- Cin - Cinelerra-GG Mailing Lists

crash in hevc alpha decoding on second asset resize in cinelerra-gg
by Andrew Randrianasulu 15 May '26

15 May '26

Unfortunately after patching our ffmpeg 8.1 with https://source.ffmpeg.org/gitweb/ffmpeg.git/commit/3b939ced79655ed084e6bebc… I still see this crash on x265 created hevc + alpha file: Thread 301 "av:hevc:df1" received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fff82ffd700 (LWP 6407)] alloc_frame (s=s@entry=0x7fff401ae8c0, l=l@entry=0x7fff401af698) at libavcodec/hevc/refs.c:194 194 AVFrame *base = s->layers[0].cur_frame->f; (gdb) bt full #0 alloc_frame (s=s@entry=0x7fff401ae8c0, l=l@entry=0x7fff401af698) at libavcodec/hevc/refs.c:194 alpha = 0x7fff0406bdc0 base = <optimized out> frame = <optimized out> vps = <optimized out> i = <optimized out> j = <optimized out> ret = <optimized out> #1 0x000000000100b7f3 in ff_hevc_set_new_ref (s=s@entry=0x7fff401ae8c0, l=l@entry=0x7fff401af698, poc=123) at libavcodec/hevc/refs.c:226 ref = <optimized out> i = <optimized out> no_output = <optimized out> #2 0x0000000000fd7dd6 in hevc_frame_start (nal_idx=1, l=0x7fff401af698, s=0x7fff401ae8c0) at libavcodec/hevc/hevcdec.c:3337 pps = 0x7fff402c5640 new_sequence = 0 prev_layers_active_output = 1 ret = <optimized out> sps = 0x7fff402c0640 pic_size_in_ctb = 2806 prev_layers_active_decode = 3 #3 decode_slice (gb=0x7fff82ffccc0, nal_idx=1, s=0x7fff401ae8c0) at libavcodec/hevc/hevcdec.c:3580 ---Type <return> to continue, or q <return> to quit--- layer_idx = <optimized out> l = 0x7fff401af698 ret = <optimized out> #4 decode_nal_unit (nal_idx=1, s=0x7fff401ae8c0) at libavcodec/hevc/hevcdec.c:3663 nal = <optimized out> gb = { buffer = 0x7fff74000e9e "\002\t\244\366W\346\024\210\022\003\001*$\004A\370\326\372\303\246\260\314\353\214c).z_=\351\021~\304\f\020'\247\222#gb\033/_\325H\261L\211v\037h\206\301\272\256[\"\377/\204|W\002pm\347\265\322\177e\030\220K\216!\224\b+mI\216\017\307\020\177\215#m\353\200\235W\215E\352K\031\275k\346\322\032g>'0\246C\274U\263v\236\215s\265y\312\245\t\255\203T\376\213Y\037\177`y\005\320\016\022R\036\035I\261\204w\202\271\311\311%S\313\322\370\306\250\241\065JW\224>1\247\234-_\243\317XO\361\003 \031\363\246X\363(s\252\214Pf\331\307o\352\266\201/\360c\020\fo\331\004\064/b"..., index = 136, size_in_bits = 5878, size_in_bits_plus8 = 5886} ret = <optimized out> #5 decode_nal_units (length=<optimized out>, buf=<optimized out>, s=0x7fff401ae8c0) at libavcodec/hevc/hevcdec.c:3779 nal = <optimized out> i = 1 ret = <optimized out> eos_at_start = <optimized out> flags = <optimized out> #6 hevc_receive_frame (avctx=<optimized out>, frame=<optimized out>) at libavcodec/hevc/hevcdec.c:3881 s = 0x7fff401ae8c0 avci = <optimized out> avpkt = <optimized out> ---Type <return> to continue, or q <return> to quit--- ret = <optimized out> sd = <optimized out> sd_size = 0 #7 0x0000000000f1263f in ff_decode_receive_frame_internal (avctx=avctx@entry=0x7fff402c6c00, frame=frame@entry=0x7fff880f2f40) at libavcodec/decode.c:625 avci = 0x7fff402c6fc0 ret = <optimized out> #8 0x000000000117b47a in frame_worker_thread (arg=0x7fff4006ff00) at libavcodec/pthread_frame.c:291 ret = 0 p = 0x7fff4006ff00 avctx = 0x7fff402c6c00 #9 0x00007ffff540755a in start_thread () from /lib64/libpthread.so.0 No symbol table info available. #10 0x00007fffedd02e5f in clone () from /lib64/libc.so.6 No symbol table info available. There was another patch moving some of those i local variables into loop, but I am not sure if or how it may help here ? Will test more. (I wonder if disabling threaded decode will avoid this?)

