Recherche avancée

Sur d’autres sites (12145)

• Pipe numpy array to virtual video device

  16 mars 2021, par sre

  I want to pipe images to a virtual video device (e.g. /dev/video0), the images are created inside a loop with the desired frame rate.

  


  In this minimal example i only two arrays which alternate in the cv2 window. Now i look for a good solution to pipe the arrays to the virtual device.

  


  I saw that ffmpeg-python can run asynchronous with ffmpeg.run_async(), but so far i could not make anything work with this package.

  


  example code without the ffmpeg stuff :

  


  #!/usr/bin/env python3

import cv2
import numpy as np
import time

window_name = 'virtual-camera'
cv2.namedWindow(window_name, cv2.WINDOW_GUI_EXPANDED)

img1 = np.random.uniform(0, 255, (1080, 1440, 3)).astype('uint8')
img2 = np.random.uniform(0, 255, (1080, 1440, 3)).astype('uint8')

for i in range(125):
    time.sleep(0.04)
    if i % 2:
        img = img1
    else:
        img = img2
    cv2.imshow(window_name, img)
    cv2.waitKey(1)
cv2.destroyAllWindows()


  


• ffmpeg pipe Invalid data found when processing input

  28 mars 2021, par Ankit Maheshwari

  Here is my configuration.


  


  


  


  const ffmpegPath = require('@ffmpeg-installer/ffmpeg').path;
const spawn = require('child_process').spawn;

ffmpeg = spawn(ffmpegPath, [
      
// Remove this line, as well as `-shortest`, if you send audio from the browser.
      // '-f', 'lavfi', '-i', 'anullsrc',

      // FFmpeg will read input video from STDIN
      '-i', '-',

      // -re flag means to Read input at native frame rate.
      '-re', '-y',

      // thread_queue_size added to avoid err: Thread message queue blocking; consider raising the thread_queue_size option, required before each input - this is for image2
      // '-thread_queue_size', '2048',

      // REF TO OVERLAY 
      // https://stackoverflow.com/questions/10438713/overlay-animated-images-with-transparency-over-a-static-background-image-using-f?rq=1
      // -loop loops the background image input so that we don't just have one frame, crucial!

      // The image file muxer writes video frames to image files, http://underpop.online.fr/f/ffmpeg/help/image2-1.htm.gz
      '-f', 'image2',

      // The -loop option is specific to the image file demuxer and gif muxer, so it can't be used for typical video files, but it can be used to infinitely loop a series of input images.
      '-loop', '1',

      // pattern_type is used to determine the format of the images contained in the files.
      // Read images matching the "*.png" glob pattern, that is files terminating with the ".png" suffix
      '-pattern_type', 'glob',

      // '-i', `images/${streamConfigData.youtube_key}/destination/image-*.png`,

      '-i', `images/${streamConfigData.youtube_key}/overlay/abc.png`,

      // '-vf', 'scale=1920x1080:flags=lanczos',
      
      // -shortest ends encoding when the shortest input ends, which is necessary as looping the background means that that input will never end.
      // 'overlay=shortest=1',

      "-filter_complex", "[1:v]format=rgba,colorchannelmixer=aa=1[fg];[0][fg]overlay",
      
      // Because we're using a generated audio source which never ends,
      // specify that we'll stop at end of other input.  Remove this line if you
      // send audio from the browser.
      // '-shortest',
      
      // If we're encoding H.264 in-browser, we can set the video codec to 'copy'
      // so that we don't waste any CPU and quality with unnecessary transcoding.
      // If the browser doesn't support H.264, set the video codec to 'libx264'
      // or similar to transcode it to H.264 here on the server.
      // '-vcodec', 'libx264',
      // it is not possible to filter and stream copy the same stream at the same time. https://stackoverflow.com/a/53526514/4057143
      '-vcodec', 'copy',
      
      // if browser not supports encoding AAC, we must transcode the audio to AAC here on the server.
      // '-acodec', 'aac',

      // Use this rate control mode if you want to keep the best quality and care less about the file size. CRF scale is 0–51, where 0 is lossless, 23 is the default, and 51 is worst quality possible. A lower value generally leads to higher quality, https://trac.ffmpeg.org/wiki/Encode/H.264
      '-crf', '23',

      // preset provide a certain encoding speed to compression ratio. A slower preset will provide better compression. medium – default preset, https://trac.ffmpeg.org/wiki/Encode/H.264
      '-preset', 'ultrafast',

      // -r set the frame rate. Generally, -r. Use the filter when you need to change framerate before applying further filters.
      // '-r', '30',
      // '-framerate', '30',

      //debug level logs
      '-loglevel', 'debug',
      '-v', 'verbose',

      // -g GOP_LEN_IN_FRAMES, -g sets the keyframe interval. https://superuser.com/a/908325
      '-g', '60',

      // video timescale, not sure what it is!
      '-video_track_timescale', '1000',

      // a live stream with more/less constant bit rate, to be able to control the bandwidth used.
      // a live stream with limited bit rate
      '-b:v', '15000k',
      // '-maxrate', '4000k',
      // '-bufsize', '8000k',

      // FLV is the container format used in conjunction with RTMP
      '-f', 'flv',
      
      // The output RTMP URL.
      // For debugging, you could set this to a filename like 'test.flv', and play
      // the resulting file with VLC.
      rtmpUrl 
    ], {
      env: {
          NODE_ENV: 'production',
          PATH: process.env.PATH
      }
    });

  


  


  



• Create silent wav and pipe it

  2 avril 2021, par Ícaro Erasmo

  I've been through many stackoverflow pages and forums trying to find the answer I want.
I created a virtual microphone and I'm trying to pipe to it some wav sounds created using FFMPEG.

