- Cin - Cinelerra-GG Mailing Lists

"Competitors" or do we want to stay alive?
by Andrew Randrianasulu 07 Apr '25

07 Apr '25

Sorry for being frustrated lately. Realities of all this "modern development" vs me tend to be not in favor of our project. I looked up two "other" projects in video editing. MLT is well-known and going on for nearly two decades. In this year release notes we see: https://github.com/mltframework/mlt/releases ===== https://github.com/mltframework/mlt/releases/tag/v7.30.0 Added support for mlt_image_yuv420p10, mlt_image_yuv444p10, and mlt_image_yuv422p16 in avfilter, swscale, and rescale filters. This facilitates using these pixel formats end-to-end when using only FFmpeg producers, certain avfilters, and avformat consumer. This means it is possible to do 10-bit end-to-end on the CPU when being careful to select compatible components and options to avoid conversions. One can pass-through HDR; however, you must set the color_trc and pix_fmt properties on the avformat consumer (see ffmpeg -h full for these values). The avformat consumer automatically converts MLT colorspace (integer value) to FFmpeg's colorspace and color_primaries (unless explicit) options. ==== I take this as "floating point pipeline still not here". Olive branch is even more depressing: https://github.com/olive-editor/olive/issues Last non-CI commit was in ... september 2023? May be developing moved somewhere, but this mean all exiting new bugs and state of being incomplete for few more years, given that even with most of hard work done by OpenGL, OCIO, ffmpeg etc remaining NLE core is not that simple, as it turned out to be. Natron does not do audio, as far as I know, and also hanging on a thread (single developer). Blender is a Big Shot now, with predictably big requirements both in software (try to build it on Slackware) and in hardware. Driven by higher end! And whole "Oh lol X is deprecated, everyone rewrite themselves to Wayland" push. I bet there might be way to resistance say Intel's or Nvidia "HDR on X" proposals in Xwayland but who will do all this work for single application? It took Valve's money effectively to get anything at all done in this area. https://www.phoronix.com/news/X11-DeepColor-Visual-RFC from 2017 .... https://lists.freedesktop.org/archives/wayland-devel/2017-December/036403.h… It took whole 8 years to get into "usable in mpv" state, and producing video usually have higher demands than playing back. I am fairly sure someone will write something in Rust (so it will break in 6 months time because Rust is for Big D Developers who have no problems with such churn) or c++ 25 because c+11 is too archaic and all new courses are about $latest (double meaning of $ here). Again, I do not think this is only individual developers fault, just more complex issue with constant stream of *new* hardware we supposed to fix in field by writing software, somehow (HDR displays are all or nothing, so HD(R), SD and GUI all must be mixed on host side). And general "culture" of individualism (because Silicon Valley, bebe!) and effectively social darwinism (where are XDTV and libquicktime? killed by constant API churn ....). I see no escape from this, at larger scale. And this makes me even more frustrated. cc ffmpeg-user/libav user because otherwise no one from "outside" will ever read this.

4 8

"Unbounded" floating point image manipulation
by Andrew Randrianasulu 04 Apr '25

04 Apr '25

I think this forum post (with still working image inserts) illustrates what Andrea trying to do inside CinGG: https://discuss.pixls.us/t/unbounded-floating-point-pipelines/6483/199 On input image a bit blown out, but after few layers of correction (_as long as you editing software does not clip to 1.0_) you get more details, in final image. Very unavoidable fact that _everything_ (camera, monitor, projector, printer, eyes) all non-linear in their transmission/sensing/emitting lightwaves of different frequencies makes all this quite complex topic!

3 5

possible patch for x265 on cmake 4.0?
by Andrew Randrianasulu 04 Apr '25

04 Apr '25

https://bitbucket.org/multicoreware/x265_git/commits/b354c009a60bcd6d7fc040… not sure libsvtav1 yet ....

