Recherche avancée

Sur d’autres sites (8738)

• FFMPEG Concat filter - Output video skips second file

  16 octobre 2018, par Rune Aspvik

  I am using the ffmpeg concat filter to join video files. I am aware the files must be exactly the same format for concat to work without encoding. However, the output of these two files are not seemless. The video will stop when it switches to second video file.

  Here is the output of the two video files.

  SKYbum:~ Aspvik$ ffmpeg -i /Users/Aspvik/Desktop/newIntro.mp4 -i /Users/Aspvik/Downloads/wetransfer-008789/Middle.mp4 
  
  ffmpeg version 4.0.2 Copyright (c) 2000-2018 the FFmpeg developers
  
    built with Apple LLVM version 9.1.0 (clang-902.0.39.2)
  
    configuration: --prefix=/usr/local/Cellar/ffmpeg/4.0.2 --enable-shared --enable-pthreads --enable-version3 --enable-hardcoded-tables --enable-avresample --cc=clang --host-cflags= --host-ldflags= --enable-gpl --enable-libmp3lame --enable-libx264 --enable-libxvid --enable-opencl --enable-videotoolbox --disable-lzma
  
    libavutil      56. 14.100 / 56. 14.100
  
    libavcodec     58. 18.100 / 58. 18.100
  
    libavformat    58. 12.100 / 58. 12.100
  
    libavdevice    58.  3.100 / 58.  3.100
  
    libavfilter     7. 16.100 /  7. 16.100
  
    libavresample   4.  0.  0 /  4.  0.  0
  
    libswscale      5.  1.100 /  5.  1.100
  
    libswresample   3.  1.100 /  3.  1.100
  
    libpostproc    55.  1.100 / 55.  1.100
  
  Input #0, mov,mp4,m4a,3gp,3g2,mj2, from '/Users/Aspvik/Desktop/newIntro.mp4':
  
    Metadata:
  
      major_brand     : isom
  
      minor_version   : 512
  
      compatible_brands: isomiso2avc1mp41
  
      encoder         : Lavf58.12.100
  
    Duration: 00:00:08.87, start: 0.000000, bitrate: 7160 kb/s
  
      Stream #0:0(eng): Video: h264 (Constrained Baseline) (avc1 / 0x31637661), yuv420p, 1280x720, 7159 kb/s, 30 fps, 30 tbr, 15360 tbn, 60 tbc (default)
  
      Metadata:
  
        handler_name    : VideoHandler
  
  Input #1, mov,mp4,m4a,3gp,3g2,mj2, from '/Users/Aspvik/Downloads/wetransfer-008789/Middle.mp4':
  
    Metadata:
  
      major_brand     : isom
  
      minor_version   : 512
  
      compatible_brands: isomiso2avc1mp41
  
      title           : Netcam Studio 1.6.1.0
  
      encoder         : Lavf57.71.100
  
    Duration: 00:00:19.47, start: 0.033008, bitrate: 676 kb/s
  
      Stream #1:0(und): Video: h264 (Constrained Baseline) (avc1 / 0x31637661), yuv420p, 1280x720, 674 kb/s, 30 fps, 30 tbr, 15360 tbn, 60 tbc (default)
  
      Metadata:
  
        handler_name    : VideoHandler

  The command I’m using for the concat is :

  ffmpeg -f concat -safe 0 -i /Users/Aspvik/Desktop/files.txt -codec copy -y /Users/Aspvik/Desktop/video.mp4

  Can anyone see what could be wrong ?

• Gstreamer convert and display video v4l2 - tee problems in rust

  27 mars 2023, par d3im

  I have USB grabber v4l2 source and I want to tee stream to autovideosink and x264enc to file (now as fake black hole)

  


  When I disable one or another branch it works but together Pipeline goes :

  


  Pipeline state changed from Null to Ready
Pipeline state changed from Ready to Paused


  


  and stays there never switches to Playing

  


  gst-launch-1.0 with similar functionality works well.

  


      gst::Element::link_many(&[&pw_video, &v_caps, &vid_queuey, &vid_tee]).unwrap();
    gst::Element::link_many(&[&vid_queue1, &autovideoconvert, &vid_queuex, &autovideosink]).unwrap();
    gst::Element::link_many(&[&vid_queue2, &autovideoconvert_x264, &vid_queue3, &x264, &vid_queue4, &fake]).unwrap();

    let tee_display_pad = vid_tee.request_pad_simple("src_10").unwrap();
    let vid_queue1_pad = vid_queue1.static_pad("sink").unwrap();

    tee_display_pad.link(&vid_queue1_pad).unwrap();

    let tee_convert_pad = vid_tee.request_pad_simple("src_20").unwrap();
    let vid_queue2_pad = vid_queue2.static_pad("sink").unwrap();

    tee_convert_pad.link(&vid_queue2_pad).unwrap();


  


  How can I use tee in rust properly to have playable pipeline with two branches ?

