Recherche avancée

Sur d’autres sites (6935)

• C++ smart pointers to FFmpeg objects

  15 octobre 2024, par Elija

  Can I create and use C++ smart pointers for different pointer types from FFmpeg ?

  


  

  1. "AVCodecContext *" which is used only as a pointer in all functions except deallocation.

  


  


  Alloc :

  


  AVCodecContext *avcodec_alloc_context3(const AVCodec *codec);


  


  Free :

  


  void avcodec_free_context(AVCodecContext **avctx);


  


  Use :

  


  int avcodec_open2(AVCodecContext *avctx, const AVCodec *codec, AVDictionary **options);


  


  Then the smart pointer :

  


  std::shared_ptr<avcodeccontext> av_codec_context(avcodec_alloc_context3(av_codec),&#xA;[](AVCodecContext* _context)&#xA;{&#xA;  if (_context) avcodec_free_context(&amp;_context);&#xA;});&#xA;avcodec_open2(av_codec_context.get(), av_codec, NULL)&#xA;</avcodeccontext>

  


  Is this correct ?

  


  

  2. "AVDictionary **" which is used in all functions only as a pointer to a pointer.

  


  


  Alloc and use :

  


  int av_dict_set(AVDictionary **pm, const char *key, const char *value, int flags);


  


  where pm is a Pointer to a pointer to a dictionary struct. If *pm is NULL a dictionary struct is allocated and put in *pm.

  


  Free :

  


  void av_dict_free(AVDictionary **m);


  


  Then a smart pointer :

  


  std::shared_ptr av_dict(new (AVDictionary*),
[](AVDictionary** _dict)
{
  if (_dict)
  {
    if(*_dict)
      av_dict_free(_dict);
    delete _dict;
  }
});
av_dict_set(av_dict.get(), "key", "value", 0);


  


  Is this correct ?

  


  

  3. "AVFormatContext *" which is used both as a pointer and as a pointer to a pointer.

  


  


  Alloc :

  


  AVFormatContext *avformat_alloc_context(void);


  


  Free :

  


  void avformat_free_context(AVFormatContext *s);


  


  Use :

  


  int avformat_find_stream_info(AVFormatContext *ic, AVDictionary **options);


  


  or

  


  int avformat_open_input(AVFormatContext **ps, const char *url, const AVInputFormat *fmt, AVDictionary **options);


  


  where ps is a Pointer to user-supplied AVFormatContext (allocated by avformat_alloc_context). May be a pointer to NULL, in which case an AVFormatContext is allocated by this function and written into ps.

  


  Then a smart pointer :

  


  std::shared_ptr<avformatcontext> av_format_context(avformat_alloc_context(),&#xA;[](AVFormatContext* _context)&#xA;{&#xA;  if(_context)&#xA;    avformat_free_context(_context);&#xA;});&#xA;avformat_find_stream_info(av_format_context.get(), NULL);&#xA;</avformatcontext>

  


  Is this correct ? But how can I use it with the avformat_open_input() function, which needs a pointer to a pointer and may want to create an object by this pointer ?

  


• Smart y coordinate to make vertical alignment for the text with typewriting effect

  31 août 2021, par Макс Шульдинер

  I'm doing animation for the typewriting effect. Here is my ffmpeg string :

  


  -i ffmpeg_inputs/output-onlinegiftools.gif  -vf "[in]drawtext=fonts/RobotoMono-Regular.ttf:text='h':fontcolor=orange:fontsize=35:x=(w-text_w)/2-200+0:y=h-th-400:enable='between(t,0.00, 7.80)', drawtext =fonts/RobotoMono-Regular.ttf:text = 'i':fontcolor=orange:fontsize=35:x=(w-text_w)/2-200+25:y=h-th-400:enable='between(t,0.80, 7.80)', drawtext =fonts/RobotoMono-Regular.ttf:text = 'g':fontcolor=orange:fontsize=35:x=(w-text_w)/2-200+75:y=h-th-400:enable='between(t,1.60, 7.80)', drawtext =fonts/RobotoMono-Regular.ttf:text = 'u':fontcolor=orange:fontsize=35:x=(w-text_w)/2-200+100:y=h-th-400:enable='between(t,2.40, 7.80)', drawtext =fonts/RobotoMono-Regular.ttf:text = 'y':fontcolor=orange:fontsize=35:x=(w-text_w)/2-200+125:y=h-th-400:enable='between(t,3.20, 7.80)', drawtext =fonts/RobotoMono-Regular.ttf:text = 's':fontcolor=orange:fontsize=35:x=(w-text_w)/2-200+150:y=h-th-400:enable='between(t,4.00, 7.80)'[out]" ffmpeg_outputs/test2.gif -y 


  


  Here is the results with different y values :

  


  

  


  

  


  As i understand, to make smooth sentence, for some letters i need top vertical alignment, and for others i need bottom vertical alignment. How can i make this "smart alignment", or the only one method is to hardcode y values for different letters ?

  


• FFmpeg : H.265/HEVC - does not work on Sony Smart-TV if you change encoder

  25 juillet 2019, par oigen90

  We have such Sony Bravia Smart-TV devices :

  • Sony KD-49XD7005 (2016) - Android
  • Sony KD-49XE7096 (2017) - Linux

  Those devices does have HEVC support, as well as 4K support.

  Also, we have a one minute Apple ProRes 4K video file. I need to transcode it into HEVC with some simple params, create DASH manifest and stream it into HTML5-player.
  The problem is that if I use the regular ’libx265’ codec passing it after "-c:v" - the video does not play on Sony Bravia (just freezes on ’loadstart’ player’s event). But if I use ’hevc_nvenc’ - it works ! What’s the reason of such behavior ? Both of codecs should work identically, hevc_nvenc just should be faster and was made to work on nvidia graphic cards (that’s all I know about those codecs’ difference so far). But in fact we have working stream encoded into hevc_nvenc and invalid stream encoded into libx265.

  See my FFmpeg command below.

  Which thoughts do you have ? If you want me to provide some more info - feel free to ask.

  ffmpeg -i input.mov -c:v hevc_nvenc -pix_fmt yuv420p -b:v 12000k -maxrate 14000k -bufsize 6000k -c:a copy -f mp4 output.mp4