November 2023 - Cin - Cinelerra-GG Mailing Lists

Forum
by Andrea paz 15 Nov '23

15 Nov '23

Via e-mail I received notification of two forum posts from user mikonass1; but I cannot see them. Do you guys see them? Is it perhaps spam?

2 2

lcms2 patch
by Andrew Randrianasulu 15 Nov '23

15 Nov '23

using code example from https://git.ffmpeg.org/gitweb/ffmpeg.git/blobdiff/f7bab37c8e66f3c77a5fbb5b5… and test files from https://littlecms.com/img/blog/check_lut.jpg https://littlecms.com/img/blog/check_full.jpg I think I was able to make png ffmpeg encoder to add built-in profile, based on (randomly-choose) colorimetry. patch itself needed to be applied with usual patch from directory just above main cinelerra-5.1 level, and not "git am" tar.gz not included due to 7Mb size! Fake colorimetry I used: ======== # this codec creates a set of image files in a directory # use an image2 file name like /dir/p%05d.png flags2=0x80000000 colorspace=smpte2085 color_primaries=bt2020 color_trc=bt2020_12bit color_range=jpeg ==== resulted in this exiftool output: exiftool /dev/shm/4.png ExifTool Version Number : 11.11 File Name : 4.png Directory : /dev/shm File Size : 29 kB File Modification Date/Time : 2023:11:12 19:28:40+03:00 File Access Date/Time : 2023:11:12 19:28:40+03:00 File Inode Change Date/Time : 2023:11:12 19:28:40+03:00 File Permissions : rw-r--r-- File Type : PNG File Type Extension : png MIME Type : image/png Image Width : 400 Image Height : 300 Bit Depth : 8 Color Type : RGB Compression : Deflate/Inflate Filter : Adaptive Interlace : Noninterlaced Pixels Per Unit X : 1 Pixels Per Unit Y : 1 Pixel Units : Unknown Profile Name : icc Profile CMM Type : Little CMS Profile Version : 4.3.0 Profile Class : Display Device Profile Color Space Data : RGB Profile Connection Space : XYZ Profile Date Time : 2023:11:12 16:28:40 Profile File Signature : acsp Primary Platform : Apple Computer Inc. CMM Flags : Not Embedded, Independent Device Manufacturer : Device Model : Device Attributes : Reflective, Glossy, Positive, Color Rendering Intent : Perceptual Connection Space Illuminant : 0.9642 1 0.82491 Profile Creator : Little CMS Profile ID : 0 Profile Description : RGB built-in Profile Copyright : No copyright, use freely Media White Point : 0.9642 1 0.82491 Chromatic Adaptation : 1.04788 0.02292 -0.05022 0.02959 0.99048 -0.01707 -0.00925 0.01508 0.75168 Red Matrix Column : 0.67348 0.27904 -0.00194 Blue Matrix Column : 0.12505 0.04561 0.79684 Green Matrix Column : 0.16566 0.67534 0.02998 Red Tone Reproduction Curve : (Binary data 32 bytes, use -b option to extract) Green Tone Reproduction Curve : (Binary data 32 bytes, use -b option to extract) Blue Tone Reproduction Curve : (Binary data 32 bytes, use -b option to extract) Chromaticity Channels : 3 Chromaticity Colorant : Unknown (0) Chromaticity Channel 1 : 0.70799 0.29201 Chromaticity Channel 2 : 0.17 0.797 Chromaticity Channel 3 : 0.131 0.04601 Image Size : 400x300 Megapixels : 0.120 Note, of course this is WRONG kind of wrong, randomly setting colorimetry tags on encode, but at least it hopefully show that ffmpeg's machinery (6.1) works at least that far. sadly, no of those two test images show correct text, so I guess display icc profile support quite missing step for us ...! but one micro/nanostep at a time?

