Recherche avancée

Sur d’autres sites (9507)

• ffmpeg drawtext support for Tibetan fonts

  19 juin 2020, par A-Circle Zhang

  I'm trying to use ffmpeg drawtext filter to render Tibetan text to videos. But some characters aren't rendered correctly.

  



  for example :

  



  


  སྤྱི་སྟེགས

  


  



  Expected render output

  



  some characters which is.. vertically too long, are not rendered correctly :

  



  Actual render output

  



  and this is my command line for generating the output :

  



  ffmpeg -i input.mp4 -filter_complex "drawtext=fontsize=80:fontcolor=black:fontfile=himalaya.ttf:shadowx=1:shadowy=1:text='སྤྱི་སྟེགས':x=0:y=0" -c:v libx264 -b:v 400k -c:a copy output.mp4


  



  The font himalaya.ttf (Microsoft Himalaya) is the default built-in font for Tibetan on Windows 10. (See https://docs.microsoft.com/en-us/typography/font-list/microsoft-himalaya )

  



  I tried text_shaping=1 and text_shaping=0, neither one is working.

  



  I've also tried other Tibetan fonts, problem is the same.

  



  But all of those fonts are rendered correctly in other places such as web browser.

  


• ffmpeg - compositing a video within a video in the centre

  1er mars 2017, par kieran

  I’m looking to composite a video with ffmpeg that places the video in the centre no matter what the composited video’s aspect ratio/size.

  The "background" video will always be 16:9 and 1920x1080px. I won’t know the aspect ratio or size of the overlay video as it will be user uploaded and could be any size/ratio.

  Here’s an example of what I’m trying to achieve :

  This is the background image :

  

  Now I want to overlay a video over the top :

  

  This should also work :

  Essentially no matter what the dimensions I want to ensure it’s always resized to fit within 1920x1080 and in addition ensure it’s always centred.

  Finally, if the uploaded video is also 16:9 it should simply overlay the entire video :
  

• ffmpeg : RTP missed packets, how to avoid ? [on hold]

  17 mai 2019, par James Adams

  I am saving H.264/5 video (from IP camera) to MP4 using ffmpeg without transcoding :

  $ ffmpeg -i  -vcodec copy -y -rtsp_transport tcp video.mp4

  I’m getting numerous missing packets messages, for example :

  [rtsp @ 0x55d6a47a8d40] RTP: missed 3 packets
  
  [rtsp @ 0x55d6a47a8d40] max delay reached. need to consume packet
  
  [rtsp @ 0x55d6a47a8d40] RTP: missed 1 packets
  
  [rtsp @ 0x55d6a47a8d40] max delay reached. need to consume packet
  
  [rtsp @ 0x55d6a47a8d40] RTP: missed 4 packets
  
  [rtsp @ 0x55d6a47a8d40] max delay reached. need to consume packet

  The above results in MP4 files that are wonky looking in some places, apparently data has been lost and this results in smeared looking video files.

  I seem to have better luck when I change the option order and put the -rtsp_transport tcp before the -i , but that’s also problematic because it seems to hang without ever finishing (at least I’ve never waited long enough for it to finish on its own before killing the process).

  I have had better luck with reading the video stream (same RTSP URL) using OpenCV (no missing packet messages) but I know of no way to avoid the transcoding step using that package so I’m stuck with ffmpeg (it seems that I’m not alone). Maybe OpenCV is doing some sort of buffering that helps it avoid dropping packets, and there’s a corresponding option that I haven’t found yet that will enable this for ffmpeg ?

  My system is a Dell XPS laptop running Ubuntu 18.04.

  How can I work around this issue ?