Recherche avancée

Sur d’autres sites (11154)

• Building FFMPEG library for iOS5.1 ARMv7

  25 octobre 2012, par Jimmy

  I'm trying to use the FFMPEG library in an XCode 4.5.1 project. And I'm trying to build it for ARMv7

  But I'm running into some major issues, I'm hoping that someone can walk me through how to do it ? I feel a little sketched out reading material thats a couple months old. I'm not aware of the changes.

  New Update (after a lot of time)

  I want to help everyone out that is having this problem because its so frustrating. heres how you do it. I've pasted the configuration stage. This is my ./configure

  ./configure --disable-doc --disable-ffmpeg --disable-ffplay --disable-ffserver --enable-cross-compile --arch=arm --target-os=darwin
  
     --cc=/Applications/Xcode.app/Contents/Developer/Platforms/iPhoneOS.platform/Developer/usr/llvm-gcc-4.2/bin/arm-apple-darwin10-llvm-gcc-4.2
  
     --as='gas-preprocessor/gas-preprocessor.pl /Applications/Xcode.app/Contents/Developer/Platforms/iPhoneOS.platform/Developer/usr/llvm-gcc-4.2/bin/arm-apple-darwin10-llvm-gcc-4.2'
  
     --sysroot=/applications/xcode.app/contents/Developer/Platforms/iPhoneOS.platform/Developer/SDKs/iPhoneOS5.1.sdk
  
     --cpu=cortex-a8
  
     --extra-ldflags='-isysroot /applications/xcode.app/contents/Developer/Platforms/iPhoneOS.platform/Developer/SDKs/iPhoneOS5.1.sdk'
  
     --enable-pic
  
     --disable-bzlib --disable-gpl --disable-shared --enable-static --disable-mmx --disable-debug --disable-neon
  
     --extra-cflags='-pipe -Os -gdwarf-2 -isysroot /applications/xcode.app/contents/Developer/Platforms/iPhoneOS.platform/Developer/SDKs/iPhoneOS5.1.sdk -m${thumb_opt:-no-thumb} -mthumb-interwork'

  This ended up working for me. I had the following problems.

  • I had to download ( https://github.com/yuvi/gas-preprocessor ) copy the file gas-preprocessor.pl at /usr/local/bin. Set permissions to read write (777)
  • Make sure I'm using the right GCC compiler : /Applications/Xcode.app/Contents/Developer/Platforms/iPhoneOS.platform/Developer/usr/llvm-gcc-4.2/bin/arm-apple-darwin10-llvm-gcc-4.2
  • Make sure I'm using the right SDK : /applications/xcode.app/contents/Developer/Platforms/iPhoneOS.platform/Developer/SDKs/iPhoneOS5.1.sdk
  • —extra-cflags="-arch armv7" causes bugs in the configure. take it out. This was causing my gcc compile errors. (specifically this one : error : unrecognized command line option “-arch”.)

  I'll let you guys know how much further I get.

• Concatenate list of streams with ffmpeg

  16 novembre 2019, par oioioi

  I’d like to write python code, automatically rearranging the scenes of a video. How do I use ffmpeg-python to concatenate a list of n streams/scenes ? What is the easiest way to merge those scenes with their audio counterpart ?
  Unfortunately, I couldn’t figure out how to do it properly.

  The status quo :

  import os
  
  
  
  import scenedetect
  
  from scenedetect.video_manager import VideoManager
  
  from scenedetect.scene_manager import SceneManager
  
  from scenedetect.frame_timecode import FrameTimecode
  
  from scenedetect.stats_manager import StatsManager
  
  from scenedetect.detectors import ContentDetector
  
  
  
  import ffmpeg
  
  
  
  STATS_FILE_PATH = 'stats.csv'
  
  
  
  def main():
  
  
  
      for root, dirs, files in os.walk('material'):
  
          for file in files:
  
              file = os.path.join(root, file)
  
              print(file)
  
              video_manager = VideoManager([file])
  
              stats_manager = StatsManager()
  
              scene_manager = SceneManager(stats_manager)
  
              scene_manager.add_detector(ContentDetector())
  
              base_timecode = video_manager.get_base_timecode()
  
              end_timecode = video_manager.get_duration()
  
