Recherche avancée

Sur d’autres sites (4431)

• How can I increase the performance of this video watermarking call which seems to be only using 6% of my system's CPU ? Currently using ffmpeg

  24 août 2023, par Codemonkey

  I have a php script to add a watermark to an uploaded video, using ffmpeg

  


  shell_exec("ffmpeg -y -i input.mp4 -i watermark.png -filter_complex $filter output.mp4")


  


  The video is 36 seconds long, 33MB and 1080x1920, 24fps. The output video comes in at 8MB, scaled down to 540x960 with a PNG overlaid as a watermark.

  


  The script takes about 3 seconds, averaging 320fps or so. For a one-off this would be ok, but I'm going to be processing hundreds/thousands of these at a time.

  


  I have 128GB of ram and an EPYC 7502p (32cores/64threads), and both input and output files are on NVMe drives.

  


  If I monitor top showing all my cores/threads in another window, with refresh delay set to a variety of things from 0.1 to 2 seconds, I'd guesstimate that this hits about 1/4 of my available threads, at about 25% utilisation each.

  


  So it kind of feels like I COULD see a 16x speed increase if it hit all of them at 100%. That actually wouldn't be ideal, locking up the server of course, but when my CPU is sat at 99% idle 99% of the time, I feel it could do better - it'd be nice to see it hitting at least half the cores or utilising at least 50% of them when it does...

  


  So, is there anything I can change with ffmpeg's config/build or the way I call it ? If this is as good as ffmpeg can do, are there better tools for the job ?

  


  Thank you

  


  ---

  


  EDIT :

  


  A few things I've learnt since posting :

  


  

  • -c:v libx264 -preset ultrafast - marginal help

  


  • nice -n -10 - marginal help (changing to -20 is makes negligible difference)

  


  • -c:a copy - significant help, which I found surprising

  


  


  FPS up from 320 to 540 using those 3 combined. I still feel there should be significantly more gains somewhere though...

  


  ---

  


  Here's the output from the original script, if it helps :

  