4 20

Crash on compile
by Andrea paz 02 Apr '25

02 Apr '25

I tried to do a compile with Andrew's patch but it fails. The patch is not to blame because I tried to do a new compile without the patch and it too fails. The last compilation I did successfully was from the March release. Then I had not attempted any others so I don't know which commit leads to the error. The appimage works, but it is from the same period as my working compile. I attach the cin.log

2 5

New March 31 AppImages
by Phyllis Smith 31 Mar '25

31 Mar '25

So that the Blend Algebra and Blend Program icons would be visible, I updated the AppImages. A few fixes are in too as noted (hopefully technically correct) in the Release Notes shown last here. Plus, I did put Andrea's experimental/testing of Gimp-like overlay algorithms for Addition, Subtraction, Divide, and Multiply using the Blend Algebra plugin at: https:/ cinelerra-gg.org/download/testing/BlendPluginExamples/gimp_.zip . Removed GIT manual reference to Overlays but a little late. *GIT for Cinelerra-GG has the following changes from 03/01/2025-03/31/2025* *Blend Algebra and Blend Program* modifications include the following: - Icons for Cinfinity and Cinfinity2 have been created by *IgorB*. - Code changes by the author, *Georgy*, to both plugins: 1) Buttons =>project, userlib, syslib will also display the target dir when the function name is blank. 2) Refresh will now immediately recompile that function (if any) which is under the timeline cursor when source is newer than shared object. This does not affect the other functions which are not under the cursor and they will still be recompiled on first access. 3) Added “output track” of Blend Algebra which was missing in copy/interpolate functions. 4) Improved robustness of code to prevent race conditions between threads of different windows when closing or resizing dialog menus. *Screencapture* for NetBSD failure has been fixed by adding fallback in case shm capture does not work by *Andrew-R*. This fix may be applicable to other Operating Systems.

1 0

Gimp's blend Vs CinGG's blend
by Andrea paz 27 Mar '25

27 Mar '25

I wanted to try using Blend Algebra to recreate Gimp's blend types inside CinGG. I couldn't because I don't really understand the formulas used by Gimp, or rather, what I understand less are the formulas used by CinGG. I understand that these are pre-multiplied while Gimp's are normal (“straight”). But even without taking this into account, I don't really understand how to compare the formulas (https://docs.gimp.org/3.0/en/gimp-concepts-layer-modes-legacy.html) For example, for the “Multiply” blend we have two different formulas, even though the manual says that the CinGG formula is the same as the Gimp formula: Gimp E = 1/255*(M*I) where “M” (foreground), if I understand correctly, corresponds to our “D” and “I” (background) corresponds to our “S” so it could be written: E = 1/255*(D*S) Instead, CinGG's formula is: [(Sa+Da-Sa*Da), Sc*(1-Da)+Dc*(1-Sa)+Sc*Dc] I cannot compare the two formulas, there is one too many Sc*(1-Da) and one too many Dc*(1-Sa), and a 1/255 is missing. Note: The presence of the parentheses like (1-Sa) is what makes me suspect that CinGG's formulas are “pre-multiplied” (https://en.wikipedia.org/wiki/Alpha_compositing#Straight_versus_premultipli….) There are three types of blends whose formulas differ (and which I wanted to implement via Blend Algebra): (Normal); Addition; Subtract and Divide. Then there are Gimp blends that CinGG doesn't have, but they don't seem interesting to me. PS: I tried to look for the formulas used by CinCV but I don't know where to find them; do they exist anywhere?

2 3

Attempt at fixing screencapture under NetBSD
by Andrew Randrianasulu 21 Mar '25

21 Mar '25

Tried to record desktop, found out screencapture does not work :( It seems even ffmpeg fall back to non-shm capture (slower) under NetBSD, so I tried to to change logic in guicast/bccapture.C so if compiled on netBSD it hard disable shm for capture (playback is fine) Please try on Linux and other systems ... Pay attention to screencapture slowdowns!

2 5

First screencapture under NetBSD!
by Andrew Randrianasulu 17 Mar '25

17 Mar '25

https://cloud.mail.ru/public/DQkK/kpDhJiyY3 9 Mbytes Testing experimental audio support fro qemu-system-ppc (running NetBSD/macppc) under NetBSD/amd64 and of course cinlerra-gg! (with my screencapture fixes) Sadly aiomixer a bit hard to get right on first try, so sound level is LOW. Captured using OSS support via /dev/audio1

2 1

Strange effects with zooming
by Stefan de Konink 14 Mar '25

14 Mar '25

With the recent git version. Did anyone observe strange artifacts on the wave forms when zooming and scrolling too? Like the offset of the wave form moves up or down? -- Stefan

4 6

Blend Algebra and Program icons
by Igor BEGHETTO 04 Mar '25

04 Mar '25

4 8