  
  
              start_time = base_timecode
  
              end_time = base_timecode + 100.0
  
              #end_time = end_timecode[2]
  
  
  
              video_manager.set_duration(start_time=start_time, end_time=end_time)
  
              video_manager.set_downscale_factor()
  
              video_manager.start()
  
              scene_manager.detect_scenes(frame_source=video_manager)
  
  
  
              scene_list = scene_manager.get_scene_list(base_timecode)
  
  
  
              print('List of scenes obtained:')
  
              for i, scene in enumerate(scene_list):
  
                  print('    Scene %2d: Start %s / Frame %d, End %s / Frame %d' % (
  
                      i+1,
  
                      scene[0].get_timecode(), scene[0].get_frames(),
  
                      scene[1].get_timecode(), scene[1].get_frames(),))
  
  
  
                  start = scene[0].get_frames()
  
                  end = scene[1].get_frames()
  
  
  
                  print(start)
  
                  print(end)
  
  
  
                  (
  
                      ffmpeg
  
                      .input(file)
  
                      .trim(start_frame=start, end_frame=end)
  
                      .setpts ('PTS-STARTPTS')
  
                      .output('scene %d.mp4' % (i+1)) 
  
                      .run()
  
                  )
  
  
  
              if stats_manager.is_save_required():
  
                  with open(STATS_FILE_PATH, 'w') as stats_file:
  
                      stats_manager.save_to_csv(stats_file, base_timecode)
  
  
  
  if __name__ == "__main__":
  
      main()

• How to change mjpeg to yuyv422 from a webcam to a v4l2loopback ?

  3 janvier 2020, par Dev Null

  Backstory : One livestreaming site I use isn’t smart enough to detect the capabilities of my webcam (Logitech Brio, 4k), and instead just uses the default frames per second settings, which is 5fps.

  (full solution walk-through in the answer)

  The best solution I could think of (besides changing livestream providers) was to create a loopback virtual webcam using v4l2loopback that I could force to have the exact settings I wanted to use on that livestream site.

  For the brio, the higher frame rates come with mjpeg, not the default yuyv.

  Problem 1 :

  I could easily read mjpeg, but unfortunately kept banging my head against the wall because v4l2loopback evidently only wanted yuyv.

  I tried things like :

  ffmpeg -f v4l2              \
  
         -input_format mjpeg  \
  
         -framerate 30        \
  
         -video_size 1280x720 \
  
         -i /dev/video0       \
  
         -vcodec copy         \
  
         -f v4l2 /dev/video6

  and

  ffmpeg -f v4l2              \
  
         -input_format mjpeg  \
  
         -framerate 30        \
  
         -video_size 1280x720 \
  
         -i /dev/video0       \
  
         -vcodec yuyv422      \ # this line changed (even tried "copy")
  
         -f v4l2 /dev/video6

  But they wouldn’t work. I got errors like :

  Unknown V4L2 pixel format equivalent for yuvj422p

  and

  ...deprecated pixel format used, make sure you did set range correctly...

  ...V4L2 output device supports only a single raw video stream...

  Eventually I got this to work :

  ffmpeg -f v4l2              \
  
         -input_format mjpeg  \
  
         -framerate 30        \
  
         -video_size 1280x720 \
  
         -i /dev/video0       \
  
         -pix_fmt yuyv422     \ # The winning entry
  
         -f v4l2 /dev/video6

  Problem 2

  The next problem was getting chrome to see the virtual webcam. It worked correctly with guvcview, and on firefox I could use webcam testing sites and it would pick the virtual camera up without a problem.

  Turns out google, in it’s overly-protective nature (while it’s siphoning off all our data, btw), doesn’t want to use webcams that can be read and written to.

  So when starting v4l2loopback you have to tell it to announce that it’s "read only" to consumers like chrome.

  Here’s the exact modprobe I use that works :

  sudo modprobe v4l2loopback devices=1 exclusive_caps=1