- Cin - Cinelerra-GG Mailing Lists

Two patches hopefully fixing build on NetBSD 10.0/amd64
by Andrew Randrianasulu 18 Dec '24

18 Dec '24

Currently from that system I still need to set export LD_LIBRARY_PATH="/usr/pkg/lib/ffmpeg6" and set audio output to oss - > /dev/audio but it seems to work. Can you check them by building on Linux? :)

2 22

Problem with X11-OpenGL
by Andrea paz 17 Dec '24

17 Dec '24

I continue from another post: >> After startup I have the following message: >> >> BC_DisplayInfo::gl_fb_config failed >> >> Do you know what it could be? >> > something between mesa and X server? > I fixed my mesa3d version to 22.1.7 due to breakage on nouveau with my > specific low end card (gt710) so I am not following latest mesa changes ... >> Another problem, if I load one or more media, when I set the >> X11-OpenGL driver, I get the following warning: >> >> void AWindowGUI::update_asset_list(): >> Warning: /home/paz/test/sad_woman.mxf >> dimensions 1920x1080 exceed OpenGL texture limit 0 >> void AWindowGUI::update_asset_list(): >> Warning: /home/paz/test/Unusual_Habitat.wav >> dimensions 0x0 exceed OpenGL texture limit 0 >> void AWindowGUI::update_asset_list(): >> Warning: /home/paz/test/video (2160p)HDR_01.mp4 >> dimensions 3840x2160 exceed OpenGL texture limit 0 >> void AWindowGUI::update_asset_list(): >> Warning: /home/paz/test/vp.webm >> dimensions 512x288 exceed OpenGL texture limit 0 >> Decoder dnxhd does not support device type vaapi. >> HW device init failed, using SW decode. >> file:/home/paz/test/sad_woman.mxf >> err: Operation not permitted >> FFMPEG::open_decoder: some stream times estimated: /home/paz/test/vp.webm >> >> editing on the timeline more or less works, but I often have freezes >> that force me to reboot. >> >> If I use the X11 driver everything works fine and I don't have the freezes. I've tried several tests, but I can't get to the bottom of the problem. On my system I have the following packages involving OpenGL: freeglut 3.6 glu 9.0.3 glslang 14.3 libepoxy 1.5.10 libglvnd 1.7 qt6-shadertools 6.7 and mesa 24.2.3 xf86-video-amdgpu 23.0.0 (for the 2D and opengl) I also have the same errors in Wayland (which, I imagine, uses xwayland) however I have better performance in timeline and no crashes. I too think it is some mesa update or, even more likely, amd drivers. However the problem is not only in the latest CinGG release but also in the previous one and also in a different distro (however always rolling and Arch based).

2 15

tried to add ubuntu24 specifically to bld_prepare.sh
by Andrew Randrianasulu 16 Dec '24

16 Dec '24

I have Ubuntu Studio 24.04.x on slow usb drive. I tried to get cingg sources there and compile her. It seems few packages changed their name (fftw3) or disappeared (python2) so I edited new case for our bld_prepare.sh script. git am patch attached, but it will be nice if anyone tested that on another ubuntu 24.04.x system

3 5

Multicam video output?
by Stefan de Konink 16 Dec '24

16 Dec '24

Given all the performane issues involved with multi track editing in Cinelerra, where every track is effectively rendered anyway. Is there any existing option to get a video track output into a separate viewer / compositor? Read: give me a window that only shows what is on this track. I am aware that <https://cinelerra-gg.org/download/CinelerraGG_Manual/Multi_Camera_Mixer.html> exists, but I am actually looking for just 'the track output' and not resource based interaction involved with mixers. Furthermore it might be a nice addition to get the multiview mode that OBS has to direct the program or multiview to a separate screen. -- Stefan

4 3

Some ideas how to speed up cingg
by Andrew Randrianasulu 16 Dec '24

16 Dec '24

Well, I only tested 1) 1) add fast lowres=3 skiploopfilter=all to your decode.opts file lowres really only work for dv/mpe2/mjpeg. 3 is really aggressive, try 1 or 2 if it works at all. 2) Set scaling method to simplest, nearest. Obviously only work if you do not do scaling, or plan to redo it anyway at higher quality. 3) Look if replacing -Ofast with " -mach=yourcpu, -mtune=yourcpu" works better (grep Makefiles to see where it set) 4) Add same march/mtune cflags to individual plugins Makefiles and see if it really speed ups them. 5) Hard - look into openmp parallel/simd directives and see if they apply to any for loops in slow plugins. 6) As I said - proxies ... It seems that in worst case scenario (g=250 frames, fps=25) for fast (less than 1 second) seeking to any point in stream you need nearly 10x more power that was needed for mere playback! It goes better for smaller g and bigger fps, but still. H265/4/2 was not designed as editing format, and you only can brute force your way so far .... 7) Look if mesa3d tracker has any issue with vaapi decoding crashing cingg. I added one case but it probably was just incomplete implementation in nouveau. I avoided some crashes by disabling those filmstrip icons as fast as I can after loading media. ymmv. I suspect way cingg drives ffmpeg and ffmpeg in turn drives vaapi driver is quite different from usual player/streaming case, but you probably need some thread debugging flags added everywhere and knowledge about how to interpret their output.