  


  When I want to pipe a keyboard noise I pipe the sound to my virtual sound capture device like this :

  


  ffmpeg -fflags &#x2B;discardcorrupt -i <keyboard sound="sound" path="path"> -f s16le -ar 44100 -ac 1 - > /tmp/gapFakeMic&#xA;</keyboard>

  


  And when I want to pipe some synthetized voice sound using Espeak to my virtual microphone, I do this :

  


  espeak -vbrazil-mbrola-4 <some random="random" text="text"> --stdout | ffmpeg -fflags &#x2B;discardcorrupt -i pipe:0 -f s16le -ar 44100 -ac 1 - > /tmp/gapFakeMic&#xA;</some>

  


  The problem is my capture device doesn't record the sound like a normal recorder that still records even when there's no sound being transmited to it. So I'm trying to append the silence to the wav which is being created while my application is running. Always when I try to send the silence to buffer, FFMPEG returns the following response :

  


  [NULL @ 0x5579f7921a00] Unable to find a suitable output format for 'pipe:'


  


  FFMPEG is a powerful tool but its documentation lacks to be useful for newbies like me. So, I'd appreciate if anyone could answer this or at least give me any direction or some resource where I could find a way of achieving this.

  


  EDIT :

  


  Here's how I'm producing the silence to my virtual microphone :

  


  ffmpeg -f lavfi -i anullsrc=channel_layout=mono:sample_rate=44100 -t <time in="in" seconds="seconds"> - > /tmp/gapFakeMic&#xA;</time>

  


  Here's the full log :

  


  ffmpeg version 4.1.6-1~deb10u1 Copyright (c) 2000-2020 the FFmpeg developers
  built with gcc 8 (Debian 8.3.0-6)
  configuration: --prefix=/usr --extra-version='1~deb10u1' --toolchain=hardened --libdir=/usr/lib/x86_64-linux-gnu --incdir=/usr/include/x86_64-linux-gnu --arch=amd64 --enable-gpl --disable-stripping --enable-avresample --disable-filter=resample --enable-avisynth --enable-gnutls --enable-ladspa --enable-libaom --enable-libass --enable-libbluray --enable-libbs2b --enable-libcaca --enable-libcdio --enable-libcodec2 --enable-libflite --enable-libfontconfig --enable-libfreetype --enable-libfribidi --enable-libgme --enable-libgsm --enable-libjack --enable-libmp3lame --enable-libmysofa --enable-libopenjpeg --enable-libopenmpt --enable-libopus --enable-libpulse --enable-librsvg --enable-librubberband --enable-libshine --enable-libsnappy --enable-libsoxr --enable-libspeex --enable-libssh --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvorbis --enable-libvpx --enable-libwavpack --enable-libwebp --enable-libx265 --enable-libxml2 --enable-libxvid --enable-libzmq --enable-libzvbi --enable-lv2 --enable-omx --enable-openal --enable-opengl --enable-sdl2 --enable-libdc1394 --enable-libdrm --enable-libiec61883 --enable-chromaprint --enable-frei0r --enable-libx264 --enable-shared

  libavutil      56. 22.100 / 56. 22.100
  libavcodec     58. 35.100 / 58. 35.100
  libavformat    58. 20.100 / 58. 20.100
  libavdevice    58.  5.100 / 58.  5.100
  libavfilter     7. 40.101 /  7. 40.101
  libavresample   4.  0.  0 /  4.  0.  0
  libswscale      5.  3.100 /  5.  3.100
  libswresample   3.  3.100 /  3.  3.100
  libpostproc    55.  3.100 / 55.  3.100

Input #0, lavfi, from 'anullsrc=channel_layout=mono:sample_rate=44100':
  Duration: N/A, start: 0.000000, bitrate: 352 kb/s
    Stream #0:0: Audio: pcm_u8, 44100 Hz, mono, u8, 352 kb/s

[NULL @ 0x560516626f40] Unable to find a suitable output format for 'pipe:'
pipe:: Invalid argument


  


  EDIT 2 :

  


  After Gyan provided a solution in the comments the error above doesn't show anymore but my result audio is being broken and doesn't come out as expected. Now the command that generates and appends the silent audio is like this :

  


  ffmpeg -f lavfi -i anullsrc=channel_layout=stereo:sample_rate=44100 -t <time in="in" seconds="seconds"> -f s16le - > /tmp/gapFakeMic&#xA;</time>

  


  Edit 3 :

  


  I've made some changes to the command I'm using to pipe silence to the virtual mic. I think the pipe is breaking because of some incompatibility in audio formats. I hope I can find a solution in the next few days. After every little change I realize some improvements. Now I can hear the silence between the keys sounds but it isn't recording all the audios I'm passing to it. Here's how the command is now :

  


  ffmpeg -f lavfi -i anullsrc=channel_layout=mono:sample_rate=44100 -t <time in="in" seconds="seconds"> -f s16le -ar 44100 -ac 1 - > /home/icaroerasmo/gapFakeMic`&#xA;</time>

  


  I also realized that when I pipe the sound to a pipe file created inside my home folder the audio quality improves.

  


  Edit 4 :

  


  After all this struggle it's clear now that the named pipe is breaking in the second time it's called. I've already googled how to flush a named pipe but I didn't find anything that worked.