Recherche avancée

Sur d’autres sites (8980)

• how to kill a thread which hang up and release resources [duplicate]

  15 octobre 2019, par DJI_lover

  This question already has an answer here :

  • How do I terminate a thread in C++11 ?
    
    5 answers
    

  I am doing some jobs in ubuntu with c++.I want to know how to kill a thread when function blocking in a thread and there is no timeout mechanism for this function. I want to force to kill the thread when function is blocking and release all the resources in the thread and then restart the thread.I can’t set a quit flag in thread because it nerver get there when function hang up.Could someone tell me the better way to do this ? Thanks !

  Code :

  static int encode_and_write_frame(AVCodecContext *codec_ctx, AVFormatContext *fmt_ctx, AVFrame *frame)
  
  {
  
      AVPacket pkt = {0};
  
      av_init_packet(&pkt);
  
      int ret = avcodec_send_frame(codec_ctx, frame);
  
      if (ret < 0)
  
      {
  
          fprintf(stderr, "Error sending frame to codec context!\n");
  
          return ret;
  
      }
  
      ret = avcodec_receive_packet(codec_ctx, &pkt);
  
      if (ret < 0)
  
      {
  
          fprintf(stderr, "Error receiving packet from codec context!\n" );
  
          return ret;
  
      }
  
      int ret1 = av_interleaved_write_frame(fmt_ctx, &pkt);
  
      av_packet_unref(&pkt);
  
      return ret1;
  
  }
  
  
  
  void thread_worker(){
  
      while(ros::ok()){
  
          encode_and_write_frame(param1,param2,param3);
  
          usleep(500000);
  
      }
  
  }
  
  
  
  thread t(&thread_worker);

  when the network situation is good, av_interleaved_write_frame() can return quickly,but when network was broken down, av_interleaved_write_frame() will hang forever, and I can’t set timeout for av_interleaved_write_frame so that I want to kill the blocking thread and restart it.

• Can't find error in function for changing sampling rate

  29 avril 2024, par kitty uwu

  I have function for changing sampling rate of audio (only one channel) :

  


  int change_sampling_rate(float *audio_input, int input_sample_rate, int output_sample_rate, int input_num_of_samples, float **audio_output, int *result_num_of_samples) {
    AVChannelLayout src_ch_layout = AV_CHANNEL_LAYOUT_MONO;
    AVChannelLayout dst_ch_layout = AV_CHANNEL_LAYOUT_MONO;

    struct SwrContext *swr_ctx;
    swr_ctx = swr_alloc();
    int ret;
    if (!swr_ctx) {
        fprintf(stderr, "Could not allocate resampler context\n");
        ret = AVERROR(ENOMEM);
    }

    av_opt_set_chlayout(swr_ctx, "in_chlayout",    &src_ch_layout, 0);
    av_opt_set_int(swr_ctx, "in_sample_rate",       input_sample_rate, 0);
    av_opt_set_sample_fmt(swr_ctx, "in_sample_fmt", AV_SAMPLE_FMT_FLT, 0);

    av_opt_set_chlayout(swr_ctx, "out_chlayout",    &dst_ch_layout, 0);
    av_opt_set_int(swr_ctx, "out_sample_rate",       output_sample_rate, 0);
    av_opt_set_sample_fmt(swr_ctx, "out_sample_fmt", AV_SAMPLE_FMT_FLT, 0);

    if ((ret = swr_init(swr_ctx)) < 0) {
        fprintf(stderr, "Failed to initialize the resampling context\n");
        return -1;
    }

    int output_samples_count = av_rescale_rnd(swr_get_delay(swr_ctx, input_sample_rate) + input_num_of_samples, output_sample_rate, input_sample_rate, AV_ROUND_UP);
    uint8_t **resampled_data = NULL;
    if (av_samples_alloc_array_and_samples(&resampled_data, NULL, 1, output_samples_count, AV_SAMPLE_FMT_FLT, 0) < 0) {
        fprintf(stderr, "Could not allocate resampled data\n");
        swr_free(&swr_ctx);
        return -1;
    }

    const uint8_t *in_samples[1] = {(const uint8_t *)audio_input};
    int frame_count = swr_convert(swr_ctx, resampled_data, output_samples_count, in_samples, input_num_of_samples);

    if (frame_count < 0) {
        fprintf(stderr, "Error while resampling\n");
        av_freep(&resampled_data[0]);
        free(resampled_data);
        swr_free(&swr_ctx);
        return -1;
    }

    *audio_output = (float *) malloc(frame_count * sizeof(float));
    if (!*audio_output) {
        fprintf(stderr, "Could not allocate memory for output\n");
        av_freep(&resampled_data[0]);
        free(resampled_data);
        swr_free(&swr_ctx);
        return -1;
    }

    memcpy(*audio_output, resampled_data[0], frame_count * sizeof(float));

    *result_num_of_samples = frame_count;
    av_freep(&resampled_data[0]);
    swr_free(&swr_ctx);
    return SUCCESS;
}


  


  When I run tests on time lag between two files (mp3) with different sampling rates, it gives answer that differs on about 15-20 ms with right answer. Can anybody, please, help me find mistakes in the code ?

  


  For example, I have two audios : [audio_1] (https://jmp.sh/s/USFPaGnHXVuKFVYarYpm) and [audio_2] (https://jmp.sh/s/jbmWbPTwkdDujAocmi56) - second audio is just a sample of first. The answer should be 35264 ms, but my function gives 35249 ms :(

  


• Concatenate two mp4 files using FFmpeg with Android

  4 juillet 2013, par vijay

  I need to merge video files in Android using FFmpeg.I used the code from halfninja and passed the following parameters to run method in Videokit.c.

  ffmpeg -i "concat:input1.mpg|input2.mpg" -c copy output.mpg

  It causes error when these parameters are passed from Android through JNI. Then i replaced -c copy with -vcodec copy -acodec copy.It worked fine.

  Now i tried to concat two .mp4 files / two .3gp files by using the above method and my output format is mp4.The output file is generated but it is empty(0kb).

  When i googled it out, i was asked to convert all mp4 or 3gp files into mpg and then concat it and again convert back to mp4 in the following post :

  Concatenate mp4 files in Android using halfninja ffmpeg (Refer Steven Penny's answer).

  But i have problems(only first input file is converted and the app exits) when converting and storing the intermediate file(i am calling run method thrice).Also this method stores all the input files in the device sdcard and this should not be done.

  -c copy works fine when passing parameters for ffmpeg executable(generated when ndk-build is made) through terminal but it throws an error "unrecognized option c" when called through JNI from android device.Can anyone let me know why this happens ?

  When we look at Steven's answer, we have to call the run method multiple times to covert all the input files and one more time to concat all these.Is it possible to it in a single call ?

  Is it possible to concat video files with different formats without storing the input files in device(mp4+mp4, mp4+3gp,3gp+3gp are my possible combinations of input) ?

  Any help is appreciated.