Recherche avancée

Sur d’autres sites (10396)

• vsync flag usage in ffmpeg while filtering

  13 octobre 2022, par antortjim

  I am trying to apply a threshold to an input video with ffmpeg, but I observe the following warning emitted for every processed frame

  


  [mp4 @ 0x56360181a200] Non-monotonous DTS in output stream 0:0; previous: 182272, current: 182272; changing to 182273. This may result in incorrect timestamps in the output file.


  


  where the previous and current are always 1 less than the value to which the DTS (Decoding Time Stamp) is changed

  


  I have noticed this warning is emitted only if I set -vsync passthrough (which I changed from the original -vsync 0 which is seen in many online examples).

  


  # input.mp4 has resolution 790x790
ffmpeg -vsync passthrough  -i input.mp4  -f lavfi -i color=808080:s=790x790 -f lavfi -i color=black:s=790x790 -f lavfi -i color=white:s=790x790 -filter_complex '[0:v][1:v][2:v][3:v]threshold' -an -c:v h264_nvenc threshold.mp4


  


  Shall I leave the vsync flag set to the default (auto or -1), or is -vsync passthrough essential to guarantee the frames are displayed in the right order ? In that case, how do I handle this warning ? Some other online examples I found of users experiencing this warning are different from mine, because in their case they are concatenating videos (see 1, 2

  


  From the documentation on the -vsync flag, at the end, I see :

  


  


  With -map you can select from which stream the timestamps should be taken. You can leave either video or audio unchanged and sync the remaining stream(s) to the unchanged one

  


  


  Maybe this warning should be handled with -map ? But I don't know how.

  


  Sidenote, I keep getting the deprecation warning asserting me to change -vsync for -fps_mode, however doing so breaks the command.

  


  FFPEG Version :

  


  commit 28ac2279adb860ea8b90d3073603912bf3eb6a83 from ffmpeg master branch

  


  ffmpeg version N-108625-g28ac2279ad Copyright (c) 2000-2022 the FFmpeg developers
built with gcc 9 (Ubuntu 9.4.0-1ubuntu1~20.04.1)
configuration: --enable-nonfree --enable-cuda-nvcc --enable-libnpp --enable-gpl --extra-cflags=-I/usr/local/cuda/include --extra-ldflags=-L/usr/local/cuda/lib64 --disable-static --enable-shared
libavutil      57. 39.101 / 57. 39.101
libavcodec     59. 50.100 / 59. 50.100
libavformat    59. 34.101 / 59. 34.101
libavdevice    59.  8.101 / 59.  8.101
libavfilter     8. 49.101 /  8. 49.101
libswscale      6.  8.112 /  6.  8.112
libswresample   4.  9.100 /  4.  9.100
libpostproc    56.  7.100 / 56.  7.100


  


  OS

  


  Ubuntu 20.04.4 LTS

  


• configure : Add -D_CRT_NONSTDC_NO_WARNINGS when building with msvc

  28 octobre 2015, par Martin Storsjö

  configure : Add -D_CRT_NONSTDC_NO_WARNINGS when building with msvc
  This silences warnings like this one :
  
  libavformat/file.c(62) : warning C4996 : ’read’ : The POSIX name for this
  
  item is deprecated. Instead, use the ISO C++ conformant name : _read.
  
  See online help for details.
  Signed-off-by : Martin Storsjö <martin@martin.st>
  

  • [DBH] configure

• ffmpeg libswresample channel mapping

  21 mars 2024, par Joseph Katrinka

  Im currently working on some c++ that uses ffmpeg to take audio from a stereo file and make a mono audio file using only the sample from one channel.
I havent been able to find and examples of people using the swr_set_channel_mapping call online, so Im wondering if anyone knows of the correct usage.
Right now Im doing something like this

  


  swr_alloc_set_opts2(&swrCtx, &out_channel_layout, AV_SAMPLE_FMT_S16, codecCtx->sample_rate, &codecCtx->ch_layout, codecCtx->sample_fmt, codecCtx->sample_rate, 0, NULL);

int* channel_mapping = (int*)av_mallocz(2 * sizeof(int));
if (useLeft)
{
channel_mapping[0] = AV_CH_FRONT_LEFT;
}
else
{
channel_mapping[0] = AV_CH_FRONT_RIGHT;
}   
swr_set_channel_mapping(swrCtx, channel_mapping);

if (swr_init(swrCtx) < 0)
    {
        DBG("failed to init resampler");
        ffmpegCleanup();
        return false;
    }



  


  Is this the correct way to do this ? Its been giving me some problems and Im worried I could be doing something wrong.
Thanks.

  


  Ive tried different ways of defining channel_mapping with no success. Im not sure what the correct way is and an example would be pretty useful.