shortest command doesnt work when using a filter_complex with audio and video filters

[phingers]

Summary of the bug: shortest command doesn't work when using a filter_complex with audio and video filters

I uploaded output.mp4 at 8:21 ET on 7/20/14 to upload.ffmpeg.org

The background audio is 4 mins long, but it should be truncated to the video length.

Shortest works if I take the drawtext off the filter_complex

How to reproduce:

ffmpeg -framerate 1/5 -i foo%03d.png -i audio.wav -i speak.wav -filter_complex '[1:a]volume=volume=0.2[aquiet];[aquiet][2:a]amix=inputs=2[a];[0:v]drawtext=fontsize=40:fontcolor=white@0.9:fontfile=FreeSerif.ttf:shadowy=2:shadowx=2:text="I am getting the hang of ffmpeg\!":x=(w-text_w)/2:y=(h-text_h-line_h)/2[v]' -map '[a]' -map '[v]' -shortest -r 32 -c:v libx264 -b:a 192k -pix_fmt yuv420p output.mp4
ffmpeg version 2.2.4-   http://johnvansickle.com/ffmpeg/    Copyright (c) 2000-2014 the FFmpeg developers
  built on Jul  1 2014 22:10:48 with gcc 4.8 (Debian 4.8.3-4)
  configuration: --enable-gpl --enable-version3 --disable-shared --disable-debug --enable-runtime-cpudetect --enable-libmp3lame --enable-libx264 --enable-libwebp --enable-libspeex --enable-libvorbis --enable-libvpx --enable-libfreetype --enable-fontconfig --enable-libxvid --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libtheora --enable-libvo-aacenc --enable-libvo-amrwbenc --enable-gray --enable-libopenjpeg --enable-libopus --disable-ffserver --enable-libass --enable-gnutls --cc=gcc-4.8
  libavutil      52. 66.100 / 52. 66.100
  libavcodec     55. 52.102 / 55. 52.102
  libavformat    55. 33.100 / 55. 33.100
  libavdevice    55. 10.100 / 55. 10.100
  libavfilter     4.  2.100 /  4.  2.100
  libswscale      2.  5.102 /  2.  5.102
  libswresample   0. 18.100 /  0. 18.100
  libpostproc    52.  3.100 / 52.  3.100
Input #0, image2, from 'foo%03d.png':
  Duration: 00:00:30.00, start: 0.000000, bitrate: N/A
    Stream #0:0: Video: png, rgb24, 640x360 [SAR 2835:2835 DAR 16:9], 0.20 fps, 1 tbr, 0.20 tbn, 0.20 tbc
Input #1, mp3, from 'audio.wav':
  Metadata:
    major_brand     : dash
    minor_version   : 0
    compatible_brands: iso6mp41
    encoder         : Lavf54.29.104
  Duration: 00:04:29.09, start: 0.025057, bitrate: 238 kb/s
    Stream #1:0: Audio: mp3, 44100 Hz, stereo, s16p, 238 kb/s
Guessed Channel Layout for  Input Stream #2.0 : mono
Input #2, wav, from 'speak.wav':
  Duration: 00:00:03.33, bitrate: 256 kb/s
    Stream #2:0: Audio: pcm_s16le ([1][0][0][0] / 0x0001), 16000 Hz, mono, s16, 256 kb/s
File 'output.mp4' already exists. Overwrite ? [y/N] y
[libx264 @ 0x2fdf7a0] using SAR=1/1
[libx264 @ 0x2fdf7a0] using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2
[libx264 @ 0x2fdf7a0] profile High, level 3.0
[libx264 @ 0x2fdf7a0] 264 - core 142 r2 a5831aa - H.264/MPEG-4 AVC codec - Copyleft 2003-2014 - http://www.videolan.org/x264.html - options: cabac=1 ref=3 deblock=1:0:0 analyse=0x3:0x113 me=hex subme=7 psy=1 psy_rd=1.00:0.00 mixed_ref=1 me_range=16 chroma_me=1 trellis=1 8x8dct=1 cqm=0 deadzone=21,11 fast_pskip=1 chroma_qp_offset=-2 threads=18 lookahead_threads=2 sliced_threads=0 nr=0 decimate=1 interlaced=0 bluray_compat=0 constrained_intra=0 bframes=3 b_pyramid=2 b_adapt=1 b_bias=0 direct=1 weightb=1 open_gop=0 weightp=2 keyint=250 keyint_min=25 scenecut=40 intra_refresh=0 rc_lookahead=40 rc=crf mbtree=1 crf=23.0 qcomp=0.60 qpmin=0 qpmax=69 qpstep=4 ip_ratio=1.40 aq=1:1.00
Output #0, mp4, to 'output.mp4':
  Metadata:
    encoder         : Lavf55.33.100
    Stream #0:0: Audio: aac (libvo_aacenc) ([64][0][0][0] / 0x0040), 44100 Hz, stereo, s16, 192 kb/s (default)
    Stream #0:1: Video: h264 (libx264) ([33][0][0][0] / 0x0021), yuv420p, 640x360 [SAR 1:1 DAR 16:9], q=-1--1, 16384 tbn, 32 tbc (default)
Stream mapping:
  Stream #0:0 (png) -> drawtext
  Stream #1:0 (mp3) -> volume
  Stream #2:0 (pcm_s16le) -> amix:input1
  amix -> Stream #0:0 (libvo_aacenc)
  drawtext -> Stream #0:1 (libx264)
Press [q] to stop, [?] for help
frame=  800 fps=116 q=-1.0 Lsize=    6606kB time=00:04:29.05 bitrate= 201.1kbits/s dup=794 drop=0
video:238kB audio:6307kB subtitle:0 data:0 global headers:0kB muxing overhead 0.933773%
[libvo_aacenc @ 0x2fde340] 1 frames left in the queue on closing
[libx264 @ 0x2fdf7a0] frame I:5     Avg QP:17.56  size: 37668
[libx264 @ 0x2fdf7a0] frame P:201   Avg QP:17.22  size:   219
[libx264 @ 0x2fdf7a0] frame B:594   Avg QP:25.96  size:    18
[libx264 @ 0x2fdf7a0] consecutive B-frames:  0.8%  0.2%  1.5% 97.5%
[libx264 @ 0x2fdf7a0] mb I  I16..4:  4.7% 64.5% 30.8%
[libx264 @ 0x2fdf7a0] mb P  I16..4:  0.1%  0.2%  0.3%  P16..4:  0.5%  0.0%  0.0%  0.0%  0.0%    skip:99.0%
[libx264 @ 0x2fdf7a0] mb B  I16..4:  0.0%  0.0%  0.0%  B16..8:  0.1%  0.0%  0.0%  direct: 0.0%  skip:99.9%  L0:16.4% L1:83.6% BI: 0.0%
[libx264 @ 0x2fdf7a0] 8x8 transform intra:58.6% inter:30.7%
[libx264 @ 0x2fdf7a0] coded y,uvDC,uvAC intra: 83.7% 87.7% 75.6% inter: 0.0% 0.1% 0.0%
[libx264 @ 0x2fdf7a0] i16 v,h,dc,p: 50% 11%  7% 32%
[libx264 @ 0x2fdf7a0] i8 v,h,dc,ddl,ddr,vr,hd,vl,hu: 15% 19% 21%  6%  6%  5%  8%  6% 13%
[libx264 @ 0x2fdf7a0] i4 v,h,dc,ddl,ddr,vr,hd,vl,hu: 24% 22% 11%  6%  7%  6%  7%  5% 12%
[libx264 @ 0x2fdf7a0] i8c dc,h,v,p: 50% 22% 20%  8%
[libx264 @ 0x2fdf7a0] Weighted P-Frames: Y:0.0% UV:0.0%
[libx264 @ 0x2fdf7a0] ref P L0: 94.2%  0.5%  3.9%  1.4%
[libx264 @ 0x2fdf7a0] ref B L0: 53.1% 45.3%  1.6%
[libx264 @ 0x2fdf7a0] ref B L1: 95.1%  4.9%
[libx264 @ 0x2fdf7a0] kb/s:77.80

