Recherche avancée

Sur d’autres sites (13007)

• Revision c37ecba7f7 : configure : factorize gcc machine option checks check_gcc_machine_option() repla

  19 mars 2014, par James Zern

  Changed Paths :
   Modify /build/make/configure.sh

  
  
  configure : factorize gcc machine option checks

  check_gcc_machine_option() replaces individual -m* checks

  Change-Id : I0f4a82020c0541b99209321907e80e071d1245e1

• Simplest way to do do video editing in C++ ?

  23 décembre 2019, par CaptainCodeman

  I have a video file (approx 30,000 frames) and want to do some processing on the individual frames with a C++ program I’ve written.

  The simplest method would be to extract the frames using ffmeg, do the processing, and then encode the video again. However, this would require a few hundred gigabytes of disk space. Is there a way to stream it ?

  Or is there some library that lets me just open a video, alter the frames, and re-encode ?

• Pipe video frames from ffmpeg to numpy array without loading whole movie into memory

  2 mai 2021, par marcman

  I'm not sure whether what I'm asking is feasible or functional, but I'm experimenting with trying to load frames from a video in an ordered, but "on-demand," fashion.

  


  Basically what I have now is to read the entire uncompressed video into a buffer by piping through stdout, e.g. :

  


  H, W = 1080, 1920 # video dimensions
video = '/path/to/video.mp4' # path to video

# ffmpeg command
command = [ "ffmpeg",
            '-i', video,
            '-pix_fmt', 'rgb24',
            '-f', 'rawvideo',
            'pipe:1' ]

# run ffmpeg and load all frames into numpy array (num_frames, H, W, 3)
pipe = subprocess.run(command, stdout=subprocess.PIPE, bufsize=10**8)
video = np.frombuffer(pipe.stdout, dtype=np.uint8).reshape(-1, H, W, 3)

# or alternatively load individual frames in a loop
nb_img = H*W*3 # H * W * 3 channels * 1-byte/channel
for i in range(0, len(pipe.stdout), nb_img):
    img = np.frombuffer(pipe.stdout, dtype=np.uint8, count=nb_img, offset=i).reshape(H, W, 3)


  


  I'm wondering if it's possible to do this same process, in Python, but without first loading the entire video into memory. In my mind, I'm picturing something like :

  


  

  1. open the a buffer

  


  2. seeking to memory locations on demand

  


  3. loading frames to numpy arrays

  


  


  I know there are other libraries, like OpenCV for example, that enable this same sort of behavior, but I'm wondering :

  


  

  • Is it possible to do this operation efficiently using this sort of ffmpeg-pipe-to-numpy-array operation ?

  


  • Does this defeat the speed-up benefit of ffmpeg directly rather than seeking/loading through OpenCV or first extracting frames and then loading individual files ?