Recherche avancée

Sur d’autres sites (15995)

• A 'clean' way to cut an MP4 movie into two sections using FFMPEG ?

  30 août 2020, par Peter in Japan

  I am attempting to use FFMPEG to make a script that can easily split a short MP4 movie with sound into two pieces at a certain point. I've searched through what feels like hundreds of posts to try to find "the" answer, but most of my attempts end up with poor results, broken video files that cause my video play to freeze, and so on. I am attempting to make a script that allows me to easily cut a short anime movie (something grabbed from Twitter or some other short source) and cut it into two sections, so it can be under the 2:20 Twitter time limit, or to cut out some scene I don't want to show my followers.

  


  The issue is that FFMPEG is good at cutting videos into segments, but bad at know where keyframes are, so most videos have two seconds of blank video at the front before some keyframe appears, which looks terrible.

  


  One example I found that works well is below, which cuts any mp4 into a bunch of chunks of n second size (six seconds in the example below). Source and documentation for this is https://moe.vg/3b8eNTs

  


  ffmpeg -i seitokai.mp4 -c:v libx264 -crf 22 -map 0 -segment_time 6 -reset_timestamps 1 -g 30 -sc_threshold 0 -force_key_frames "expr:gte(t,n_forced*1)" -f segment output%03d.mp4

  


  This code works great, at least allowing me to access the "left" side of a video that I want, in this case a six-second segment I want. Can anyone tell me how to accomplish the above, but starting at the 13-second period in said video, so I could get right "right" (not starting from 00:00:00) video cut cleanly ?

  


  Alternately, a single, unified and elegant way to split (re-encode) an MP4 into two segments that forces keyframes from the very beginning of the cut pieces would be wonderful. I can't believe how hard this seems to be.

  


  Thanks in advance for any help you can give ! Greetings from rural Japan !

  


• libswscale bad dst image pointers cgo

  28 août 2020, par SolskGaer

  I am trying to use libswscale to scale image before encoding to h264 using cgo. Here I wrote a simple demo(sorry for the bad code style, I just want to do quick verification) :

  


  func scale(img []byte, scaleFactor int) {
    input, _, _ := image.Decode(bytes.NewReader(img))
    if a, ok := input.(*image.YCbCr); ok {
        width, height := a.Rect.Dx(), a.Rect.Dy()
        var format C.enum_AVPixelFormat = C.AV_PIX_FMT_YUV420P
        context := C.sws_getContext(C.int(width), C.int(height), format, C.int(width/scaleFactor), C.int(height/scaleFactor), 0, C.int(0x10), nil, nil, nil)
        in := make([]uint8, 0)
        in = append(in, a.Y...)
        in = append(in, a.Cb...)
        in = append(in, a.Cr...)
        stride := []C.int{C.int(width), C.int(width / 2), C.int(width / 2), 0}
        outstride := []C.int{C.int(width / scaleFactor), C.int(width / scaleFactor / 2), C.int(width / scaleFactor / 2), 0}
        out := make([]uint8, width*height/scaleFactor/scaleFactor*3/2)
        C.sws_scale(context, (**C.uint8_t)(unsafe.Pointer(&in[0])), (*C.int)(&stride[0]), 0,
            C.int(height), (**C.uint8_t)(unsafe.Pointer(&out[0])), (*C.int)(&outstride[0]))
        min := image.Point{0, 0}
        max := image.Point{width / scaleFactor, height / scaleFactor}
        output := image.NewYCbCr(image.Rectangle{Min: min, Max: max}, image.YCbCrSubsampleRatio420)
        paneSize := width * height / scaleFactor / scaleFactor
        output.Y = out[:paneSize]
        output.Cb = out[paneSize : paneSize*5/4]
        output.Cr = out[paneSize*5/4:]
        opt := jpeg.Options{
            Quality: 90,
        }
        f, _ := os.Create("img.jpeg")
        jpeg.Encode(f, output, &opt)
    }
}



  


  Everytime I run the code snippet, I got an error saying bad dst image pointers, what is the problem of my code. I am new to cgo, so the code is probably silly to you, I apology for that.
If you have more elegant way to achieve the functionality, I am all ears. Any suggestion would be appreciated.

