Recherche avancée

Sur d’autres sites (12753)

• Where is moviepy getting the video fps from ?

  7 février 2017, par Gloin

  I am using the Python 3 moviepy module for video editing, and I have a few videos that are taken in slow motion. When imported into moviepy, they are massively sped up, and then sit on the last frame for the rest of their duration. Note that, the videos are supposed to be normal for the first couple and last couple of seconds, then slow in the middle.

  Unfortunately, I cannot provide the actual video for you to test with, but here is the relevant metadata (fetched with the command ffprobe -v quiet -print_format json -show_format -show_streams slo-mo_movie.mov)

  "r_frame_rate": "240/1",                               
  
  "avg_frame_rate": "1679400/39481",
  
  "time_base": "1/2400",

  For comparison, here is the equivalent metadata from a video taken, I think, from the same phone, but without slo-mo :

  "r_frame_rate": "30/1",
  
  "avg_frame_rate": "143160/4771",
  
  "time_base": "1/600",

  I can import the videos in to moviepy with clip = VideoFileClip("path/to/file.mp4"), and then for each run print(clip.fps). The first video prints 2400 (not a typo from me !), and the second 30.

  Here is the moviepy code (in moviepy/video/io/ffmpeg_reader.py) at line 293) that gets the fps :

  # Get the frame rate. Sometimes it's 'tbr', sometimes 'fps', sometimes
  
  # tbc, and sometimes tbc/2...
  
  # Current policy: Trust tbr first, then fps. If result is near from x*1000/1001
  
  # where x is 23,24,25,50, replace by x*1000/1001 (very common case for the fps).
  
  
  
  try:
  
      match = re.search("( [0-9]*.| )[0-9]* tbr", line)
  
      tbr = float(line[match.start():match.end()].split(' ')[1])
  
      result['video_fps'] = tbr
  
  
  
  except:
  
      match = re.search("( [0-9]*.| )[0-9]* fps", line)
  
      result['video_fps'] = float(line[match.start():match.end()].split(' ')[1])
  
  
  
  
  
  # It is known that a fps of 24 is often written as 24000/1001
  
  # but then ffmpeg nicely rounds it to 23.98, which we hate.
  
  coef = 1000.0/1001.0
  
  fps = result['video_fps']
  
  for x in [23,24,25,30,50]:
  
      if (fps!=x) and abs(fps - x*coef) < .01:
  
          result['video_fps'] = x*coef
  
  
  
  if check_duration:
  
      result['video_nframes'] = int(result['duration']*result['video_fps'])+1
  
      result['video_duration'] = result['duration']
  
  else:
  
      result['video_nframes'] = 1
  
      result['video_duration'] = None
  
  # We could have also recomputed the duration from the number
  
  # of frames, as follows:
  
  # >>> result['video_duration'] = result['video_nframes'] / result['video_fps']

  If I set the slo-mo video’s fps using moviepy to 24, it outputs it the same (very fast, then still on the last frame), but if I set the slo-mo video’s fps to 20, then it outputs it correctly.

  Obviously video players like VLC player and Quicktime can correctly work out what frame speed to play, but moviepy/ffmpeg fails. Moviepy/ffmpeg is getting the wrong fps from somewhere.

  So, how can I get moviepy to automatically output them as they are supposed to be without human trial and error like above ?

• FFMPEG Keyframes Exception

  3 avril 2017, par newToRacket

  I was hoping someone could help me, when pulling keyframes from a .mp4 file using FFMPEG in python I get this error :

       C:\Python27\python.exe "C:/Coll. Detection/UAS_Detection/GUI.py"
  
  in GUI main
  
  ('filePath', PyQt4.QtCore.QString(u'C:\\Users\\Razor\\Downloads\\video_test_512kb.mp4'), '\n')
  
  File Path -> C:\Users\Razor\Downloads\video_test_512kb.mp4
  
  SingleDetect C:\Users\Razor\Downloads\video_test_512kb.mp4
  
  ffmpeg version N-84679-gd65b595 Copyright (c) 2000-2017 the FFmpeg developers
  
    built with gcc 6.3.0 (GCC)
  
    configuration: --enable-gpl --enable-version3 --enable-cuda --enable-cuvid --enable-d3d11va --enable-dxva2 --enable-libmfx --enable-nvenc --enable-avisynth --enable-bzlib --enable-fontconfig --enable-frei0r --enable-gnutls --enable-iconv --enable-libass --enable-libbluray --enable-libbs2b --enable-libcaca --enable-libfreetype --enable-libgme --enable-libgsm --enable-libilbc --enable-libmodplug --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libopenh264 --enable-libopenjpeg --enable-libopus --enable-librtmp --enable-libsnappy --enable-libsoxr --enable-libspeex --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvo-amrwbenc --enable-libvorbis --enable-libvpx --enable-libwavpack --enable-libwebp --enable-libx264 --enable-libx265 --enable-libxavs --enable-libxvid --enable-libzimg --enable-lzma --enable-zlib
  
