Advanced search

On other websites (4358)

• FFMPEG h.264 stream to VLC from raw YUV444 produces black screen

  30 November 2018, by GerbGerb

  Not sure if this fits better here or over on Super User, so please redirect if here is the wrong place.

  I am piping raw YUV444 640x480 frames to ffmpeg from inside a C++ program using stdout. ffmpeg encodes them using h.264 and streams the video via UDP to an IP of my choice.

  When I am not streaming and saving the video as -mp4 file, everything works - But when I stream over the network and try to receive in VLC, the stream is received and the playing timer under the video screen runs, but the image is black, none of the video is displayed. Also I have some crackling sounds on my audio output.

  My FFMPEG settings are as follows:

  ffmpeg -y -loglevel info -f rawvideo -r 25 -c:v rawvideo -pix_fmt yuv444p -s 640x480 -i - -c:v libx264 -preset superfast -framerate 25 -f mpegts udp://192.168.150.1:20001

  in VLC, I just use "open network stream" with udp://@:20001. The port is open, I tested it with a simple dummy sender/receiver.
  Almost the same cli arguments for the video recording lead to a perfectly playable video when writing to a file and downloading it:

  ffmpeg -y -loglevel info -f rawvideo -r 25 -c:v rawvideo -pix_fmt yuv444p -s 640x480 -i - -c:v libx264 -preset superfast -framerate 25 video.mp4

  If I try to stream the same video using

  ffmpeg -i video.mp4 -v 0 -vcodec mpeg4 -f mpegts udp://192.168.150:20001

  the screen remains black.

  VLC debug output changes between different runs, because other modules are selected, so I assume stream type recognition goes haywire. But why?

  What am I doing wrong?

  Here the ffmpeg log
  and the VLC log

• Extremely slow ffmpeg/sws_scale() - only on heavy duty

  28 September 2020, by user2328447

  I am writing a video player using ffmpeg (Windows only, Visual Studio 2015, 64 bit compile).
With common videos (up to 4K @ 30FPS), it works pretty good. But with my maximum target - 4K @ 60FPS, it fails. Decoding still is fast enough, but when it comes to YUV/BGRA conversion it is simply not fast enough, even though it's done in 16 threads (one thread per frame on a 16/32 core machine).

  



  So as a first countermeasure I skipped the conversion of some frames and got a stable frame rate of 40 that way. Comparing the two versions in Concurrency Visualizer, I found a strange issue I don't know the reason of.

  



  .

  



  Here's an image of the frameskip version:

You see that the conversion is pretty quick (average roughly 35ms)
Thus, as multiple threads are used, it also should be quick enough for 60FPS, but it isn't!

  



  .

  



  The image of the non-frameskip version shows why:

The conversion of a single frame has become ten times slower than before (average roughly 350ms). Now a heavy workload on many cores would of course cause a minor slowdown per core due to reduced turbo - let's say 10 or 20%. But never an extreme slowdown of 1000%.

  



  .

  



  Interesting detail is, that the stack trace of the non-frameskip version shows some system activity I don't really understand - beginning with ntoskrnl.exe!KiPageFault+0x373. There are no exceptions, other error messages or such - it just becomes extremely slow.

  



  Edit: A colleague just told me that this looks like a memory problem with paged-out memory at first glance - but my memory utilization is low (below 1GB, and more than 20GB free)

  



  Can anyone tell me what could be causing this?

  


• ffmpeg: Create a fake shadow below alpha channel webm/png sequence

  6 May 2021, by Beneos Battlemaps

  Purpose: I'd like to render out animated 3D meshes as png sequence to use them as animated tokens for virtual tabletop games. To make the mesh looks more natural I'd like to create a fake show beneath the actual token.

  


  Problem: I have a png sequence 1 (as well as a webm file created with ffmpet out of this png sequence if it makes it easier) with alpha channel. To create the webm I use:
ffmpeg -framerate 24 -f image2 -i Idle_Top.%04d.png -c:v libvpx-vp9 -crf 25 -pix_fmt yuva420p Idle_Top.webm (If its relevant). I'd like to render out the png sequence to a webm file that have the current images as well as the transparent shadow beneath the token combined.

  


  Possible workflow: I think a good way to achieve the wanted shadow effect is to use the alpha channel image as a mask on a black picture with the same resolution as the source image 2. Then you have a complete black version of the image. Then you need to place this image beneath the colored image and make a offset of 10px left and 10px down to create the ilusion of perspective 3. At the end the black image below the colored image must have a transparency as well ( 30% visibility should be enough) 4.

  


  Workflow overview

  


  Assets: I've put the webm file and the png files on my gDrive https://drive.google.com/drive/folders/1wznGaPwhKc2UyPpSZBSISa1gs3oixsHR?usp=sharing

  


  Though I work with ffmpeg on a regular basis I have no clue where to start. Can you please help me out with this interesting problem?

  


  Best regards
Ben