  


• ffmpeg av_interleaved_write_frame() : Broken pipe under windows

  7 avril 2016, par Allen

  I am using ffmpeg to convert original media file to rawvideo yuv format, ouputed the yuv to pipe, then my command tool receive the raw yuv as input, do some processing.

  e.g :

  D:\huang_xuezhong\build_win32_VDNAGen>ffmpeg -i test.mkv -c:v rawvideo -s 320x240 -f rawvideo - | my_tool -o output

  every time, when run the command, ffmpeg will dump this av_interleaved_write_frame(): Broken pipe error msg :

  Output #0, rawvideo, to 'pipe:':
  
    Metadata:
  
    encoder         : Lavf56.4.101
  
    Stream #0:0: Video: rawvideo (I420 / 0x30323449), yuv420p, 320x240 [SAR 120:91 DAR 160:91], q=2-31, 200 kb/s, 24 fps, 24 tbn, 24 tbc (default)
  
    Metadata:
  
        encoder         : Lavc56.1.100 rawvideo
  
    Stream mapping:
  
        Stream #0:0 -> #0:0 (h264 (native) -> rawvideo (native))
  
  Press [q] to stop, [?] for help
  
  av_interleaved_write_frame(): Broken pipe
  
  frame=    1 fps=0.0 q=0.0 Lsize=     112kB time=00:00:00.04 bitrate=22118.2kbits/s
  
  video:112kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing o
  
  verhead: 0.000000%
  
  Conversion failed!

  in my souce code : it take stdin as the input file, every time, it read a frame size content from it, if the content read is less than a frame size, then continue read, until a frame is fetched, then it use the frame content to generate something.

  int do_work (int jpg_width, int jpg_height) 
  
  {
  
  int ret = 0;
  
  
  
  FILE *yuv_fp = NULL;
  
  unsigned char * yuv_buf = NULL;
  
  int frame_size = 0;
  
  int count = 0;
  
  int try_cnt = 0;
  
  
  
  frame_size = jpg_width * jpg_height * 3 / 2;
  
  va_log (vfp_log, "a frame size:%d\n", frame_size);
  
  
  
  yuv_fp = stdin;
  
  
  
  yuv_buf = (unsigned char *) aligned_malloc_int(
  
          sizeof(char) * (jpg_width + 1) * (jpg_height + 1) * 3, 128);
  
  
  
  if (!yuv_buf) {
  
      fprintf (stderr, "malloc yuv buf error\n");
  
      goto end;
  
  } 
  
  
  
  memset (yuv_buf, 0, frame_size);
  
  while (1) {
  
  
  
      try_cnt++;
  
      va_log (vfp_log, "try_cnt is %d\n", try_cnt);
  
  
  
      //MAX_TRY_TIMES = 10
  
      if (try_cnt > MAX_TRY_TIMES) {
  
          va_log (vfp_log, "try time out\n");
  
          break;
  
      }
  
  
  
      count = fread (yuv_buf + last_pos, 1, frame_size - last_pos, yuv_fp);
  
      if (last_pos + count < frame_size) {
  
          va_log (vfp_log, "already read yuv: %d, this time:%d\n", last_pos + count, count);
  
          last_pos += count;
  
          continue;
  
      }
  
  
  
      // do my work here
  
  
  
      memset (yuv_buf, 0, frame_size);
  
      last_pos = 0;
  
      try_cnt = 0;
  
  }
  
  
  
  end:
  
  if (yuv_buf) {
  
      aligned_free_int (yuv_buf);
  
  }
  
  
  
  return ret;
  
  }

  my log :

  2016/04/05 15:20:38 : a frame size:115200
  2016/04/05 15:20:38 : try_cnt is 1
  2016/04/05 15:20:38 : already read yuv : 49365, this time:49365
  2016/04/05 15:20:38 : try_cnt is 2
  2016/04/05 15:20:38 : already read yuv : 49365, this time:0
  2016/04/05 15:20:38 : try_cnt is 3
  2016/04/05 15:20:38 : already read yuv : 49365, this time:0
  2016/04/05 15:20:38 : try_cnt is 4
  2016/04/05 15:20:38 : already read yuv : 49365, this time:0
  2016/04/05 15:20:38 : try_cnt is 5
  2016/04/05 15:20:38 : already read yuv : 49365, this time:0
  2016/04/05 15:20:38 : try_cnt is 6
  2016/04/05 15:20:38 : already read yuv : 49365, this time:0
  2016/04/05 15:20:38 : try_cnt is 7
  2016/04/05 15:20:38 : already read yuv : 49365, this time:0
  2016/04/05 15:20:38 : try_cnt is 8
  2016/04/05 15:20:38 : already read yuv : 49365, this time:0
  2016/04/05 15:20:38 : try_cnt is 9
  2016/04/05 15:20:38 : already read yuv : 49365, this time:0
  2016/04/05 15:20:38 : try_cnt is 10
  2016/04/05 15:20:38 : already read yuv : 49365, this time:0
  2016/04/05 15:20:38 : try_cnt is 11
  2016/04/05 15:20:38 : try time out
  ```

  my question :

  when piping used ,does ffmpeg write content to pipe buffer as soon as it has content, or it will buffer some size content, then flushes them to pipe ?Maybe some internal logic that I misunderstood,any one could help explain or fix my code ?

  PS : this command run ok under linux.