3 4

Rendering, encoding and decoding with Nvidia
by David Koski 14 Nov '23

14 Nov '23

I've started using CinGG-20230930-x86_64.AppImage and I'm getting the hang of the basics. I have gotten my nvidia hardware engaged with Cinelerra for encoding but still the rendering is not all that great and my CPUs are loaded down. On finding that I need environment variables to engage decoding, I thought maybe that is hindering my rendering. So with the following setting I started a rendering job: export CIN_HW_DEV=vdpau Now I get an error: recvmsg(22, Failed to open VDPAU backend libvdpau_nvidia.so: cannot open shared object file: No such file or directory libvdpau is installed: ii libvdpau-va-gl1 0.4.2-1build2:amd64 VDPAU driver with OpenGL/VAAPI backend ii libvdpau1 1.4-3build2:amd64 Video Decode and Presentation API for Unix (libraries) The files exist: drwxr-xr-x 21 root root 4096 Nov 1 19:29 /./ drwxr-xr-x 154 root root 12288 Nov 8 21:08 /etc/ -rw-r--r-- 1 root root 51 Apr 9 2020 /etc/vdpau_wrapper.cfg drwxr-xr-x 14 root root 4096 Mar 8 2023 /usr/ drwxr-xr-x 118 root root 4096 Nov 9 20:40 /usr/lib/ drwxr-xr-x 125 root root 122880 Nov 8 21:08 /usr/lib/x86_64-linux-gnu/ lrwxrwxrwx 1 root root 17 Mar 24 2022 /usr/lib/x86_64-linux-gnu/libvdpau.so.1 -> libvdpau.so.1.0.0 -rw-r--r-- 1 root root 18592 Mar 24 2022 /usr/lib/x86_64-linux-gnu/libvdpau.so.1.0.0 drwxr-xr-x 2 root root 4096 Sep 30 09:40 /usr/lib/x86_64-linux-gnu/vdpau/ lrwxrwxrwx 1 root root 23 Mar 24 2022 /usr/lib/x86_64-linux-gnu/vdpau/libvdpau_trace.so.1 -> libvdpau_trace.so.1.0.0 -rw-r--r-- 1 root root 59544 Mar 24 2022 /usr/lib/x86_64-linux-gnu/vdpau/libvdpau_trace.so.1.0.0 drwxr-xr-x 306 root root 12288 Nov 8 21:08 /usr/share/ drwxr-xr-x 145 root root 4096 Nov 8 21:07 /usr/share/bug/ -rwxr-xr-x 1 root root 553 Jan 6 2021 /usr/share/bug/libvdpau1* drwxr-xr-x 2302 root root 81920 Nov 8 21:08 /usr/share/doc/ drwxr-xr-x 2 root root 4096 Jul 27 09:19 /usr/share/doc/libvdpau1/ -rw-r--r-- 1 root root 1447 Mar 24 2022 /usr/share/doc/libvdpau1/changelog.Debian.gz -rw-r--r-- 1 root root 3208 Jan 6 2021 /usr/share/doc/libvdpau1/copyright Regards, David Koski david(a)kosmosisland.com

5 10

Re: [Cin] Cinelerra-GG packages build farm
by Phyllis Smith 12 Nov '23

12 Nov '23

This is very much appreciated as there are some definite advantages to having packages as opposed to the appimage -- in my opinion the most beneficial is the ability to get the updates applied automatically when there is a new version available (if enabled by the user). I assume it is OK for me to add this to the manual in the section: Distro with CINELERRA-GG Included . On Sun, Oct 8, 2023 at 5:14 AM Андрей Спицын <spitsyn.andrey(a)gmail.com> wrote: > Hello Phyllis and everyone, > > I've created a build farm for cinelerra deb and rpm packages. Although it > is at an early stage of development, it can build packages on every git > change in the main repo. Feel free to use these packages at > https://github.com/einhander/cin-gg-packages/releases. Note that releases > correspond to a build date, not a git commit date. > Current build hosts are debian 12 and alma linux 8. I'm open to > suggestions for a new distro and packages. > > > Best regards, > Andrey Spitsyn >

4 10

openexr test images
by Andrew Randrianasulu 11 Nov '23

11 Nov '23

https://openexr.com/en/latest/_test_images/index.html esp. https://raw.githubusercontent.com/AcademySoftwareFoundation/openexr-images/… where description says "This RGB image contains a number of rather bright rings, with pixel values over 1000, on a gray background. The image is useful for testing how filtering and resampling algorithms react to high- dynamic-range data. (Some filters, for example, convolution kernels with negative lobes, tend to produce objectionable artifacts near high-contrast edges.) Note that the rings in the image are smooth, although on most displays clamping of the pixel values introduces aliasing artifacts. To see that the rings really are smooth, view the image with and exposure of -10."

1 0

Fwd: [heroineworshiper/hvirtual] question about clipping in Histogram plugin (Issue #8)
by Andrew Randrianasulu 10 Nov '23

