Recherche avancée

Sur d’autres sites (6103)

• aac : Split function to parse ADTS header data into public and private part

  26 avril 2017, par Diego Biurrun

  aac : Split function to parse ADTS header data into public and private part
  This makes the currently semi-public avpriv_aac_parse_header() function
  
  private to libavcodec and adds a proper public API function to return
  
  the parts of the ADTS header required in libavformat.
  

  • [DBH] configure
  • [DBH] doc/developer.texi
  • [DBH] libavcodec/Makefile
  • [DBH] libavcodec/aac_adtstoasc_bsf.c
  • [DBH] libavcodec/aac_parser.c
  • [DBH] libavcodec/aacdec.c
  • [DBH] libavcodec/adts_header.c
  • [DBH] libavcodec/adts_header.h
  • [DBH] libavcodec/adts_parser.c
  • [DBH] libavcodec/adts_parser.h
  • [DBH] libavformat/spdifdec.c
  • [DBH] libavformat/spdifenc.c

• ffmpeg, first second of cut video part freezed

  12 avril 2017, par Hugo

  I need to split video produced by raspivid ( rasbperry pi rasbian video output h264 ) into two parts and second part is always starting with about one second long frozen image. Based on what i found problem is in keyframes.

  I did try following solutions i found :

  • use -ss parameter before -i which should use only keyframes to split - not working, result is same when -ss is before or after -i
  • i did try to extract exact keyframe using ffprobe and to use value pkt_dts_time to specify exact microsecond to split video - again result is same, video is starting with frozen image

  I also try to use -vcodec copy option for both methods but again - result is exactly the some. Only difference i see is delay how long is inicial picture frozen.

• record every step of selenium test case to a video proplem

  16 février 2017, par user7574253

  i use pytest and selenium to test web site, i am writing a base class, every test case run should be recording to a video, so i can see the fail test case in the video and know what’s happing. below is my code

  ffmpeg_command = os.environ['ffmpeg_command']
  
  video_base_path = os.environ['video_base_path']
  
  
  
  
  
  class BaseRecordTest(BaseTest):
  
      def set_up_00(self, method):
  
  
  
          # video save path should be "base path" + "test case module path" +      "test case class" + "test case name"
  
          module_name = self.__module__
  
          if module_name == '__main__':
  
               filename = sys.modules[self.__module__].__file__
  
               module_name = os.path.splitext(os.path.basename(filename))[0]
  
          module_path = module_name.replace('.', '\\')
  
          dir_name = video_base_path + module_path
  
          if not os.path.exists(dir_name):
  
              os.makedirs(dir_name)
  
  
  
          video_path = dir_name + '\{}-{}.avi'.format(self.__class__.__name__,   method.__name__)
  
  
  
          if os.path.exists(video_path):
  
              os.remove(video_path)
  
  
  
          process = Popen(
  
          "{} {}".format(ffmpeg_command, video_path),
  
          shell=True, stdout=PIPE, stderr=PIPE)
  
          self.__process = process
  
  
  
      def tear_down_99(self, method):
  
          # self.__process.terminate()
  
          # os.killpg(self.__process.pid, signal.CTRL_C_EVENT)
  
          os.kill(self.__process.pid, signal.CTRL_C_EVENT)
  
          # terminate(self.__process)

  my system is windows 10 , i have try send signal to kill the recording process, but if use Popen terminate or kill method , the ffmpeg_command does not real stop, if send ctrl_c_event signal, video can be saved but the next test case not execute forever, so i want to ask there is a way to resolve this, that can execute each case and each case execute process can be recording to a vedio, thanks !

  ffmpeg cammand like that :

  D:\\ffmpeg-20170214-8fa18e0-win64-static\\bin\\ffmpeg -video_size 1366x768 -f gdigrab -offset_x 0 -offset_y 0 -framerate 60 -i desktop