Recherche avancée

Sur d’autres sites (4846)

• Changing bit-rate with timestamp copy still offsets events by 1 or 2 frames

  10 avril 2018, par harkmug

  As a novice user, I am trying to understand the following. I have an mp4 (encoder=Lavf57.66.104) that is 127 MB in size and I use the following to reduce its size :

  ffmpeg -i original.mp4 -start_at_zero -copyts -b:v 1000k -c:a copy output.mp4

  The duration and number of frames stay the same after this process, however, when I annotate (using ELAN) the same events (at millisecond level, e.g. a blink) the output video (encoder=Lavf57.55.100) seems to be offset relative to the original by 1 or 2 frames.

  Can someone help me understand this shift ? Thanks !

  UPDATE (2018-04-10) :
  As per @Mulvya’s suggestion, ran :

  ffmpeg -i original.mp4 -copyts -b:v 1000k -c:a copy output.mp4

  Looking at the two files :

  fprobe -v error -select_streams v:0 -show_frames -show_entries frame=key_frame,pkt_pts_time,pict_type,coded_picture_number -of default=noprint_wrappers=1:nokey=1 -of csv=p=0 original.mp4 | head -n 5
  
  
  
  1,5292.861000,I,0
  
  0,5292.894333,P,1
  
  0,5292.927667,P,2
  
  0,5292.961000,P,3
  
  0,5292.994333,P,4

  Same for output :

  1,5292.866016,I,0
  
  0,5292.899349,P,1
  
  0,5292.932682,P,2
  
  0,5292.966016,P,3
  
  0,5292.999349,P,4

  Trying to understand how to get the same time-stamps for the same frames. Maybe this is not possible ?

• Multiple overlays using ffmpeg

  23 mars 2018, par lhan

  I’m trying to satisfy a few layering scenarios for building video files using ffmpeg.

  Scenario 1 : Overlay a video (specifying opacity of the video) on top of an image, creating a new video as the result.

  I solved this with :

  ffmpeg -i video.mp4 -i image.jpg -filter_complex '[0]format=rgba,colorchannelmixer=aa=0.7,scale=w=3840:h=2160[a];[1][a]overlay=0:0' -t 30 output.mp4

  I’m scaling the video to 3840x2160 to match my image (ideally I’d have them matching beforehand).

  Scenario 2 : 3 layers now, video - image - image. The middle image layer is a transparent image with text. So we have a base image, with text overlaid, and a video on top of that at a certain opacity.

  I solved this with :

  ffmpeg -i video.mp4 -i image.jpg -i text.png -filter_complex '[0]format=rgba,colorchannelmixer=aa=0.7,scale=w=3840:h=2160[a];[2][a]overlay=0:0,scale=w=3840:h=2160[b];[1][b]overlay=0:0' -t 30 output.mp4

  Scenario 3 (which I can’t get working) : Same as Scenario #2, but with text on top of the video.

  I tried re-arranging my filter, hoping to affect the layering order :

  ffmpeg -i video.mp4 -i image.jpg -i text.png -filter_complex '[2]overlay=0:0,scale=w=3840:h=2160[a];[0][a]format=rgba,colorchannelmixer=aa=0.7,scale=w=3840:h=2160[b];[1][b]overlay=0:0' -t 5 output.mp4

  But that gives the following error :

  Too many inputs specified for the "format" filter. Error initializing complex filters. Invalid argument

  Full Error :

  Input #0, mov,mp4,m4a,3gp,3g2,mj2, from ’video.mp4’ :
  Metadata :
  major_brand : mp42
  minor_version : 0
  compatible_brands : mp42mp41
  creation_time : 2018-03-09T20:52:18.000000Z

  Duration : 00:00:30.00, start : 0.000000, bitrate : 8002 kb/s

  Stream #0:0(eng) : Video : h264 (High) (avc1 / 0x31637661), yuv420p(tv, bt709), 1920x1080 [SAR 1:1 DAR 16:9], 7997 kb/s, 24 fps,
  24 tbr, 24k tbn, 48 tbc (default)

  Metadata :
  creation_time : 2018-03-09T20:52:18.000000Z
  handler_name : Alias Data Handler
  encoder : AVC Coding Input #1, image2, from ’image.jpg’ :