10 Nov '23

well, Andrea, can I copy your email to GitHub issues or you prefer to explain it there yourself? I think your use case goes a bit beyond just testing and into actual workflow category? ---------- Forwarded message --------- От: Adam Williams <notifications(a)github.com> Date: пт, 10 нояб. 2023 г., 21:16 Subject: Re: [heroineworshiper/hvirtual] question about clipping in Histogram plugin (Issue #8) To: heroineworshiper/hvirtual <hvirtual(a)noreply.github.com> Cc: Andrew Randrianasulu <randrianasulu(a)gmail.com>, Author < author(a)noreply.github.com> The big problem with histogram is histograms are expected to clamp. The graph has to be changed to show a Y range beyond 0-1, you need some more parameters to define min & max Y. You could make a new histogram plugin which implemented the new options. It gets a bit bloated so the question is again why you need histogram to support more than 0-1? Sent: Thursday, November 09, 2023 at 11:14 AM From: "Andrew Randrianasulu" ***(a)***.***> To: "heroineworshiper/hvirtual" ***(a)***.***> Cc: "Adam Williams" ***(a)***.***>, "Comment" ***(a)***.***> Subject: Re: [heroineworshiper/hvirtual] question about clipping in Histogram plugin (Issue #8) A case of is it worth the effort to support values <0 & >1 & the answer was no. If you can find a good reason & actually fix everything including the opengl functions no-one uses, maybe it could work. well, I looked into playback3d.C and some shaders there surely uses clip? But biggest problem for me I do not know how to init fp32 fbo/pbo, so it all remained clipped. But just switching textures to 32f made big difference with test file Andrea provided and fade out. https://www.youtube.com/watch?v=7GOhtJciHBk I tried to follow advice given in https://community.khronos.org/t/glx-ati-pixel-format-float-under-linux/4768… (but using more modern GLX_RGBA_FLOAT_BIT_ARB). use-case is exr/tiff from some photography software or rasterizer, where values > 1.0f exist. (ffmpeg seems to convert most/all video to integer formats, but may be in last ffmpeg they added it - https://git.ffmpeg.org/gitweb/ffmpeg.git/commit/479747645f795b6f4f376578ea1… - "avutil/pixfmt.h: add native-endian RGB32F and RGBA32F formats") — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: ***(a)***.***> — Reply to this email directly, view it on GitHub <https://github.com/heroineworshiper/hvirtual/issues/8#issuecomment-18061989…>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AJSS7TJ5T7MREFWQWTBCADLYD…> . You are receiving this because you authored the thread.Message ID: <heroineworshiper/hvirtual/issues/8/1806198961(a)github.com>

2 7

Clipping to 1.0 investigation.
by Andrew Randrianasulu 09 Nov '23

09 Nov '23

I tried cingg compiled on Nov, 1 2020 and issues still here - you need native exr reader to see it with Andrea's sample. Either by compositor's vectoroscope or color picker - in clipping case it will show 1.0 for much of window area, in non-clipping mode (no plugins) it will show up to 60.something in rgb channels! Sadly, 2018 version had no vectoroscope on compositor, need to double-test if native exr reader + any plugin will fail there in the same way via color picker.....

3 15

Trying to understand colour, try N!
by Andrew Randrianasulu 09 Nov '23

09 Nov '23

https://blender.stackexchange.com/questions/46825/render-with-a-wider-dynam… well, this one says EXR can hold quite a big value per channel, way above 1.0f, obviously. But for displaying (and encoding?) it must be compressed back to device-specific 0 - 1.0 range. Thing is, if I open exr example file in gimp 2.10 and re-export it - it retain its above 100% values. It does not do this if I do this same roundtrip via cingg. So yea, we seems to be clamping at output. Still not sure where, may be in fader? (where fade can't be more than 100%) . May be its time to ask Lion (Adam).

3 12

GIT check in of Tiff 4.6.0
by Phyllis Smith 08 Nov '23

08 Nov '23

Updated tiff from 4.5.0 to 4.6.0 and checked into GIT. Ran 2 render tests of different video files and loaded a bunch of previously created .tif files with no seen anomalies. (I should have tested on Ubuntu 16 and Debian 32-bit also, but will do so tomorrow). Tried to update libjpeg-turbo from 2.1.5.1 to 3.0.1 but having Errors in building which I have not figured out yet but will keep trying.

3 3

Floating point numbers between 0 and 1.0
by Andrew Randrianasulu 07 Nov '23

07 Nov '23

https://stackoverflow.com/questions/17949796/how-many-unique-values-are-the… ppl here come up with this number: 1,065,353,215. so ... even between 0.0f and 1.0f there is quite a number of ... numbers. Still, out of range numbers useful in photo work like here https://ninedegreesbelow.com/photography/gimp-tone-map-with-levels.html found via bugzilla bug filled for gimp: https://bugzilla.gnome.org/show_bug.cgi?id=737925 "*Bug 737925* <https://bugzilla.gnome.org/show_bug.cgi?id=737925> - Some filters / operations provide float values superior to 1.0" on CinGG plugins I found only exactly one instance of clamp line as Andrea suggested: grep "CLAMP(output, 0, 1.0)" -r plugins plugins/histogram/histogram.C: CLAMP(output, 0, 1.0); plugins/histogram/histogram.C: CLAMP(output, 0, 1.0); and nothing more.

2 1