[cehoyos]

Is this a regression?

[cehoyos]

Apparently not important.

[c-14]

Replying to cehoyos:

Is this a regression?

It doesn't look like a regression to me. I checked it with several of the release versions including the 2.2 branch, the 2.1 branch, the 2.0 branch, the 1.2 branch and the 1.1 branch. Same results with each version.

I've also managed to simplify the command by a bit.
This is the simplest version of the command I've found so far that reproduces the issue:

ffmpeg -i file -filter_complex '[0:a]anull[a];[0:v]trim=duration=25[v]' -map '[v]' -map '[a]' -shortest output

Note that it does not matter if the audio stream is shorter or if the video stream is shorter, -shortest does not work in either case. -filter_complex - on the other hand - is required. Using:

ffmpeg -i file -af anull -vf trim=duration=25 -shortest output

does not trigger the bug.
Input file format and output file format do not seem to influence the issue nor does the codec used.

[Li0n5]

Confirmed, got this bug too. It just encodes infinitely. Could we have some attention to it?

[gabrielstuff]

I can confirm the same error. I'm not using audio, but an overlay option and this is the same issue. File keeps on growing and growing.

[asavah]

Confirmed here too.
Can we finally get some dev love on this bug, please?

[jb_alvarado]

Current version has still the problem. This function is in combination with audio filter apad very useful, but sadly this bug exists.

[mrskman]

I had the same issue. Workaround is using separate -filter_complex argument for every output stream:

ffmpeg -i AUDIO -i VIDEO \
 -filter_complex '[0:a]apad[audio]' \
 -filter_complex '[0:v]scale=-2:720[video]' \
 -map '[audio]' -map '[video]' \
 -shortest \
 OUTPUT

[vstavrinov]

Replying to mrskman:

I had the same issue. Workaround is using separate -filter_complex argument for every output stream:

Thank You a lot. One clarification: in my case where the video is endless, the video filter should be placed first, i.e. before audio.