Recherche avancée

Sur d’autres sites (6831)

• FFmpeg dnn_processing with tensorflow backend : difficulties applying a filter on an image

  9 avril 2023, par ArnoBen

  I am trying to perform a video segmentation for background blurring similar to Google Meet or Zoom using FFmpeg, and I'm not very familiar with it.

  


  Google's MediaPipe model is available as a tensorflow .pb file here (using download_hhxwww.sh).

  


  I can load it in python and it works as expected, though I do need to format the input frames : scaling to the model input dimension, adding a batch dimension, dividing the pixel values by 255 to have a range 0-1.

  


  FFmpeg has a filter that can use tensorflow models thanks to dnn_processing, but I'm wondering about these preprocessing steps. I tried to read the dnn_backend_tf.c file in ffmpeg's github repo, but C is not my forte. I'm guessing it adds a batch dimension somewhere otherwise the model wouldn't run, but I'm not sure about the rest.

  


  Here is my current command :

  


  ffmpeg \
    -i $FILE -filter_complex \
    "[0:v]scale=160:96,format=rgb24,dnn_processing=dnn_backend=tensorflow:model=$MODEL:input=input_1:output=segment[masks];[masks]extractplanes=2[mask]" \
    -map "[mask]" output.png


  


  

  • I'm already applying a scaling to match the input dimension.

  


  • I wrote this [masks]extractplanes=2[mask] because the model outputs a HxWx2 tensor (background mask and foreground mask) and I want to keep the foreground mask.

  


  


  The result I get with this command is the following (input-output) :

  


  

  


  I'm not sure how to interpret the problems in this output. In python I can easily get a nice grayscale output :

  


  

  


  I'm trying to obtain something similar with FFmpeg.

  


  Any suggestion or insights to obtain a correct output with FFmpeg would be greatly appreciated.

  


  PS : If I try to apply this on a video file, it hits a Segmentation Fault somewhere before getting any output so I stick with testing on an image for now.

  


• PHP : Return response immediately, but run multiple lengthy shell_exec functions in the background

  15 janvier 2016, par Jake

  I am creating a simple web service that accepts a video upload, runs multiple different encodes on the video (mp4, webm, ogv), and then uploads the newly created files to our video host.

  Let’s say I have multiple commands..

  shell_exec('ffmpeg -i input.mp4 -f mp4 -c:v libx264 -preset slow -crf 24 -s 1280x720 -c:a libfdk_aac -profile:a aac_he -ar 22050 -b:a 64k -movflags +faststart output-1280x720.mp4');
  
  
  
  shell_exec('ffmpeg -i input.mp4 -f mp4 -c:v libx264 -preset slow -crf 24 -s 1920x1080 -c:a libfdk_aac -profile:a aac_he -ar 22050 -b:a 64k -movflags +faststart output-1920x1080.mp4');
  
  
  
  shell_exec('ffmpeg -i input.mp4 -f ogg -c:v libtheora -q:v 5 -s 1280x720 -c:a libvorbis -ar 22050 -b:a 64k -movflags +faststart output-1280x720.ogv');
  
  
  
  shell_exec('ffmpeg -i input.mp4 -f ogg -c:v libtheora -q:v 5 -s 1920x1080 -c:a libvorbis -ar 22050 -b:a 64k -movflags +faststart output-1920x1080.ogv');

  In summary, I want to...

  1. Print an immediate response : success : true
  2. Kick off multiple ffmpeg jobs synchronously.
  3. After each job is complete, send a POST to another server (one POST for each shell_exec).

  It would also be nice to only send the POST if the job was successful, but I could easily work around that by just checking to see if the output file exists on the server.

  I know that I can force shell_exec to run in the background by simply appending >/dev/null 2>/dev/null & to each command — which would allow me to print a response immediately — but I think by doing this, this would cause all of the jobs to run in parallel, and also, since this is diverting the output, I do not get any true callbacks when jobs are complete.

  Any ideas ??

• pthread : Avoid spurious wakeups

  18 octobre 2013, par Ben Jackson

  pthread : Avoid spurious wakeups
  pthread_wait_cond can wake up unexpectedly (Wikipedia : Spurious_wakeup).
  The FF_THREAD_SLICE thread mechanism could spontaneously execute
  
  jobs or allow the caller of avctx->execute to return before all
  
  jobs were complete.
  Test both cases to ensure the wakeup is real.
  Signed-off-by : Ben Jackson <ben@ben.com>
  
  Signed-off-by : Michael Niedermayer <michaelni@gmx.at>
  
  Signed-off-by : Derek Buitenhuis <derek.buitenhuis@gmail.com>
  
  Signed-off-by : Luca Barbato <lu_zero@gentoo.org>
  

  • [DBH] libavcodec/pthread.c