  Duration : 00:00:00.04, start : 0.000000, bitrate : 526829 kb/s

  Stream #1:0 : Video : mjpeg, yuvj444p(pc, bt470bg/unknown/unknown), 3840x2160 [SAR 96:96 DAR 16:9], 25 tbr, 25 tbn, 25 tbc Input #2,
  png_pipe, from ’text.png’ : Duration : N/A, bitrate : N/A

  Stream #2:0 : Video : png, rgba(pc), 1500x1500, 25 tbr, 25 tbn, 25 tbc [AVFilterGraph @ 0x7fc37d402de0]

  Too many inputs specified for the "format" filter. Error initializing complex filters. Invalid argument

  I can sort of get around that by tweaking the command so that the text isn’t an input to the overlay :

  ffmpeg -i lightTexture.mp4 -i image.jpg -i textSample.png -filter_complex '[2]overlay=0:0,scale=w=3840:h=2160;[0]format=rgba,colorchannelmixer=aa=0.7,scale=w=3840:h=2160[b];[1][b]overlay=0:0' -t 5 output_text_on_top.mp4

  But then my output video is all messed up. I suspect I am on the wrong track by trying to cram all of this into the -filter_complex. I’m wondering if I need to create two overlays and then overlay those (i.e overlay Text onto the Video, and then overlay that onto the base image) though I’m not sure how to accomplish that.

  If anyone could point me in the right direction here, I’d be super grateful.

• Transparent PNG to WEBM

  5 avril 2018, par Dan

  I am trying to make a partially transparent video for a Unity application, however, I can’t seem to get my commands to retain the alpha of the PNGs. Instead, it just makes the transparent parts white.

  Currently the command I am using is

  ffmpeg -r 30 -f image2 -s 2048x2048 -i frames%04d.png -vcodec vp8 -crf 63 -b:v 0 -pix_fmt yuva420p -auto-alt-ref 0 vid/output.webm

  Whilst I am aware I could try the VP9 codec, AFAIK, this is not supported in Unity 2017.2, so I have to stick to VP8

  The output this gives is

  D:\development\batch - Copy>ffmpeg -r 30 -f image2 -s 2048x2048 -i frames%04d.png -vcodec vp8 -crf 63 -b:v 0 -pix_fmt yuva420p -auto-alt-ref 0 vid/output.webm
  
  ffmpeg version N-90315-gf706cdda56 Copyright (c) 2000-2018 the FFmpeg developers
  
    built with gcc 7.3.0 (GCC)
  
    configuration: --enable-gpl --enable-version3 --enable-sdl2 --enable-bzlib --enable-fontconfig --enable-gnutls --enable-iconv --enable-libass --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-libmfx --enable-amf --enable-ffnvcodec --enable-cuvid --enable-d3d11va --enable-nvenc --enable-nvdec --enable-dxva2 --enable-avisynth
  
    libavutil      56.  9.100 / 56.  9.100
  
    libavcodec     58. 14.100 / 58. 14.100
  
    libavformat    58. 10.100 / 58. 10.100
  
    libavdevice    58.  2.100 / 58.  2.100
  
    libavfilter     7. 13.100 /  7. 13.100
  
    libswscale      5.  0.102 /  5.  0.102
  
    libswresample   3.  0.101 /  3.  0.101
  
    libpostproc    55.  0.100 / 55.  0.100
  
  Input #0, image2, from 'frames%04d.png':
  
    Duration: 00:00:03.50, start: 0.000000, bitrate: N/A
  
      Stream #0:0: Video: png, pal8(pc), 2048x2048, 30 fps, 30 tbr, 30 tbn, 30 tbc
  
  Stream mapping:
  
    Stream #0:0 -> #0:0 (png (native) -> vp8 (libvpx))
  
  Press [q] to stop, [?] for help
  
  [libvpx @ 00000209d9a6e680] v1.7.0
  
  Output #0, webm, to 'vid/output.webm':
  
    Metadata:
  
      encoder         : Lavf58.10.100
  
      Stream #0:0: Video: vp8 (libvpx), yuva420p, 2048x2048, q=-1--1, 30 fps, 1k tbn, 30 tbc
  
      Metadata:
  
        encoder         : Lavc58.14.100 libvpx
  
      Side data:
  
        cpb: bitrate max/min/avg: 0/0/0 buffer size: 0 vbv_delay: -1
  
  frame=  105 fps=4.1 q=0.0 Lsize=     144kB time=00:00:03.46 bitrate= 339.6kbits/s speed=0.134x
  
  video:94kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 52.357399%