Recherche avancée

Sur d’autres sites (9727)

• given a list of mp4, generate fmp4 and concatenate

  6 août 2021, par onion

  my input is a list of mp4 (for example each one is 10s), each mp4 will have correct timestamp, for example the second mp4 represents from 10s - 20s data.

  


  To simulate my input, I generate a list of mp4 this way

  


  ffmpeg -i ../origin-long-video.mp4 -map 0 -c copy -f segment -segment_time 10 -force_key_frames "expr:gte(t,n_forced*2)" -reset_timestamps 0 videos/output_%03d.mp4


  


  Note I use reset_timestamps 0 so that the timestamp is preserved.

  


  Then I convert each mp4 to fragment mp4 by using

  


  ffmpeg -y   -i videos/output_001.mp4   -force_key_frames "expr:gte(t,n_forced*2)"   -sc_threshold 0   -s 1280x720   -c:v libx264 -b:v 1500k   -c:a copy   -hls_time 100   -hls_playlist_type vod   -hls_segment_type fmp4   -hls_segment_filename "hls1/file%d.m4s"  -copyts hls1/index.m3u8


  


  The above cmd is for the first mp4 file, i did same operation for other mp4 in the list.
Note I used a large hls_time so that each mp4 will result in one fmp4, and also I use copyts to preserve the timestamp.

  


  Then I want to concatenate for example the 2nd and the 3rd fmp4 to generate one mp4, I used

  


  cat init.mp4 > rs.mp4
cat 2nd fmp4 >> rs.mp4
cat 3rd fmp4 >> rs.mp4


  


  However when trying to play the generated rs.mp4, it is 20s black screen then 10s video (corresponding to the 3rd mp4).

  


  I tried another approach which just generated a list of fmp4 directly like this :

  


  ffmpeg -y   -i ../origin-long-video.mp4   -force_key_frames "expr:gte(t,n_forced*2)"   -sc_threshold 0   -s 1280x720   -c:v libx264 -b:v 1500k   -c:a copy   -hls_time 10   -hls_playlist_type vod   -hls_segment_type fmp4   -hls_segment_filename "videos/file%d.m4s"   videos/index.m3u8


  


  And then if I concatenate the 2nd and 3rd generated fmp4 use the same way above, the result mp4 plays well.

  


  I wonder what is the difference between the generated fragment mp4 of the two approach so that they leads to different behavior when concatenation. Thank you !

  


• Revision 4e2732c3c3 : Separate encode_block for pass 1 and 2 The encode_block for pass 1 takes simple

  23 octobre 2013, par Jingning Han

  Changed Paths :
   Modify /vp9/encoder/vp9_encodemb.c

  
  
  Separate encode_block for pass 1 and 2

  The encode_block for pass 1 takes simpler functionalities and can
  save a few branches. The main reason is to make encode_block only
  used after running rate-distortion optimization search in pass 2,
  hence allowing dual buffer stack approach later.

  Change-Id : I9e549ffb758e554fe185e48a07d6e0e01e475bcf

• How to use FFMPEG to write image as bytes using ProcessBuilder

  7 mai 2022, par ljnoah

  I have a callback function that gives me frames as bytes type to which I would like to pass as FFMPEG parameter to write them to a rtmp URL. but I don't really have any experience with ffmpeg, thus far I was not able to find an example on how to do it. Basically, I would like to know can I use use the bytes array that is FrameData that holds the images I am getting and write to ffmpeg as a parameter to be sent via streaming to a server using ProcessBuilder.

  


  private byte[] FrameData = new byte[384 * 288 * 4];
private final IFrameCallback mIFrameCallback = new IFrameCallback() {
      @Override
      public void onFrame(final ByteBuffer frameData) {
          frameData.clear();
          frameData.get(FrameData, 0, frameData.capacity());
          ProcessBuilder pb = new ProcessBuilder(ffmpeg ,  "-y",  "-f", "rawvideo",  "vcodec", "rawvideo",  "-pix_fmt", "bgr24",
                "-r", "25",
                "-i", "-",
                "-c:v", "libx264",
                "-pix_fmt", "yuv420p",
                "-preset", "ultrafast",
                "-f", "flv",
                "rtmp://192.168.0.13:1935/live/test");
          }
        Log.e(TAG, "mIFrameCallback: onFrame------");
        try {
            pb.inheritIO().start().waitFor();
        } catch (InterruptedException | IOException e) {
            e.printStackTrace();
        }
  };


  


  This callback gives me the frames from my camera on the fly and writes it to FrameData, which I can compress to a bitmap if needed. The current attempt isn't working as I have no idea how to pass my byte array as a parameter to ffmpeg to be streamed via rtmp as above

  


  to push my frames from the camera that are stored FrameData byte buffer via RTMP/RTSP to my server IP. I would use a similar approach in python like this :

  


  import subprocess
fps = 25
width = 224
height = 224
command = ['ffmpeg',  '-y',  '-f', 'rawvideo',  '-vcodec', 'rawvideo',  '-pix_fmt', 'bgr24',
           '-s', "{}x{}".format(width, height),
           '-r', str(fps),
           '-i', '-',
           '-c:v', 'libx264',
           '-pix_fmt', 'yuv420p',
           '-preset', 'ultrafast',
           '-f', 'flv',
           'rtmp://192.168.0.13:1935/live/test']
p = subprocess.Popen(command, stdin=subprocess.PIPE)
while(True):
    frame = np.random.randint([255], size=(224, 224, 3))
    frame = frame.astype(np.uint8)
    p.stdin.write(frame.tobytes())


  


  I really don't understand how to write my byte arrays to the ffmpeg as I would in this Python example above