1 1

Update x265 to 4.2, enable HEVC alpha encoding
by Andrew Randrianasulu 14 May '26

14 May '26

https://bitbucket.org/multicoreware/x265_git/downloads/x265_4.2.tar.gz Patch (git am) attached put file in thirdparty/src, patch sources, autogen.sh then build as usual In theory now for h265.mp4 you can select yuva420/yuva420p10, but I am not sure it works as intended ... decoding says alpha present (if I set loglevel=debug in decode.opts) but other sources including ffprobe from ffmpeg-8.1 detect just yuvj420p ? Tested only on Rosa 2016 x86-64 chroot

2 3

hevc alpha decoding in cingg
by Andrew Randrianasulu 12 May '26

12 May '26

Sending to ffmpeg-devel because I do not know if my workaround makes any sense? For some reason hevc alpha decoding was not working for us. I picked up https://source.ffmpeg.org/gitweb/ffmpeg.git/commit/eedf8f0165fe6523be692c5a… avcodec/hevc: workaround hevc-alpha videos generated by VideoToolbox but videos were still not showing their alpha channel. So I added this to libavcodec/hevc/hevcdec.c: --- ./libavcodec/hevc/hevcdec.c 2026-03-16 18:13:04.000000000 +0000 +++ ./libavcodec/hevc/hevcdec.c.new 2026-05-09 05:22:45.098799320 +0000 @@ -712,9 +712,11 @@ break; } + if (alpha_fmt != AV_PIX_FMT_NONE) *fmt++ = alpha_fmt; - *fmt++ = sps->pix_fmt; + if(alpha_fmt == AV_PIX_FMT_NONE) + *fmt++ = sps->pix_fmt; *fmt = AV_PIX_FMT_NONE; // export multilayer information from active VPS to the caller, ==== basically only adding main layer (?) formats if there was no alpha) It seems to work with two hevc-alpha files and one normal hevc file May be this is only cingg specific problem - due to way we pick pixel formats on decoding? ffmpeg itself works ok without this workaround Even with this patch ffmpeg cli still picks correct format, on encoding to png: thirdparty/ffmpeg-8.1/ffmpeg -i /dev/shm/av1--test.mp4 -vframes 1 1.png 1.png contain transparency test files from https://developer.apple.com/documentation/avfoundation/using-hevc-video-wit… more test files: https://lf-tk-sg.ibytedtos.com/obj/tcs-client-sg/resources/video_demo_hevc.… possibly related https://trac.ffmpeg.org/ticket/10901 HEVC with alpha to PNG does not maintain alpha

3 2

Patch: do not build turbojpeg-specific library, we do not use it
by Andrew Randrianasulu 10 May '26

10 May '26

Hopefully I got this correct, did partial rebuild on x86-64 Rosa 2016 Will try to integrate x265 4.2.0 next

