Recherche avancée

Sur d’autres sites (8707)

• Pydub dir error . dynamic file path doesn't play sound but with specific file path the code works fine

  20 avril 2021, par anshul raj

  I don't know what's wrong with this code - it works fine when I provide an exact file path to the music file, but if I pass it in a dynamic way, it doesn't work.

  


  Actually my code is that user give a music file name it get downloaded then with the meta id it find the file in downloads dir then play it

  


   def songplayer(self,meta):
        def playmmusic(name):
            from pydub.playback import play
            from pydub import AudioSegment
            AudioSegment.converter = "C:\\ffmpeg\\bin\\ffmpeg.exe"
            AudioSegment.ffmpeg = "C:\\ffmpeg\\bin\\ffmpeg.exe"
            AudioSegment.ffprobe = "C:\\ffmpeg\\bin\\ffprobe.exe"
            sound = AudioSegment.from_file(name)
            play(sound)

        print(colored("Currently Playing : " + meta['title'],'yellow'))
        r=meta['id']
        tt='./downloads/'+r
        playmmusic(tt)


  


• FFMPEG Encoder error when playback with 1936 width video size

  5 février 2017, par AnhTuan.Ng

  I have a 4096x1760 YUV-NV12 video frames stream and i have to split it to 4 frames have resolution 1936x872, then encode them to 4 video file (Like bellow picture)
  Crop frame
  My algorithm is :
  1. Using openCV to crop and splitting (4096x1760) NV12 frame to 4 (1936x872) down size.
  2. Using FFMPEG swscale() to convert RGBA to I420 buffer
  3. Using FFMpeg to encode 4 (1936x872) frames to 4 video files.

  But i have a trouble, that 4 video encoded have error signals, Just horizontal stripes on screen (such as zebra skin, So sad !)
  Video error
  I think error caused by encoder, because when i change crop size to 1920x872, All thing seem be ok (By one step, I dump buffer data to bmp image to check, and every thing look good !)

  And, what a surprise ! When I change my crop famre size from 1936x872 to 1920x872. All things work done, Video output can be playing so smooth !

  So I think this error because my FFMPEG encoder setting up wrong. Please tell me why and show me a way to fix it ! Thanks you !

  this is my code :

  -My define :

  #define SIDE_FRAME_WIDTH        4096
  
  #define SIDE_FRAME_HEIGHT       1760
  
  
  
  #define SINGLE_SIDE_WIDTH       1936    //1920 is woking!
  
  #define SINGLE_SIDE_HEIGHT      872     //872
  
  
  
  #define SIDE_CROP_WIDTH         1936
  
  #define SIDE_CROP_HEIGHT        872

  My Splitting thread :

  void splittingThread(SplitingThreadParam &param)
  
  {
  
      Converter *pConverter = (Converter*)param.parentParamClass;
  
      BYTE* cropRGBABuff[4];
  
      for (int i = 0; i < 4; i++)
  
          cropRGBABuff[i] = new BYTE[SINGLE_SIDE_WIDTH*SINGLE_SIDE_HEIGHT * 4];
  
      //Split:
  
      pConverter->OpenCVSplittingSideFrame(param.inputBuff, SIDE_FRAME_WIDTH, SIDE_FRAME_HEIGHT, SINGLE_SIDE_WIDTH, SINGLE_SIDE_HEIGHT, cropRGBABuff[0], cropRGBABuff[1], cropRGBABuff[2], cropRGBABuff[3]);
  
  
  
      //Convert to I420:
  
      pConverter->ConvertColor(cropRGBABuff[0], param.out1Buff, SINGLE_SIDE_WIDTH, SINGLE_SIDE_HEIGHT, AV_PIX_FMT_RGBA, AV_PIX_FMT_YUV420P);
  
      pConverter->ConvertColor(cropRGBABuff[1], param.out2Buff, SINGLE_SIDE_WIDTH, SINGLE_SIDE_HEIGHT, AV_PIX_FMT_RGBA, AV_PIX_FMT_YUV420P);
  
      pConverter->ConvertColor(cropRGBABuff[2], param.out3Buff, SINGLE_SIDE_WIDTH, SINGLE_SIDE_HEIGHT, AV_PIX_FMT_RGBA, AV_PIX_FMT_YUV420P);
  
      pConverter->ConvertColor(cropRGBABuff[3], param.out4Buff, SINGLE_SIDE_WIDTH, SINGLE_SIDE_HEIGHT, AV_PIX_FMT_RGBA, AV_PIX_FMT_YUV420P);
  
  
  
      //pConverter->DumpBufferData(param.out1Buff, SINGLE_SIDE_WIDTH, SINGLE_SIDE_HEIGHT, 1);
  
  }

  My Splitting function :

  void Converter::OpenCVSplittingSideFrame(BYTE *input, unsigned int srcWidth, unsigned int srcHeight,
  
      unsigned int cropWidth, unsigned int cropHeight,
  
      BYTE *out1, BYTE *out2, BYTE *out3, BYTE *out4)
  
  {
  
      Mat nv12Mat;
  
      Mat rgbMat;
  
      Mat rgbCropImg[4];
  
  
  
      //Create YUV Matrix:
  
      nv12Mat.create(srcHeight * 3 / 2, srcWidth, CV_8UC1);
  
      memcpy(nv12Mat.data, input, srcWidth * srcHeight * 3 / 2);
  
      //Create RGB matrix:
  
      rgbMat.create(srcHeight, srcWidth, CV_8UC4);
  
  
  
      //Convert YUV to RGB:
  
      cvtColor(nv12Mat, rgbMat, COLOR_YUV2RGBA_NV21);
  
  
  
