Recherche avancée

Sur d’autres sites (18642)

• FFMPEG H264 with custom overlay per frame

  4 octobre 2020, par La bla bla

  We have a stream that is stored in the cloud (Amazon S3) as individual H264 frames. The frames are stored as framexxxxxx.264, the numbering doesn't start from 0 but rather from some larger number, say 1000 (so, frame001000.264)

  


  The goal is to create a mp4 clip which is either timelapse or just faster for inspection and other checking (much faster, compressing around 3 hours of video down to < 20 minutes), this also requires we overlay the frame number (the filename) on the frame itself

  &#xA;

  At first I was creating a timelapse by pulling from S3 only the keyframes (i-frames ? still rather new to codecs & stuff) and overlaying the filename on them and saving as png (which probably isn't needed, but that's what I did) using (this command is used inside a python script)

  &#xA;

  ffmpeg -y -i {h264_name} -vf \"scale=1920:-1, 
drawtext=fontfile=/usr/share/fonts/truetype/ubuntu-font-family/Ubuntu-B.ttf:fontsize=34:text={txt}:fontcolor=white:x=50:y=50:bordercolor=black:borderw=2\" 
-c:a copy -pix_fmt yuv420p {basename}.png


  &#xA;

  after this I combined all the frames by using python to convert the lowest numbered frame to 0.png and incrementing (so it would be continuous, because I only used keyframes the numbers originally weren't sequential) and running

  &#xA;

  ffmpeg -y -f image2 -i %d.png -r {self.params.fps} -vcodec libx264 -crf {self.params.crf} -pix_fmt yuv420p {out_file}


  &#xA;

  and this worked great, but the difference between keyframes was too long to allow for proper inspection

  &#xA;

  so now for the question(s)

  &#xA;

  since I know frames that are not keyframes (p-frames ?) can't be used alone by ffmpeg, the method of overlaying the file name and converting it to png (or keep as h264, same thing) won't work, or at least, I couldn't find a way for it to work, maybe there's a way to specify a frame's keyframe ?, how can one overlay the filename (and not the frame number as shown here for example)

  &#xA;

  Also, is it possible to skip some p-frames between the keyframes ? (so if a keyframe is every 30 frames, we would take a keyframe, a frame 15 frames later, and next another keyframe)

  &#xA;

  I thought about using ffmpeg's pipe option to feed it with the files as they're being downloaded, but I'm not sure if I can specify drawtext this way

  &#xA;

  Also, if there's another alternative that can achieve that (at first I was converting to png, using python and OpenCV to add the filename and then merging the pngs to mp4, but then I found drawtext can do that in a single command so I used it)

  &#xA;

• Android FFMPEG Build with librtmp

  20 décembre 2013, par Dan Turner

  I have been pulling my hair out trying to compile FFMPEG for Android with librtmp enabled. I have successfully built FFMPEG using the Guardian Project here, but it doesn't have librtmp support. The only answer I have found to this issue is on a previous Stack Overflow question (HERE), but it doesn't want to work for me.

  At the moment, I have the cross-compiled librtmp.so.0 file from the official rtmpdump android build in an rtmpdump/librtmp folder sitting in the android-ffmpeg folder. An extract from my configure_ffmpeg.sh file reads as follows :

  ./configure \
  
  $DEBUG_FLAG \
  
  --arch=arm \
  
  --cpu=cortex-a8 \
  
  --target-os=linux \
  
  --enable-runtime-cpudetect \
  
  --prefix=$prefix \
  
  --enable-pic \
  
  --disable-shared \
  
  --enable-static \
  
  --cross-prefix=$NDK_TOOLCHAIN_BASE/bin/$NDK_ABI-linux-androideabi- \
  
  --sysroot="$NDK_SYSROOT" \
  
  --extra-cflags="-I../x264 -mfloat-abi=softfp -mfpu=neon" \
  
  --extra-ldflags="-L../x264" \
  
  --extra-cflags="-I/home/dan/android-ffmpeg/rtmpdump" \
  
  --extra-ldflags="-L/home/dan/android-ffmpeg/rtmpdump -lrtmp"
  
  \
  
  
  
  --enable-version3 \
  
  --enable-gpl \
  
  \
  
  --disable-doc \
  
  --enable-yasm \
  
  \
  
  --enable-decoders \
  
  --enable-encoders \
  
  --enable-muxers \
  
  --enable-demuxers \
  
  --enable-parsers \
  
  --enable-protocols \
  
  --enable-filters \
  
  --enable-avresample \
  
  --enable-libfreetype \
  
  \
  
  --disable-indevs \
  
  --enable-indev=lavfi \
  
  --disable-outdevs \
  
  \
  
  --enable-hwaccels \
  
  \
  
  --enable-ffmpeg \
  
  --disable-ffplay \
  
  --disable-ffprobe \
  
  --enable-ffserver \
  
  --enable-network \
  
  \
  
  --enable-libx264 \
  
  --enable-zlib \
  
  --enable-librtmp \

  When I try to compile this, it eventually displays an error and my FFMPEG config.log file tells me that it can't find -lrtmp. I'm positive I'm directing it to the right directory... does anyone have any ideas ?