  ffmpeg version 4.2.9 Copyright (c) 2000-2023 the FFmpeg developers
  built with gcc 8 (GCC)
  configuration: --prefix=/usr --bindir=/usr/bin --datadir=/usr/share/ffmpeg --docdir=/usr/share/doc/ffmpeg --incdir=/usr/include/ffmpeg --libdir=/usr/lib64 --mandir=/usr/share/man --arch=x86_64 --optflags='-O2 -g -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -fexceptions -fstack-protector-strong -grecord-gcc-switches -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -mtune=generic -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection' --extra-ldflags='-Wl,-z,relro -Wl,-z,now -specs=/usr/lib/rpm/redhat/redhat-hardened-ld ' --extra-cflags=' ' --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libvo-amrwbenc --enable-version3 --enable-bzlib --disable-crystalhd --enable-fontconfig --enable-frei0r --enable-gcrypt --enable-gnutls --enable-ladspa --enable-libaom --enable-libdav1d --enable-libass --enable-libbluray --enable-libcdio --enable-libdrm --enable-libjack --enable-libfreetype --enable-libfribidi --enable-libgsm --enable-libmp3lame --enable-nvenc --enable-openal --enable-opencl --enable-opengl --enable-libopenjpeg --enable-libopus --enable-libpulse --enable-librsvg --enable-libsrt --enable-libsoxr --enable-libspeex --enable-libssh --enable-libtheora --enable-libvorbis --enable-libv4l2 --enable-libvidstab --enable-libvmaf --enable-version3 --enable-vapoursynth --enable-libvpx --enable-libx264 --enable-libx265 --enable-libxvid --enable-libzimg --enable-libzvbi --enable-avfilter --enable-avresample --enable-libmodplug --enable-postproc --enable-pthreads --disable-static --enable-shared --enable-gpl --disable-debug --disable-stripping --shlibdir=/usr/lib64 --enable-libmfx --enable-runtime-cpudetect
  libavutil      56. 31.100 / 56. 31.100
  libavcodec     58. 54.100 / 58. 54.100
  libavformat    58. 29.100 / 58. 29.100
  libavdevice    58.  8.100 / 58.  8.100
  libavfilter     7. 57.100 /  7. 57.100
  libavresample   4.  0.  0 /  4.  0.  0
  libswscale      5.  5.100 /  5.  5.100
  libswresample   3.  5.100 /  3.  5.100
  libpostproc    55.  5.100 / 55.  5.100
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from 'input.mp4':
  Metadata:
    major_brand     : M4V
    minor_version   : 1
    compatible_brands: isomavc1mp42
    creation_time   : 2023-07-03T14:55:37.000000Z
  Duration: 00:00:36.50, start: 0.000000, bitrate: 7315 kb/s
    Stream #0:0(und): Video: h264 (High) (avc1 / 0x31637661), yuv420p(tv, bt709/bt709/unknown), 1080x1920 [SAR 1:1 DAR 9:16], 7213 kb/s, 24 fps, 24 tbr, 24k tbn, 48 tbc (default)
    Metadata:
      creation_time   : 2023-07-03T14:55:37.000000Z
      handler_name    : ETI ISO Video Media Handler
      encoder         : Elemental H.264
    Stream #0:1(und): Audio: aac (LC) (mp4a / 0x6134706D), 48000 Hz, stereo, fltp, 96 kb/s (default)
    Metadata:
      creation_time   : 2023-07-03T14:55:37.000000Z
      handler_name    : ETI ISO Audio Media Handler
Input #1, png_pipe, from 'watermark.png':
  Duration: N/A, bitrate: N/A
    Stream #1:0: Video: png, rgba(pc), 1232x804 [SAR 11811:11811 DAR 308:201], 25 tbr, 25 tbn, 25 tbc
Stream mapping:
  Stream #0:0 (h264) -> scale (graph 0)
  Stream #1:0 (png) -> scale (graph 0)
  overlay (graph 0) -> Stream #0:0 (libx264)
  Stream #0:1 -> #0:1 (aac (native) -> aac (native))
Press [q] to stop, [?] for help
[libx264 @ 0x56510d488e00] using SAR=1/1
[libx264 @ 0x56510d488e00] using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2 AVX FMA3 BMI2 AVX2
[libx264 @ 0x56510d488e00] profile High, level 3.1, 4:2:0, 8-bit
[libx264 @ 0x56510d488e00] 264 - core 157 r2980 34c06d1 - H.264/MPEG-4 AVC codec - Copyleft 2003-2019 - 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=30 lookahead_threads=5 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=24 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:
    major_brand     : M4V
    minor_version   : 1
    compatible_brands: isomavc1mp42
    encoder         : Lavf58.29.100
    Stream #0:0: Video: h264 (libx264) (avc1 / 0x31637661), yuv420p, 540x960 [SAR 1:1 DAR 9:16], q=-1--1, 24 fps, 12288 tbn, 24 tbc (default)
    Metadata:
      encoder         : Lavc58.54.100 libx264
    Side data:
      cpb: bitrate max/min/avg: 0/0/0 buffer size: 0 vbv_delay: -1
    Stream #0:1(und): Audio: aac (LC) (mp4a / 0x6134706D), 48000 Hz, stereo, fltp, 128 kb/s (default)
    Metadata:
      creation_time   : 2023-07-03T14:55:37.000000Z
      handler_name    : ETI ISO Audio Media Handler
      encoder         : Lavc58.54.100 aac
frame=  876 fps=324 q=-1.0 Lsize=    7768kB time=00:00:36.45 bitrate=1745.5kbits/s speed=13.5x
video:7177kB audio:566kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.336181%
[libx264 @ 0x56510d488e00] frame I:9     Avg QP:21.46  size: 37765
[libx264 @ 0x56510d488e00] frame P:237   Avg QP:23.78  size: 17862
[libx264 @ 0x56510d488e00] frame B:630   Avg QP:28.03  size:  4405
[libx264 @ 0x56510d488e00] consecutive B-frames:  2.9%  2.5%  3.8% 90.9%
[libx264 @ 0x56510d488e00] mb I  I16..4: 13.6% 65.9% 20.5%
[libx264 @ 0x56510d488e00] mb P  I16..4:  5.2% 16.6%  4.1%  P16..4: 32.8% 18.0% 12.5%  0.0%  0.0%    skip:10.8%
[libx264 @ 0x56510d488e00] mb B  I16..4:  1.1%  2.7%  0.2%  B16..8: 35.3%  7.2%  2.5%  direct: 3.3%  skip:47.7%  L0:37.0% L1:52.2% BI:10.8%
[libx264 @ 0x56510d488e00] 8x8 transform intra:65.0% inter:71.5%
[libx264 @ 0x56510d488e00] coded y,uvDC,uvAC intra: 54.7% 62.1% 11.5% inter: 17.9% 12.9% 0.5%
[libx264 @ 0x56510d488e00] i16 v,h,dc,p: 16% 28%  8% 47%
[libx264 @ 0x56510d488e00] i8 v,h,dc,ddl,ddr,vr,hd,vl,hu: 21% 23% 18%  6%  6%  7%  7%  7%  7%
[libx264 @ 0x56510d488e00] i4 v,h,dc,ddl,ddr,vr,hd,vl,hu: 23% 16% 11%  7%  9% 10%  8%  8%  8%
[libx264 @ 0x56510d488e00] i8c dc,h,v,p: 55% 20% 20%  5%
[libx264 @ 0x56510d488e00] Weighted P-Frames: Y:5.5% UV:3.4%
[libx264 @ 0x56510d488e00] ref P L0: 65.0% 18.8% 13.0%  3.0%  0.2%
[libx264 @ 0x56510d488e00] ref B L0: 95.0%  4.0%  1.1%
[libx264 @ 0x56510d488e00] ref B L1: 98.2%  1.8%
[libx264 @ 0x56510d488e00] kb/s:1610.57
[aac @ 0x56510d4865c0] Qavg: 22331.562


  


• Why the result of FFmpeg capture process has no audio to create a webm file ?