    libavutil      55. 51.100 / 55. 51.100
  
    libavcodec     57. 86.103 / 57. 86.103
  
    libavformat    57. 67.100 / 57. 67.100
  
    libavdevice    57.  3.101 / 57.  3.101
  
    libavfilter     6. 78.100 /  6. 78.100
  
    libswscale      4.  3.101 /  4.  3.101
  
    libswresample   2.  4.100 /  2.  4.100
  
    libpostproc    54.  2.100 / 54.  2.100
  
  Input #0, mov,mp4,m4a,3gp,3g2,mj2, from 'C:\Users\Razor\Downloads\video_test_512kb.mp4':
  
    Metadata:
  
      major_brand     : isom
  
      minor_version   : 512
  
      compatible_brands: mp41
  
      creation_time   : 1970-01-01T00:00:00.000000Z
  
      title           : test file mp4 - http://www.archive.org/details/Pbtestfilemp4videotestmp4
  
      encoder         : Lavf51.10.0
  
      comment         : license:  http://creativecommons.org/licenses/by-nc-sa/2.5/
  
    Duration: 00:00:16.27, start: 0.000000, bitrate: 562 kb/s
  
      Stream #0:0(und): Video: h264 (Constrained Baseline) (avc1 / 0x31637661), yuv420p, 320x240 [SAR 1:1 DAR 4:3], 512 kb/s, 15 fps, 15 tbr, 15 tbn, 30 tbc (default)
  
      Metadata:
  
        creation_time   : 1970-01-01T00:00:00.000000Z
  
        handler_name    : VideoHandler
  
      Stream #0:1(und): Audio: aac (LC) (mp4a / 0x6134706D), 44100 Hz, stereo, fltp, 45 kb/s (default)
  
      Metadata:
  
        creation_time   : 1970-01-01T00:00:00.000000Z
  
        handler_name    : SoundHandler
  
  Stream mapping:
  
    Stream #0:0 -> #0:0 (h264 (native) -> mjpeg (native))
  
  Press [q] to stop, [?] for help
  
  [swscaler @ 0000000003500e00] deprecated pixel format used, make sure you did set range correctly
  
  Output #0, image2, to 'tmp/%d.jpeg':
  
    Metadata:
  
      major_brand     : isom
  
      minor_version   : 512
  
      compatible_brands: mp41
  
      comment         : license:  http://creativecommons.org/licenses/by-nc-sa/2.5/
  
      title           : test file mp4 - http://www.archive.org/details/Pbtestfilemp4videotestmp4
  
      encoder         : Lavf57.67.100
  
      Stream #0:0(und): Video: mjpeg, yuvj420p(pc), 320x240 [SAR 1:1 DAR 4:3], q=2-31, 200 kb/s, 15 fps, 15 tbn, 15 tbc (default)
  
      Metadata:
  
        creation_time   : 1970-01-01T00:00:00.000000Z
  
        handler_name    : VideoHandler
  
        encoder         : Lavc57.86.103 mjpeg
  
      Side data:
  
        cpb: bitrate max/min/avg: 0/0/200000 buffer size: 0 vbv_delay: -1
  
  2017-04-02 18:36:07.832000
  
  Traceback (most recent call last):
  
    File "Detection.py", line 59, in <module>
  
      draw_detections(frame, found)
  
    File "Detection.py", line 32, in draw_detections
  
      result = Result.objects.create(picture=img, x=x, y=y, w=w, h=h)
  
  AttributeError: type object 'Result' has no attribute 'objects'
  
  frame=   21 fps=0.0 q=19.3 Lsize=N/A time=00:00:16.06 bitrate=N/A speed= 191x    
  
  video:160kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: unknown
  
  
  
  Process finished with exit code 0
  
  </module>

  I’m trying to run these keyframes through an opencv human shape recognition api. It saves close to 300 keyframes to the folder I specified but for some reason the program just hangs afterwords instead of running the open cv hogger api on the newly saved keyframes. Will post script if it will help. Thank you for any help you can provide !!!

• swscale : use dithering in DITHER_COPY only if not set -sws_dither none

  24 octobre 2017, par Mateusz

  swscale : use dithering in DITHER_COPY only if not set -sws_dither none
  This patch uses dithering in DITHER_COPY macro only if
  
  it was not used option '-sws_dither none'.
  
  With option '-sws_dither none' it uses downshift.
  For human eye dithering is OK, for video codecs not necessarily.
  
  If user don't want to use dithering, we should respect that.
  Signed-off-by : Mateusz Brzostek <mateuszb@poczta.onet.pl>
  
  Signed-off-by : Michael Niedermayer <michael@niedermayer.cc>
  

  • [DH] libswscale/swscale_unscaled.c