1 0

Mesa 24.3 and Vulkan
by Terje J. Hanssen 14 Dec '24

14 Dec '24

In August this year we had a thread "h264 vulkan encoding" initiated by Andrea https://lists.cinelerra-gg.org/pipermail/cin/2024-August/008517.html Mesa 24.3 is released with many Vulkan Driver Improvement, and implementations are expected around the corner in rolling distributions. Some articles about this in Phoronix https://www.phoronix.com/news/Mesa-24.3-Released https://www.phoronix.com/news/Mesa-24.3-rc2 https://www.phoronix.com/news/Mesa-24.3-rc1-Released https://www.phoronix.com/news/Intel-Vulkan-Video-H264-H265 With ffmpeg (7.1) h264/5 vulkan encoder support, I wonder if Cingg is ready to put them into use? ffmpeg -hide_banner -encoders | grep vulkan V....D h264_vulkan H.264/AVC (Vulkan) (codec h264) V....D hevc_vulkan H.265/HEVC (Vulkan) (codec hevc)

3 10

AMD's amf in ffmpeg
by Andrew Randrianasulu 12 Dec '24

12 Dec '24

As Terje requested I looked up into amf compilation, _it seems_ it just require headers at compilation stage BUT amdgpu-pro runtime for any real use, and it goes via vulkan anyway ? https://ffmpeg.org/general.html#AMD-AMF_002fVCE

3 9

Video accelerators support, builds and use
by Terje J. Hanssen 11 Dec '24

11 Dec '24

To clarify some pieces once again, I put up some basic statements or questions: For an end-user to utilize video acceleration support, he/she need a computer with supported graphical hardware with libs/API and drivers for it(?) The libs and drivers can be dynamical linked (enabled) to the system or static built (embedded) in CinGG? So what happened when adding oneVPL (qsv) support to the build system; dynamic linked to system or static added embedded into the build? If oneVPL was dynamic linked, the qsv support may be be distribution specific, or if static built it will be generic available on compliant hardware? Is it correct to say the build machine does not need the specific graphical hardware, but needs the actual graphic libs installed to build Cingg with it? Could in principle similar methods be extended to include broader video acceleration support for AMD/amf and NVIDIA/nvenc? -------------- So a confusing piece if "oneVPL" instead should have been replaced with "libvpl?, because I just read /Note for Users of Intel® oneAPI Video Processing Library (oneVPL) and for Intel® Media SDK/ /https://www.intel.com/content/www/us/en/developer/tools/vpl/overview.html/ /oneVPL is now called the Intel® Video Processing Library (Intel® VPL). The library will no longer be part of the oneAPI specification so that Intel can focus on providing video processing features on Intel GPUs./ In comparision on openSUSE/Slowroll on Intel, there are libvpl(2) (and no oneVPL). /The oneAPI Video Processing Library (oneVPL) provides a single video processing API for encode, decode, and video processing that works across a wide range of accelerators. / ffmpeg similar has --enable-libvpl --enable-vaapi --enable-vdpau --enable-vulkan / /

2 2

Checked into GIT several hardware specific mods
by Phyllis Smith 10 Dec '24

10 Dec '24

Outcome from all of the hard work that Andrew did with Terje for the newer Intel hardware performance enhancements has been checked into GIT. Here is a rough description of what was checked in. - 12 new or replaced "vaapi" encoding render formats (vainfo shows if supported) - 12 new or replaced "qsv" encoding render formats (requires Intel HW build) - patches to ffmpeg.C and ffmpeg.h to support the above - 3 new Termux (Android) encoding render formats - 2 mediacodec which will only work on Android + an mpeg one that works everywhere - mjpegtools patch to accommodate clang compiling continuance There should be no impact to standard usage of CinGG but users need to be especially aware of the fact that using a render encoding format that requires specific hardware/software implementation will give an error. In addition, the standard appimage releases will not provide the availability to use this Intel hardware. This is the same as the inability to use vdpau/vaapi from those same appimages since the appimages are created on a computer without the end users same hardware. I need to update the Manual to make this clear in the Hardware video acceleration section.

1 0

x265 4.1 is out for some time ...
by Andrew Randrianasulu 09 Dec '24

09 Dec '24

https://x265.readthedocs.io/en/4.1/releasenotes.html reverted some of 4.0 api changes, so our ffmpeg patch for this probably not needed anymore if we decide to update. Memleak fixes also. https://bitbucket.org/multicoreware/x265_git/downloads/x265_4.1.tar.gz

1 0