Recherche avancée

Sur d’autres sites (5756)

• Big question of Target size, or video bitrate must be specified when using two-pass

  14 juillet 2022, par Rojojun

  I tried to solve the error which is Target size, or video bitrate must be specified when using two-pass.

  


  But I couldn't find how to solve it and how to find path of video exactly I attached my code below this post

  


  Please give me some tips of solving the problem !

  


  @Service
public class ThumbnailService {

    public HashMap exportThumbnail(File file) throws Exception {
        // file is from controller and form-data
        //String inputPath = "Users/hojunna/Download/my/";
        //String outputPath = "/usr/local/video/thumbnail/";

        String ffmpegBasePath = "/opt/homebrew/bin/";
        FFmpeg ffmpeg = new FFmpeg(ffmpegBasePath+"ffmpeg");        
        FFprobe ffprobe = new FFprobe(ffmpegBasePath+"ffprobe");    
        
        FFmpegBuilder builder = new FFmpegBuilder()
                .setInput("/Users/hojunna/Desktop/" + file)                         
                //.overrideOutputFiles(true)                    
                //.addExtraArgs("-ss", "00:00:01")          
                .addOutput("/Users/hojunna/Desktop/test.jpg")       
                .setFrames(1)
                .setVideoFilter("select='gte(n\\,10)',scale=200:-1")
                .done();

        FFmpegExecutor executor = new FFmpegExecutor(ffmpeg, ffprobe);      
        executor.createJob(builder).run();                                  
        executor.createTwoPassJob(builder).run();                           

        HashMap resultMap = new HashMap();
        return resultMap;
    }
}


  


• Why does FFMPEG b-frames and b_pyramid offset start_pts (and start_time) in fragmented output ?

  13 juillet 2022, par Vans S

  It seems when transcoding something into fragmented or segmented output the start_pts is not 0, and no combination of options can make it 0 other than setting bframes to 0 and bpyramid to 0. This does not happen with regular nonfragmented output.

  


  Does anyone know why this is, and how to prevent it as I believe this is causing weird timesync issues with playback in browsers (every fragment slightly delays the video more and more) where after 2-3 hours the stream can end up 15minutes+ delayed.

  


  Example where each segment start_pts is not 0 :

  


  ffmpeg -i in.mp4 -pix_fmt yuv420p -an -f yuv4mpegpipe -frames:v 150 - | ffmpeg -f yuv4mpegpipe -i - -y -force_key_frames 1,2,3,4 -map 0 -codec:v libx264 -f segment -segment_list out.csv -segment_times 2,4 -segment_time_delta 0.05 -preset:v fast -segment_format_options movflags=+faststart out%03d.mp4


  


  start_pts is 0 here if we add :

  


  -x264opts b_pyramid=0 -bf 0 
#or change codec to
-codec:v mpeg4
#or output regular mp4
ffmpeg -i in.mp4 -pix_fmt yuv420p -an -f yuv4mpegpipe -frames:v 150 - | ffmpeg -f yuv4mpegpipe -i - -y out.mp4


  


  EDIT : Looking into this further I am starting to think this is a bug with how empty_moov interacts with the negative_cts_offsets flag. (when empty_moov is used, negative_cts_offsets seems to be ignored, and we need empty_moov for full webbrowser support.)

  


• Why does FFMPEG b-frames and b_pyramid offset start_pts (and start_time) in fragmented output ?

  13 juillet 2022, par Vans S

  It seems when transcoding something into fragmented or segmented output the start_pts is not 0, and no combination of options can make it 0 other than setting bframes to 0 and bpyramid to 0. This does not happen with regular nonfragmented output.

  


  Does anyone know why this is, and how to prevent it as I believe this is causing weird timesync issues with playback in browsers (every fragment slightly delays the video more and more) where after 2-3 hours the stream can end up 15minutes+ delayed.

  


  Example where each segment start_pts is not 0 :

  


  ffmpeg -i in.mp4 -pix_fmt yuv420p -an -f yuv4mpegpipe -frames:v 150 - | ffmpeg -f yuv4mpegpipe -i - -y -force_key_frames 1,2,3,4 -map 0 -codec:v libx264 -f segment -segment_list out.csv -segment_times 2,4 -segment_time_delta 0.05 -preset:v fast -segment_format_options movflags=+faststart out%03d.mp4


  


  start_pts is 0 here if we add :

  


  -x264opts b_pyramid=0 -bf 0 
#or change codec to
-codec:v mpeg4
#or output regular mp4
ffmpeg -i in.mp4 -pix_fmt yuv420p -an -f yuv4mpegpipe -frames:v 150 - | ffmpeg -f yuv4mpegpipe -i - -y out.mp4


  


  EDIT : Looking into this further I am starting to think this is a bug with how empty_moov interacts with the negative_cts_offsets flag. (when empty_moov is used, negative_cts_offsets seems to be ignored, and we need empty_moov for full webbrowser support.)