Recherche avancée

Sur d’autres sites (6607)

• 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.

  


• I am streaming mp3 music via ffmpeg to a local rtmp server then converting to hls, but am having difficulties end to end testing

  12 avril 2020, par SquirrelSenpai

  I am streaming mp3 music via ffmpeg to a local rtmp server then converting to hls, but am having difficulties end to end testing. I am know test.m3u8 playlist should be produce, however I am unable to check inside /nginx/hls/ as it is locked by www-data during operation. I have tried multiple permutation of what I thought the output hls stream would be in vlc with no luck. localhost:8080/live/test.m3u8, localhost:8080/hls/test.m3u8

  



  Any tips on effective testing would be much appreciated.

  



  Technologies involved :

  



  

  • FFMPEG

  


  • NGINX (This and the below 3 are part of a module)

  


  • HLS

  


  • RTMP

  


  



  Working :

  



  ffmpeg -hide_banner -i http://149.255.59.164:8138 -f mp3 test.mp3


  



  Seemingly working, correctly reads files, shows conversion of some kind
  
size= 362kB time=00:00:23.09 bitrate= 128.3kbits/s speed=3.21x

  



  fmpeg -hide_banner -i http://x.x.x.x:8138 -f mp3 rtmp://localhost:1935/live/test


  



  Nginx.conf

  



  user www-data;
worker_processes auto;
pid /run/nginx.pid;
include /etc/nginx/modules-enabled/*.conf;

events {
        worker_connections 768;
        # multi_accept on;
}

rtmp_auto_push on;

rtmp{

        server{

                listen 1935;

                chunk_size 4000;

                #one publisher, many subscribers

                application live {

                        # enable live streaming
                        live on;
                        record off;

                        # publish only from localhost
                        allow publish 127.0.0.1;
                        deny publish all;

                        # hls - required for web browser consumption
                        hls on;
                        hls_path /tmp/hls;
                        hls_fragment 3;
                        hls_playlist_length 60;

                        # disable consuming the streaming from nginx as rtmp
                        deny play all;

                }

        }

}

# HTTP can be used for accessing RTMP stats
http {

    server {

        listen      8080;

        # This URL provides RTMP statistics in XML
        location /stat {
            rtmp_stat all;

            # Use this stylesheet to view XML as web page
            # in browser
            rtmp_stat_stylesheet stat.xsl;
        }

        location /stat.xsl {
            # XML stylesheet to view RTMP stats.
            # Copy stat.xsl wherever you want
            # and put the full directory path here
            root /path/to/stat.xsl/;
        }

        location /hls {
            # Serve HLS fragments
            types {
                application/vnd.apple.mpegurl m3u8;
                video/mp2t ts;
            }
            root /tmp;
            add_header Cache-Control no-cache;
        }

        location /dash {
            # Serve DASH fragments
            root /tmp;
            add_header Cache-Control no-cache;
        }
    }
}


  


• ffmpeg tile screenshot file with uniform distribution over video length

  23 octobre 2017, par Michael Yousef

  I’m trying to take a video and create a screenshot file for it. I want it to cover the entire duration of the video and be uniformly distributed over the video. I have a command right now that I found online, it can produce a screenshot file, but it doesn’t cover the entire duration

  ffmpeg -ss 00:05:00 -i video.mp4 -frames 1 -vf "select=not(mod(n\,8000)),scale=320:240,tile=4x8" out.png

  Instead of having it cap every 80 seconds, I want it to determine what the duration is use that. It should cap the first screen at my initial offset, then however far ahead as necessary.

  Also, if anyone knows how to add information to the outputted file at the top, like filename, duration, bitrate, etc., that’s also something I want to output.