Recherche avancée

Sur d’autres sites (10245)

• Facebook live video can only be previewed while streaming my Mac screen

  24 janvier 2017, par iownthegame

  I use the following ffmpeg command to share my screen streaming to Facebook.

  

  ffmpeg -f avfoundation -r 10 -video_size 352x288 -i "0" -c:v h264 -f flv 
  
  "rtmp://rtmp-api.facebook.com:80/rtmp/1269140699772419?ds=1&a=AaYsXcYcdHQrrrUF"

  then I can see the video preview scenes
  

  however, once I press the ’Go Live’ button, and go to my facebook page, I can find a post said I am Live Now. When I press the ’play button’, there are only black scenes that last for 1 2 seconds, and the video ends.

  

  And there is also a weird thing, the video can be played when I end up the live stream. When refreshing the page, I can find a post said I was Live, the video of this post can be played successfully.

  Anyone knows why the video can be watched when previewing and finishing live but not the exact live moment ?

  [updated]

  During the live streaming period, if I press the play button, there are only 1 2 seconds black scenes and the live stream ends.
  

  However if I press the video frame again, another modal pops up then I can see the live streaming. Is it a facebook bug ?
  

• lavc/opusdsp : RISC-V V (128-bit) postfilter

  5 octobre 2022, par Rémi Denis-Courmont

  lavc/opusdsp : RISC-V V (128-bit) postfilter
  This is implemented for a vector size of 128-bit. Since the scalar
  
  product in the inner loop covers 5 samples or 160 bits, we need a group
  
  multipler of 2.
  To avoid reconfiguring the vector type, the outer loop, which loads
  
  multiple input samples sticks to the same multipler. Consequently, the
  
  outer loop loads 8 samples per iteration. This is safe since the minimum
  
  period of the CELT codec is 15 samples.
  The same code would also work, albeit needlessly inefficiently with a
  
  vector length of 256 bits. A proper implementation will follow instead.
  

  • [DH] libavcodec/opusdsp.c
  • [DH] libavcodec/opusdsp.h
  • [DH] libavcodec/riscv/Makefile
  • [DH] libavcodec/riscv/opusdsp_init.c
  • [DH] libavcodec/riscv/opusdsp_rvv.S

• Gstreamer pipeline to scale down video before streaming

  20 novembre 2014, par r3dsm0k3

  Here is what Im trying to achieve.

  Im streaming from a Logitech C920 camera on beaglebone black with gstreamer. I have to save a copy of the video saved locally while it is streaming. I have achieved that with tee.
  Logitech camera gives h264 encoded video at a certain bitrate, mostly very high.

  Im streaming from a moving car on 3G, and the network is not good enough to send the stream to nginx-rtmp server Im using to re-distribute thus gives strong artifacts in the result.

  Im able to alter the bitrate of captured video using uvch264.
  But then, the locally saved video also would have lower bitrate.

  Is there anyway of capturing a higher bitrate 1080p video from the camera and sending a lower resolution, lower bitrate video the streaming server ?

  Following is the pipeline I have currently.

  gst-launch-1.0 -v -e uvch264src initial-bitrate=400000 average-bitrate=400000 iframe-period=3000  device=/dev/video0 name=src auto-start=true  src.vidsrc ! queue ! video/x-h264,width=1920,height=1080,framerate=30/1 ! h264parse ! flvmux streamable=true  name=flvmuxer ! queue ! tee name=t ! queue ! filesink location=/mnt/test.flv t. ! queue ! rtmpsink location=$SERVER/hls/$CAM1

  I could also try sending the higher bitrate video to a udpsink instead of rtmpsink and with another gstreamer process parallely and takes the data using a udpsink and probably post process/ re-encode and send to rtmp server.

  Im also limited by the processing speed BeagleBone has to do for encoding the videos. Currently Im trying for 1 camera and in the finished project I would like to have 2 cameras connected. Upload speed Im getting for the network is under 1Mbps.

  How do I solve this with less load on the BeagleBone ? Im very open to a new architecture as well.