      //Crop:
  
      CropMatrix(rgbMat, rgbCropImg[0], 0, 0, cropWidth, cropHeight);
  
      CropMatrix(rgbMat, rgbCropImg[1], 0, SIDE_CROP_HEIGHT, cropWidth, cropHeight);
  
      CropMatrix(rgbMat, rgbCropImg[2], SIDE_CROP_WIDTH, 0, cropWidth, cropHeight);
  
      CropMatrix(rgbMat, rgbCropImg[3], SIDE_CROP_WIDTH, SIDE_CROP_HEIGHT, cropWidth, cropHeight);
  
  
  
      //Copy from matrix to output buffer:
  
      memcpy(out1, rgbCropImg[0].data, cropWidth * cropHeight * 4);
  
      memcpy(out2, rgbCropImg[1].data, cropWidth * cropHeight * 4);
  
      memcpy(out3, rgbCropImg[2].data, cropWidth * cropHeight * 4);
  
      memcpy(out4, rgbCropImg[3].data, cropWidth * cropHeight * 4);
  
  }

  Convert color range to RGBA to I420 function :

  int Converter::ConvertColor(unsigned char *srcbuff, unsigned char *targetBuff, unsigned int width, unsigned int height, AVPixelFormat srcFormat, AVPixelFormat targetFormat)
  
  {
  
  
  
      int ret = 0;
  
      //create the conversion context
  
      struct SwsContext *swsContext = sws_getContext(
  
          width, height, srcFormat,    // AV_PIX_FMT_NV12, // source 
  
          width, height, targetFormat, // AV_PIX_FMT_YUV420P,    // target: GL_RGBA
  
          SWS_BILINEAR,
  
          NULL, NULL, NULL
  
          );
  
      if (swsContext < 0) {
  
          return -1;
  
      }
  
  
  
      // allocate frame 
  
      AVFrame *pSrcFrame = av_frame_alloc();
  
      AVFrame *pTargetFrame = av_frame_alloc();
  
  
  
      // source frame
  
      avpicture_fill(
  
          (AVPicture*)pSrcFrame,
  
          (const uint8_t *)srcbuff,
  
          srcFormat,
  
          width,
  
          height
  
          );
  
  
  
      // target frame
  
      if (srcFormat != AV_PIX_FMT_RGBA)
  
          avpicture_fill(
  
          (AVPicture*)pTargetFrame,
  
          (const uint8_t *)targetBuff,
  
          targetFormat,
  
          width,
  
          height
  
          );
  
      else
  
          avpicture_fill(
  
          (AVPicture*)pTargetFrame,
  
          (const uint8_t *)targetBuff,
  
          targetFormat,
  
          1936,
  
          872
  
          );
  
      ret = sws_scale(
  
          swsContext,             // sws context
  
          // source
  
          pSrcFrame->data,        // source
  
          pSrcFrame->linesize,    // source stride
  
          0,                      // slice Y
  
          height,                 // slice H
  
          // target
  
          pTargetFrame->data,        // target
  
          pTargetFrame->linesize     // target stride
  
          );
  
  
  
      av_frame_free(&pSrcFrame);
  
      av_frame_free(&pTargetFrame);
  
      sws_freeContext(swsContext);
  
  
  
      return ret;
  
  }

  My encoder setting :

  EncoderContext encoderCtx;
  
  encoderCtx.vCodecId = VIDEO_CODEC_H264;
  
  encoderCtx.aCodecId = AUDIO_CODEC_NONE;
  
  encoderCtx.numAudioChannel = 0;
  
  encoderCtx.eFormat = BEYOND_ENCODER_MP4;
  
  encoderCtx.videoBitrate = 10000000;
  
  encoderCtx.audioBitrate = 0;
  
  encoderCtx.audioBitrate = 0;
  
  encoderCtx.timeBaseDen = config->time_base_den;
  
  encoderCtx.timeBaseNum = config->time_base_num;
  
  encoderCtx.width = SINGLE_SIDE_WIDTH;
  
  encoderCtx.height = SINGLE_SIDE_HEIGHT;
  
  CString filePath(config->savePath.c_str());
  
  CString filePath2;
  
  filePath2.Format(_T("%s_%d.mp4"), filePath, i);
  
  CT2A multibyte(filePath2, CP_UTF8);
  
  encoderCtx.outFileName = multibyte;

  Thanks all for your help !

• LGPL Xuggle : ConverterFactory.createConverter not supported

  24 mars 2013, par jlengrand

  I am using Xuggle to create an mpeg-ts stream.
  Everything was working nice until I realized that I was using the GPL version of ffmpeg, which I cannot.

  So I recompiled the whole library, and removed the —enable-nonfree and —enable-gpl flags from ffmpeg.

  The thing is that now my code would throw an error at me.

  The lines in question are :

  image = new BufferedImage(width, height, BufferedImage.TYPE_3BYTE_BGR);
  
  IVideoPicture outFrame = ConverterFactory.createConverter(image, IPixelFormat.Type.YUV420P).toPicture(image, timeStamp);

  And the error I get is

  java.lang.UnsupportedOperationException: Converter class com.xuggle.xuggler.video.BgrConverter constructor failed with: java.lang.IllegalArgumentException: IVideoResampler not supported in this build
  
      at com.xuggle.xuggler.video.ConverterFactory.createConverter(ConverterFactory.java:347)

  The thing is that we were using the same code with the 3.3 version of Xuggle, and even though ffmpeg was throwing warnings at us saying that we had no hardware acceleration, everything was working fine.

  So did something important change between those versions ?
  And how can I overcome this ? I searched in the archive but couldn't find related issues so far.

  Thanks !