hints:

• ffmpeg writes to stderr. This means, you need to use ProcessBuilder and redirectErrorStream(true).
• secondly, you need to extract the silence_start and silence_end information.
• then you might use the timestamps to cut the video

Following code may be helpful: Using Java and FFMPEG with silencedetect to remove audio silence

Solution 3

After reading the FFmpeg silenceremove documentation, this is how you remove silence at the beginning and end of an audio file (keeps silence in the middle).

ffmpeg -i "INPUT.mp3" -af silenceremove=start_periods=1:stop_periods=1:detection=peak "OUTPUT.mp3"