  Regards

  Dan

• Why motion vector extracted from b frames are unchange ?

  19 août 2019, par 霍宇琦

  I am writing c codes to extract motion vectors from b frame in MPEG4(Part2) compressed video format. But some motion vector seems wrong.

  I use a raw video clips, using the extract_mvs.c from ffmpeg 4.2. For example if the frame seq is IPBPPBPP.... i can get all the side data for all frame. But when inspecting the mv->src_x, mv->src_y, mv->dst_x, mv->dst_y i find that all the srcs are equal to dsts for some individual frames, there must be sth wrong in it, but i change little from the official code.

  //modify from ffmpeg/doc/example/extract_mvs.c:
  
  
  
  while(getting frames one by one) {
  
      AVFrameSideData *sd;
  
      video_frame_count++;
  
  
  
      //printf("%d", video_frame_count);
  
      if(video_frame_count < 19){
  
          if (frame->pict_type == AV_PICTURE_TYPE_I ) printf("\nI");
  
          if (frame->pict_type == AV_PICTURE_TYPE_B ) printf("B"); 
  
          if (frame->pict_type == AV_PICTURE_TYPE_P ) printf("P");
  
  
  
          sd = av_frame_get_side_data(frame, AV_FRAME_DATA_MOTION_VECTORS);
  
          if (sd) {
  
              const AVMotionVector *mvs = (const AVMotionVector *)sd->data;
  
              for (i = 0; i < sd->size / sizeof(*mvs); i++) {
  
                  const AVMotionVector *mv = &mvs[i];
  
                  if (mv->dst_x - mv->src_x != 0 || mv->dst_y - mv->src_y != 0) {
  
                  printf("%d,%2d,%2d,%2d,%4d,%4d,%4d,%4d,0x%"PRIx64"\n",
  
                      video_frame_count, mv->source,
  
                      mv->w, mv->h, mv->src_x, mv->src_y,
  
                      mv->dst_x, mv->dst_y, mv->flags);
  
                      break;
  
                  }
  
              }
  
          }
  
          av_frame_unref(frame);
  
      }
  
  }

  Outputs are :

  Input #0, avi, from 'origin.avi':
  
    Duration: 00:00:13.63, start: 0.000000, bitrate: 492 kb/s
  
      Stream #0:0: Video: mpeg4 (DX50 / 0x30355844), yuv420p, 320x240 [SAR 1:1 DAR 4:3], 486 kb/s, 30 fps, 30 tbr, 30 tbn, 30k tbc
  
      Metadata:
  
        title           : H:\HumanActionDB\MotionClips\hmdb51_30fps_wBrd_10off_divx\brush_??
  
  framenum,source,blockw,blockh,srcx,srcy,dstx,dsty,flags
  
  
  
  [mpeg4 @ 0x55739dc29580] Video uses a non-standard and wasteful way to store B-frames ('packed B-frames'). Consider using the mpeg4_unpack_bframes bitstream filter without encoding but stream copy to fix it.
  
  IP2,-1,16,16, 137,  24, 136,  24,0x0
  
  BP4,-1,16,16, 152,  57, 152,  56,0x0
  
  P5,-1,16,16,  56, 155,  56, 152,0x0
  
  BP7,-1,16,16, 151,  40, 152,  40,0x0
  
  PB9,-1,16,16, 151,  40, 152,  40,0x0
  
  P10,-1,16,16, 152,  39, 152,  40,0x0
  
  P11,-1,16,16,  26,  55,  24,  56,0x0
  
  B12,-1,16,16, 152,  39, 152,  40,0x0
  
  P13,-1,16,16, 152,  39, 152,  40,0x0
  
  P14,-1,16,16,  41, 168,  40, 168,0x0
  
  B15,-1,16,16, 152,  39, 152,  40,0x0
  
  P16,-1,16,16, 153,  39, 152,  40,0x0
  
  PB18,-1,16,16, 168,  55, 168,  56,0x0

  you can see that the third frame(B) and the sixth, eighth frame(B, P) and the 17th frame (P) can be read, and the data can be extract from them, but

  mv->dst_x == mv->src_x && mv->dst_y - mv->src_y

  Can someone help me ? Thanks.