  


  Update : I read some posts about increasing queue size, so I tried for this and then all queues :

  


      let vid_queue1 = gst::ElementFactory::make("queue")
        .name("queue1")
        .property("max-size-buffers", 5000 as u32)
        .property("max-size-bytes", 1048576000 as u32)
        .property("max-size-time", 60000000000 as u64)
        .build()
        .expect("queue1");


  


  but it didn't help so I tried set zero latency :

  


      let x264 = gst::ElementFactory::make("x264enc")
        .name("x264")
        .property_from_str("speed-preset", "ultrafast")
        .property_from_str("pass", "qual")
        .property_from_str("tune", "zerolatency")
        .property("quantizer", 0 as u32)
        .property("threads", 8 as u32)
        .build()
        .expect("!x264");


  


  and it works now. But comparable gst-launch-1.0 settings didn't had such option - only queues sizes increased.

  


  Is there any other option than setting zerolatency ?

  


• ffmpeg burned-in subtitles render in the wrong font

  21 août 2024, par dv151

  Trying to burn in subtitles to a video in FFMPEG in GothamProBold font. No matter what I do it keeps reverting to Helvetica. From the console, I see that FFMPEG seems to load the font without error. Then switches over to font provider "coretext"

  


  [Parsed_subtitles_0 @ 0x7fed054048c0] Loading font file '/Projects/Fonts/GothaProBol.otf'
[Parsed_subtitles_0 @ 0x7fed054048c0] Using font provider coretext
[Parsed_subtitles_0 @ 0x7fed054048c0] fontselect: (GothaProBol.otf, 400, 0) -> /System/Library/Fonts/Helvetica.ttc, -1, Helvetica


  


  It seems like it has my font loaded, then loads what is likely a system default of Helvetica instead. My guess is that my chosen font isn't actually loading after all.

  


  FFMPEG command (called from python) is as follows :

  


  ffmpeg_cmd = ["ffmpeg", 
              "-i", self.source_video_uri, 
              "-y",
              "-c:v", "prores", "-profile:v", "1", 
              "-c:a", "pcm_s16be", 
              "-vf", f"subtitles={srt_uri}:fontsdir=/Projects/Fonts:force_style='Fontname=GothaProBol.otf'",
              f"{self.source_video_uri}_render.mov"]

subprocess.call(ffmpeg_cmd)


  


  Any ideas ?

  


  UPDATE : Found this setting in libass header file "ass.h" - which ffmpeg calls when using the subtitle filter. Don't know how to actually set this variable when ffmpeg calls libass, but here it is. Line 182 :

  


   * \brief Default Font provider to load fonts in libass' database
 *
 * NONE don't use any default font provider for font lookup
 * AUTODETECT use the first available font provider
 * CORETEXT force a CoreText based font provider (OS X only)
 * FONTCONFIG force a Fontconfig based font provider
 *
 * libass uses the best shaper available by default.
 */
typedef enum {
    ASS_FONTPROVIDER_NONE       = 0,
    ASS_FONTPROVIDER_AUTODETECT = 1,
    ASS_FONTPROVIDER_CORETEXT,
    ASS_FONTPROVIDER_FONTCONFIG,
    ASS_FONTPROVIDER_DIRECTWRITE,
} ASS_DefaultFontProvider;


  


  RE : ANSWER BELOW : For the most part, it seems that if your font is installed in /System/Fonts or /Library/Fonts then CoreText can find it. Though in some cases, the naming conventions can be quite particular and non-intuitive. It also can't seem to find all fonts, necessarily.

  


  For example : Gotham Pro Bold, in the /Library/Fonts folder on my system, file named "GothaProBol.otf" is correctly passed to fontname as : GothamPro-Bold or just Gotham Pro. Gotham Pro Bold, GothamPro, Gotham Pro-Bold, GothaProBol, and GothaProBol.otf do NOT work.

  


  For most fonts it seems the preferred convention is FontName-Style/Weight as displayed in Mac OS's FontBook, not the filename.

  


  That said, I have a novelty 'Game of Thrones.ttf' font in the same folder as Gotham Pro, and I can't get CoreText to connect to it under any of the above naming conventions.