Recherche avancée

Sur d’autres sites (3819)

• build : add support for building .cu files via nvcc

  11 mai 2017, par Timo Rothenpieler

  build : add support for building .cu files via nvcc
  Original work by Yogender Gupta <ygupta@nvidia.com>
  

  • [DH] .gitignore
  • [DH] Makefile
  • [DH] compat/cuda/ptx2c.sh
  • [DH] configure
  • [DH] ffbuild/common.mak

• Best Settings for H264_Nvenc To Minimize Latency ? (FFMPEG)

  1er novembre 2022, par M. Ying

  I'm currently using Libavcodec to encode video frames using H.264. Because I'm streaming this video frames after they're encoded, minimizing latency is crucial to me. My current settings are :

  



  avcodec_register_all();
codec = avcodec_find_encoder(AV_CODEC_ID_H264);

// allocate and set ffmpeg context
context = avcodec_alloc_context3(encoder->codec);
context->bit_rate = bitrate;
context->width = out_width;
context->height = out_height;
context->time_base.num = 1;
context->time_base.den = 30;
context->gop_size = 1; // send SPS/PPS headers every packet
context->max_b_frames = 0;
context->pix_fmt = AV_PIX_FMT_YUV420P;

// set encoder parameters to max performance
av_opt_set(context->priv_data, "preset", "ultrafast", 0);
av_opt_set(context->priv_data, "tune", "zerolatency", 0);

// open capture encoder
avcodec_open2(context, codec, NULL);


  



  These settings work well, but I am trying to switch to hardware-based encoding to take the workload off my CPU. I currently have an NVIDIA GPU, so I tried using the following settings with h264_nvenc :

  



  codec = avcodec_find_encoder_by_name("h264_nvenc");

// allocate and set ffmpeg context
context = avcodec_alloc_context3(encoder->codec);
context->dct_algo = FF_DCT_FASTINT;
context->bit_rate = bitrate;
context->width = out_width;
context->height = out_height;
context->time_base.num = 1;
context->time_base.den = 30;
context->gop_size = 1; // send SPS/PPS headers every packet
context->max_b_frames = 0;
context->pix_fmt = AV_PIX_FMT_YUV420P;

// set encoder parameters to max performancen
av_opt_set(context->priv_data, "preset", "llhq", 0);
av_opt_set(context->priv_data, "tune", "zerolatency", 0);


  



  The issue is, I noticed that the latency with h264_nvenc is significantly more than the latency with AV_CODEC_ID_H264 (the software-based version). I think that my settings or setup of h264_nvenc must be wrong, since GPU-based encoding should be faster than software-based encoding. Could anyone point me in the right direction ? Thanks so much !

  


• avcodec/cuvid : Add support for P010/P016 as an output surface format

  22 novembre 2016, par Philip Langdale

  avcodec/cuvid : Add support for P010/P016 as an output surface format
  The nvidia 375.xx driver introduces support for P016 output surfaces,
  
  for 10bit and 12bit HEVC content (it’s also the first driver to support
  
  hardware decoding of 12bit content).
  The cuvid api, as far as I can tell, only declares one output format
  
  that they appear to refer to as P016 in the driver strings. Of course,
  
  10bit content in P016 is identical to P010, and it is useful for
  
  compatibility purposes to declare the format to be P010 to work with
  
  other components that only know how to consume P010 (and to avoid
  
  triggering swscale conversions that are lossy when they shouldn’t be).
  For simplicity, this change does not maintain the previous ability
  
  to output dithered NV12 for 10/12 bit input video - the user will need
  
  to update their driver to decode such videos.
  

  • [DH] compat/cuda/dynlink_cuviddec.h
  • [DH] libavcodec/cuvid.c
  • [DH] libavcodec/version.h