Recherche avancée

Sur d’autres sites (9153)

• ffmpeg watermark without background [closed]

  22 décembre 2022, par Edwin Pitters

  I have a problem, I am trying to add a watermark to my videos with ffmpeg using a gtx 1060 graphics card, the process works well and very fast, the problem is that the watermark appears with a black background , the image has no background, it is transparent, the problem happens only when I use the nvidia graphics card, because if I do the process with my processor the watermark is placed correctly as expected, so I am sure it is a problem in my configuration when running ffmpeg

  


  Here I leave the command that I am using :

  


  .\ffmpeg.exe -y -hide_banner -init_hw_device cuda=cuda -filter_hw_device cuda -hwaccel cuda -hwaccel_output_format cuda -i test.mp4 -i watermark.png -filter_complex "[1:v]colorchannelmixer=aa=0.3,scale=iw*0.6:-1,format=nv12,hwupload[img];[0:v][img]overlay_cuda=x='if(lt(mod(t\,16)\,8)\,W-w-W*10/100\,W*10/100)':y='if(lt(mod(t+4\,16)\,8)\,H-h-H*5/100\,H*5/100)'[out]" -map [out] -c:v h264_nvenc -b:v 6M -an -preset fast out_overlay.mp4


  


  If I use my processor with the following command, the guide mark without background is added, that is, as expected

  


  for %%a in ("*.m*") do ffmpeg -y -hide_banner -threads 4 -i "%%a" -preset ultrafast -vcodec libx264 -b:v 4000k -minrate 4000k -maxrate 4000k -bufsize 4000k -c:a aac -b:a 64k -pass 1 -f mp4 NUL && ffmpeg -y -hide_banner -threads 8 -i "%%a" -i watermark.png -preset ultrafast -vcodec libx264 -b:v 4000k -minrate 4000k -maxrate 4000k -bufsize 4000k -filter_complex "[1]colorchannelmixer=aa=0.3,scale=iw*0.8:-1[a];[0][a]overlay=x='if(lt(mod(t\,8)\,4)\,W-w-W*10/100\,W*10/100)':y='if(lt(mod(t+2\,8)\,4)\,H-h-H*10/100\,H*10/100)'" -c:a copy -tune film -movflags +faststart -pass 2 "watermark/%%a"
pause


  


  I also tried changing colorchannelmixer=aa=0.3 for lut=a=val*0.3 but it seems that this command is not having any effect

  


  I find that the image is well reviewed to discard, in fact I tried with other images also with a transparent background and I have the same result, a watermark but with a black background

  


• ffmpeg : concat videos and images

  23 mai 2016, par Yosko

  I have 2 videos (same resolution, same encoding) files that I want to concat and I want to insert some text for 3 seconds between them, as a splitter. I’m doing this with ffmpeg on Windows.

  Optional ideas that I would be interested in :

  • avoid reencoding the video in the process
  • having a fade in / fade out at the intersection of each part

  For now, I made the text as an image (but I am open to other suggestions). Let’s say I have :

  • video1.mp4 : 6:33
  • splitter.png (same resolution as video1.mp4)
  • video2.mp4 : 16:44

  I have tried a few things, but I always end up with the same problem : the video is 23:20 (video1 + 3 seconds + video2), but the 3 seconds gap is just the last video1 frame frozen instead of my image/text...

  Any Idea what I did wrong or how I should achieve this ?

  Here is what I tried so far :

  Method 1 : image to video

  Turn the image into a 3 seconds mp4 film, then concat (demuxer) it with the others :

  ffmpeg -loop 1 -f image2 -i splitter.png -r 30 -t 3 splitter.mp4
  
  ffmpeg -f concat -i input.txt -codec copy output.mp4

  Where the input.txt looks like :

  file 'E:\video1.mp4'
  
  file 'E:\splitter.mp4'
  
  file 'E:\video2.mp4'

  The content of splitter.png is visible in the splitter.mp4, but not in the output.mp4. Also I’m not entirely sure the splitter.mp4 respects the exact same encoding as the 2 videos, and I don’t know how to verify that.

  Method 2 : insert image frames

  Directly run the concat (demuxer) 90 times (30fps -> 3 seconds) on the image

  ffmpeg -f concat -i input.txt -codec copy output.mp4

  Where the input.txt looks like :

  file 'E:\video1.mp4'
  
  file 'E:\splitter.png'
  
  ...
  
  file 'E:\splitter.png'
  
  file 'E:\video2.mp4'

  Edit : possible solution ?

  Since all I’m doing is screencasting, I might as well screencast my splitter image. This way I would be sure of the audio & video encoding and wouldn’t have any problem merging and it wouldn’t need any reencoding... I know it might sound dumb, but it would probably do the trick...

  Note : I didn’t have try it, since I already worked through Openshot.

• vulkan_decode : use a single execution pool

  3 décembre 2024, par Lynne

  vulkan_decode : use a single execution pool
  Originally, the decoder had a single execution pool, with one
  
  execution context per thread. Execution pools were always intended
  
  to be thread-safe, as long as there were enough execution contexts
  
  in the pool to satisfy all threads.
  Due to synchronization issues, the threading part was removed at some
  
  point, and, for decoding, each thread had its own execution pool.
  
  Having a single execution pool per context is hacky, not to mention
  
  wasteful.
  
  Most importantly, we *cannot* associate single shaders across multiple
  
  execution pools for a single application. This means that we cannot
  
  use shaders to either apply film grain, or use this framework for
  
  software-defined decoders.
  The recent commits added threading capabilities back to the execution
  
  pool, and the number of contexts in each pool was increased. This was
  
  done with the assumption that the execution pool was singular, which
  
  it was not. This led to increased parallelism and number of frames
  
  in flight, which is taxing on memory.
  This commit finally restores proper threading behaviour.
  
  The validation layer has isses that are reported and addressed in the
  
  earlier commit.
  

  • [DH] libavcodec/vulkan_decode.c
  • [DH] libavcodec/vulkan_decode.h