Recherche avancée

Sur d’autres sites (4389)

• ffmpeg cannot write single-channel float32 exr image ?

  24 janvier 2024, par Hugues

  I am trying to use ffmpeg to write an EXR image containing a single channel of float32 values.
Here is the code :

  


  def test_ffmpeg_float_gray_to_exr():
  import numpy as np
  import subprocess

  array1 = np.array([[0.1, 0.2], [0.4, 0.7]], np.float32)  # Single-channel 2x2 image.
  array1.tofile('temp1.raw')
  ffmpeg_command_write = [
      'ffmpeg', '-hide_banner',
      '-f', 'rawvideo',
      '-pix_fmt', 'grayf32le',
      '-s', '2x2',
      '-i', 'temp1.raw',
      '-pix_fmt', 'grayf32le',
      '-compression', 'zip1',  # ['none', 'rle', 'zip1', zip16']
      '-format', 'float',  # ['half', 'float']
      '-vcodec', 'exr',
      '-y', 'output.exr',
  ]
  subprocess.run(ffmpeg_command_write, check=True)

  ffmpeg_command_read = [
      'ffmpeg', '-hide_banner',
      '-i', 'output.exr',
      '-f', 'image2pipe',
      '-pix_fmt', 'grayf32le',
      '-y', 'temp2.raw',
  ]
  subprocess.run(ffmpeg_command_read, check=True)
  array2 = np.fromfile('temp2.raw', dtype=np.float32).reshape(2, 2)
  print(array2)

test_ffmpeg_float_gray_to_exr()


  


  When writing, ffmpeg gives the message Incompatible pixel format 'grayf32le' for codec 'exr', auto-selecting format 'gbrpf32le' and then reading back the values gives incorrect results.

  


  If I create a single-channel exr image (using OpenCV), it does read correctly using the ffmpeg_command_read command in the above code, so only the writing part is broken.

  


  And, I am able to create 3-channel (gbrpf32le) and 4-channel (gbrapf32le) exr images using ffmpeg and variants of the above code. Just not 1-channel.

  


  The source code https://github.com/FFmpeg/FFmpeg/blob/master/libavcodec/exrenc.c suggests that it should be possible to encode grayf32le (but I'm no expert).

  


  Am I missing some command-line option to enable this ?

  


• Serving a single JPEG using ffserver

  4 mai 2016, par Magnus

  I have a setup where a local application writes a sequence of JPEG images into a FIFO (Unix named pipe on Linux). On the other end I have ffmpeg picking up the sequence and passing it into an instance of ffserver :

  % ffmpeg -i fifo.mjpeg http://127.0.0.1:8090/feed.ffm

  The configuration for ffserver looks like this :

  HTTPPort 8090
  
  HTTPBindAddress 0.0.0.0
  
  MaxHTTPConnections 20
  
  MaxClients 10
  
  MaxBandwidth 1000
  
  
  
  <feed>
  
    File /tmp/feed.ffm
  
    FileMaxSize 200k
  
  
  
    ACL allow 127.0.0.1
  
  </feed>
  
  
  
  <stream>
  
    Format mpjpeg
  
    Feed feed.ffm
  
  
  
    VideoSize 960x600
  
    VideoFrameRate 2
  
    VideoIntraOnly
  
    Strict -1
  
  
  
    NoAudio
  
    NoDefaults
  
  </stream>

  This works fine, I can point my web browser to http://127.0.0.1:8090/stream.mpjpeg and see the video.

  Now I want to add a way to download a single JPEG (I think of it as snapshot of the video). I added the following to the ffserver configuration :

  <stream>
  
    Format singlejpeg
  
    Feed feed.ffm
  
  
  
    VideoSize 960x600
  
    VideoFrameRate 2
  
    VideoIntraOnly
  
    Strict -1
  
  
  
    NoAudio
  
    NoDefaults
  
  </stream>

  That only sort of works. Sure, if I point my browser to http://127.0.0.1:8090/image.jpg I do so a still picture from the video, but the browser never stops loading !

  Indeed, if I run wget http://127.0.0.1:8090/image.jpg I see that the MIME type is good (image/jpeg), but there seems to be no end to the image.

  Am I missing something in my configuration that makes ffserver sending more than a single image ?

  I should add I’ve tried this setup on both 2.8.6 (Debian Jessie, package comes from jessie-backports) and 3.0 (Arch Linux), with the same result in both cases.

• Not able to concatenate encrypted .ts segments into single .ts segment using FFMPEG encoder

  19 juillet 2017, par Shivam657

  We are trying to concatenate segments encrypted with AES-128 bit encryption keys into a single .ts segment by ffmpeg encoder.

  Command used to concatenate :-

  ffmpeg -i "concat:segment_namev4_0.ts|segment_name_v4_1.ts" -c copy  output.ts

  But unfortunately, we are getting the following error :

  [concat @ 0x7fe02f800000] Impossible to open ’/Users/userName/Documents/uploads/segment_name_v4_aef7e05a-dfb2-4d88-a007-8be9ebad9600/./segment_name_v4_0.ts’
  /dev/fd/63 : Invalid data found when processing input

  We also tried using cat command,
  cat sample_0_av.ts sample_1_av.ts sample_2_av.ts sample_3_av.ts > alltest.ts

  though this concatenated the file, but when we tried to stream the file using the same AES-128 bit key, the video is stopping at merging points.

  Any advice/suggestions are most welcome :)