Recherche avancée

Sur d’autres sites (5805)

• Android sws_scale RGB Frame taking long time and cause of latency in the video

  17 avril 2018, par AJit

  I am reading frames from FFMPEG and try to directly draw to surface window in Native android, If i scale image to exact size what we are getting from camera and apply YUV420P to RGBA it take 1ms to scale through av_image_fill_arrays but if i try to scale image what i need for surface, then i take 25 to 30ms to scale same frame. so latency is the problem.

  In below example getting YUV420P pixel format.

  Low latency :[ 1ms by sws_scale]

  swsContext = sws_getContext(videoCodecContext->width,
  
                                      videoCodecContext->height,
  
                                      videoCodecContext->pix_fmt,
  
                                      videoCodecContext->width,
  
                                      videoCodecContext->height,
  
                                      AV_PIX_FMT_RGB0,
  
                                      SWS_FAST_BILINEAR, NULL, NULL, NULL);
  
  av_image_fill_arrays()
  
  av_read_frame()
  
  avcodec_decode_video2(
  
  sws_scale(swsContext,
  
            (const uint8_t *const *) videoFrame->data,
  
            videoFrame->linesize,
  
            0,
  
            videoCodecContext->height,
  
            pictureFrame->data,
  
            pictureFrame->linesize);
  
  ANativeWindow_lock()
  
  Write all buffer bytes to window.
  
  ANativeWindow_unlockAndPost()

  Low latency :[ 30ms by sws_scale]

  [videoContext Width: 848 Height: 608]
  
  swsContext = sws_getContext(videoCodecContext->width,
  
                                      videoCodecContext->height,
  
                                      videoCodecContext->pix_fmt,
  
                                      1080,
  
                                      608,
  
                                      AV_PIX_FMT_RGB0,
  
                                      SWS_FAST_BILINEAR, NULL, NULL, NULL);

  Whenever we change context width and height other then videoContext it will take more then 30ms and we are delaying the video.

  TRY 1 : Pass the buffer from JNI to Java and create bitmap there to scale later on but createBitmap itself take 500ms so no useful.

  TRY 2 : Direct YUV420P to RGB conversion. still long time then sws_scale.

  TRY 3 : Direct write YUV to window bites but it show without color(If anyone have solution here will might helpful).

  TRY 4 : Use yuvlib, no luck.

  TRY 5 : Different color combinations and flags in swsContext.

  Any help would appreciated.

• ffmpeg av_read_frame is blocked long time

  10 avril 2018, par geeeek

  while (av_read_frame(pFormatCtx, packet) >= 0) {
  
      if (packet->stream_index == videoindex) {
  
          ret = avcodec_decode_video2(pCodecCtx, pFrame, &got_picture, packet);
  
          if (ret < 0) {
  
              printf("Decode Error.\n");
  
              return -1;
  
          }
  
          printf("Succeed to decode 1 frame! %d\n", tt++);
  
      }
  
      av_packet_unref(packet);
  
  }

  I am using ffmpeg to provide live audio/video RTSP streaming in android.

  But I have a trouble from blocking av_read_frame function forever.
  Android Log :

  I/libnav: Succeed to decode 1 frame! 70
  
  I/libnav: Succeed to decode 1 frame! 71
  
  I/libnav: Succeed to decode 1 frame! 72
  
  I/libnav: Succeed to decode 1 frame! 73
  
  I/libnav: Succeed to decode 1 frame! 74
  
  I/libnav: Succeed to decode 1 frame! 75
  
  blocking forever...

  Thus I
  have experimented some code stuffs. I have copied Android jni code about ffmpeg to pc(ubuntu). Above my code is worked without any blocking.

  Succeed to decode 21122 frame!
  
  Succeed to decode 21123 frame!
  
  Succeed to decode 21124 frame!
  
  ...
  
  without blocking ...

  I guess that this problem is about delay. I have removed avcodec_decode_video2 call in android jni. my code is executed without any blocking.

  I have found ffmpeg av_read_frame() need very long time to stop threads. But it can’t help me.

  If you have a similiar problem, please help me.

• Elixir long running background task doesn't complete, sometimes crashes

  29 mars 2018, par Kulix

  As a little bit of background, I have an endpoint that is responsible for video upload. What I want to do, is copy the video file to a temporary location from the request, and spin up an asynchronous task (ffmpeg shell process) to transcode that video in the background so that my endpoint can return a 200 in a timely manner, and the response does not wait for ffmpeg to finish transcoding that video.

  Here is my code from the controller.

  def create(conn, %{"file" => file ... })
  
      uuid = Video.uuid()
  
      tmp_path = Application.get_env(:myapp, :video_tmp_path) <> "/" <> uuid
  
      :ok = File.cp(file.path, tmp_path)
  
      VideoService.process(tmp_path, final_path)

  The inside of VideoService looks like the following.

  defmodule MyApp.Services do
  
    defmodule VideoService do
  
      def process(tmp_path, final_path) do
  
        Task.start_link fn ->
  
          System.cmd("ffmpeg", ["-i", tmp_path, final_path, "-hide_banner"])
  
          File.rm(tmp_path)
  
        end
  
      end
  
    end
  
  end

  The problem I am having here is that no matter what, nothing past the System.cmd("ffmpeg") call executes in the VideoService, and sometimes the System.cmd call doesn’t even spin up an ffmpeg process. Am I doing something wrong here ? What I need is a way to spin this ffmpeg shell process out in the background from the controller / service and respond with a 200 on video upload. Thanks for help in advance. I am new to elixir / OTP, so I’m sure I’m doing something stupid.

  I also randomly see the following error

  erl_child_setup: failed with error 32 on line 240