  23 mars 2019, par RAM

  The result of my bellow FFmpeg command has no audio (is silent) :

  ffmpeg -f gdigrab -framerate 30 -i desktop -video_size 720x480 -c:v libvpx-vp9 -c:a libopus -b:v 1M -b:a 128K -auto-alt-ref 0 -crf 10 -preset ultrafast output.webm

  But this one has audio :

  ffmpeg -f gdigrab -i desktop -f dshow -i audio="Microphone (4- High Definition Audio Device)" output.mkv

  • How should I capture as webm file by using libopus or libvorbis ?
  • What is the problem in my first command ?

  ---

  My FFmpeg version :

  ffmpeg version N-93439-gb073fb9eea Copyright (c) 2000-2019 the FFmpeg developers
  
    built with gcc 8.2.1 (GCC) 20190212
  
    configuration: --enable-gpl --enable-version3 --enable-sdl2 --enable-fontconfig --enable-gnutls --enable-iconv --enable-libass --enable-libdav1d --enable-libbluray --enable-libfreetype --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libopenjpeg 
  
                   --enable-libopus --enable-libshine --enable-libsnappy --enable-libsoxr --enable-libtheora --enable-libtwolame --enable-libvpx --enable-libwavpack --enable-libwebp --enable-libx264 --enable-libx265 --enable-libxml2 --enable-libzimg --enable-lzma --enable-zlib --enable-gmp --enable-libvidstab 
  
                   --enable-libvorbis --enable-libvo-amrwbenc --enable-libmysofa --enable-libspeex --enable-libxvid --enable-libaom --enable-libmfx --enable-amf --enable-ffnvcodec --enable-cuvid --enable-d3d11va --enable-nvenc --enable-nvdec --enable-dxva2 --enable-avisynth --enable-libopenmpt
  
    libavutil      56. 26.100 / 56. 26.100
  
    libavcodec     58. 47.105 / 58. 47.105
  
    libavformat    58. 26.101 / 58. 26.101
  
    libavdevice    58.  7.100 / 58.  7.100
  
    libavfilter     7. 48.100 /  7. 48.100
  
    libswscale      5.  4.100 /  5.  4.100
  
    libswresample   3.  4.100 /  3.  4.100
  
    libpostproc    55.  4.100 / 55.  4.100

• ffmpeg clean all noise background silences in a poscast

  23 mars 2019, par fireDevelop.com

  I have hundreds of podcast without music, just the voice and the room silence.
  In the silences, I have many clicks, respirations, etc...
  I need to clean all silences with a script, keeping intact the voice.

  In this picture you can see my dirty silences

  And here the result I want in all my audios

  When I use some scripts of sox. I don`t get the result I spect because the voice is affected by the script, the room-silence disappear and some clic still in the silences.

  Then in order to keep intact the voice, I want to do this :

  1. Delete all the silences longer than 3 seconds.

  ---

  2. Split all the audio and silences with in a sequence numbers. ie. :

     • 001-Silence-2.0seconds.wav
     • 002-voice.wav
     • 003-Silence-0.25seconds.wav
     • 004-voice.wav
     • 005-Silence-0.75seconds.wav
     • 006-voice.wav
     • ...
     • ...

  ---

  3. Before, run the script I created manually many files with silences of diferents silences I will use :

     • myManuallySilence-0.25seconds.wav
     • myManuallySilence-0.50seconds.wav
     • myManuallySilence-0.75seconds.wav
     • myManuallySilence-0.1seconds.wav
     • myManuallySilence-1.25seconds.wav
     • ...
     • ...
     • myManuallySilence-2.50seconds.wav
     • myManuallySilence-2.75seconds.wav
     • myManuallySilence-3.0seconds.wav

  ---

  4. the script will check the dirty silences duration and replace by the files myManuallySilence-x.xseconds.wav

  ---

  5. merge all files in one wav file, with the original voice and all the silences cleanned.

  ---

  At the moment I have only this script :

  # get the path of Adobe Audition and add timestamp in the output 
  
  filename
  
  fileName=out
  
  current_time=$(date "+%Y.%m.%d-%H.%M.%S")
  
  newFileName=$fileName.$current_time.wav
  
  #yourPathAPP=/Applications/Adobe\ Audition\ CC\ 2019/Adobe\ Audition\ 
  
  CC\ 2019.app
  
  yourPathAPP=/Volumes/6TB/Applications/ocenaudio.app
  
  # # First denoise audio
  
  
  
  # ## Get noise sample
  
   ffmpeg -i in.wav -vn -ss 00:00:00 -t 00:00:01 noise-sample.wav
  
  
  
  # ## Create noise profile
  
   sox noise-sample.wav -n noiseprof noise.prof
  
  
  
  # ## Clean audio from noise
  
   sox in.wav $newFileName noisered noise.prof 0.50
  
  # # Split audio by noise
  
   sox -V3 $newFileName output.wav silence 1 00:00:02.000 - 80d 1 
  
   00:00:02.000 -80d : newfile : restart
  
  
  
  # ####### (these settings worked for my computer mic - maybe we need to 
  
  finetune them later) #######

  Is getting all the voice in separate files like this :
  output001.wav
  output002.wav
  output003.wav
  output004.wav
  ...
  output00x.wav

  Please, any suggestion will be appreciated.
  Thanks so much in advance !