2 2

HEIC tiled images
by Andrew Randrianasulu 09 May '26

09 May '26

Today I learned some HEIC images present themselves as bunch (up to 49!) of smaller tiles in cinelerra-GG and mpv. https://github.com/mpv-player/mpv/issues/13585 ffmpeg itself automerges them now, so this work (ffmpeg git cli): /dev/shm/ffmpeg/ffmpeg -i /dev/shm/IMG_4453.HEIC /dev/shm/grid.png bash-5.1$ /dev/shm/ffmpeg/ffmpeg -i /dev/shm/IMG_4453.HEIC /dev/shm/grid.png ffmpeg version N-123229-gf2e5eff3ff Copyright (c) 2000-2026 the FFmpeg developers built with gcc 11.2.0 (GCC) configuration: --enable-opencl --disable-debug --enable-libx265 --enable-libx264 --enable-gpl --enable-libplacebo --enable-vulkan --enable-libshaderc --enable-libzimg --enable-libaom --enable-libdav1d --enable-libsoxr --enable-libfontconfig --enable-libfreetype --enable-libfribidi --enable-gnutls --enable-libass --enable-libbluray --enable-libcdio --enable-frei0r --enable-libgsm --enable-openal --enable-libopus --enable-librtmp --enable-libsnappy --enable-libspeex --enable-libssh --enable-libtheora --enable-libtwolame --enable-libv4l2 --enable-libvidstab --enable-libvorbis --enable-libvpx --enable-libwebp libavutil 60. 30.100 / 60. 30.100 libavcodec 62. 30.100 / 62. 30.100 libavformat 62. 15.100 / 62. 15.100 libavdevice 62. 4.100 / 62. 4.100 libavfilter 11. 17.100 / 11. 17.100 libswscale 9. 7.100 / 9. 7.100 libswresample 6. 4.100 / 6. 4.100 Input #0, mov,mp4,m4a,3gp,3g2,mj2, from '/dev/shm/IMG_4453.HEIC': Metadata: major_brand : heic minor_version : 0 compatible_brands: mif1heic Duration: N/A, start: 0.000000, bitrate: N/A Stream group #0:0[0x31]: Tile Grid: hevc (Main Still Picture) (hvc1 / 0x31637668), yuvj420p(pc), 4032x3024 (default) Side data: EXIF metadata: (2012 bytes) Stream #0:48[0x32]: Video: hevc (Main Still Picture) (hvc1 / 0x31637668), yuvj420p(pc), 320x240, 1 fps, 1 tbr, 1 tbn Side data: ICC Profile: (548 bytes) Stream mapping: Stream #0:0 (hevc) -> xstack Stream #0:1 (hevc) -> xstack Stream #0:2 (hevc) -> xstack Stream #0:3 (hevc) -> xstack Stream #0:4 (hevc) -> xstack Stream #0:5 (hevc) -> xstack Stream #0:6 (hevc) -> xstack Stream #0:7 (hevc) -> xstack Stream #0:8 (hevc) -> xstack Stream #0:9 (hevc) -> xstack Stream #0:10 (hevc) -> xstack Stream #0:11 (hevc) -> xstack Stream #0:12 (hevc) -> xstack Stream #0:13 (hevc) -> xstack Stream #0:14 (hevc) -> xstack Stream #0:15 (hevc) -> xstack Stream #0:16 (hevc) -> xstack Stream #0:17 (hevc) -> xstack Stream #0:18 (hevc) -> xstack Stream #0:19 (hevc) -> xstack Stream #0:20 (hevc) -> xstack Stream #0:21 (hevc) -> xstack Stream #0:22 (hevc) -> xstack Stream #0:23 (hevc) -> xstack Stream #0:24 (hevc) -> xstack Stream #0:25 (hevc) -> xstack Stream #0:26 (hevc) -> xstack Stream #0:27 (hevc) -> xstack Stream #0:28 (hevc) -> xstack Stream #0:29 (hevc) -> xstack Stream #0:30 (hevc) -> xstack Stream #0:31 (hevc) -> xstack Stream #0:32 (hevc) -> xstack Stream #0:33 (hevc) -> xstack Stream #0:34 (hevc) -> xstack Stream #0:35 (hevc) -> xstack Stream #0:36 (hevc) -> xstack Stream #0:37 (hevc) -> xstack Stream #0:38 (hevc) -> xstack Stream #0:39 (hevc) -> xstack Stream #0:40 (hevc) -> xstack Stream #0:41 (hevc) -> xstack Stream #0:42 (hevc) -> xstack Stream #0:43 (hevc) -> xstack Stream #0:44 (hevc) -> xstack Stream #0:45 (hevc) -> xstack Stream #0:46 (hevc) -> xstack Stream #0:47 (hevc) -> xstack xstack:default -> Stream #0:0 (png) Press [q] to stop, [?] for help [swscaler @ 0xf3300040] deprecated pixel format used, make sure you did set range correctly Output #0, image2, to '/dev/shm/grid.png': Metadata: major_brand : heic minor_version : 0 compatible_brands: mif1heic encoder : Lavf62.15.100 Stream #0:0: Video: png, rgb24(pc, gbr/unknown/unknown, progressive), 4032x3024, q=2-31, 200 kb/s, 1 fps, 1 tbn Metadata: encoder : Lavc62.30.100 png Side data: ICC Profile: (548 bytes) EXIF metadata: (2012 bytes) [image2 @ 0xb8a1dc0] The specified filename '/dev/shm/grid.png' does not contain an image sequence pattern or a pattern is invalid. [image2 @ 0xb8a1dc0] Use a pattern such as %03d for an image sequence or use the -update option (with -frames:v 1 if needed) to write a single image. [out#0/image2 @ 0xbeb0e80] video:9222KiB audio:0KiB subtitle:0KiB other streams:0KiB global headers:0KiB muxing overhead: unknown frame= 1 fps=0.1 q=-0.0 Lsize=N/A time=00:00:01.00 bitrate=N/A speed=0.0853x elapsed=0:00:11.73 https://trac.ffmpeg.org/ticket/11170 was closed with this ffmpeg commit in oct 2025 https://git.ffmpeg.org/gitweb/ffmpeg.git/commit/5ff8395e7806ad27743829b0470… "fftools/ffmpeg_demux: create a filtegraph to merge HEIF tiles automatically" some test files from https://github.com/nokiatech/heif/tree/gh-pages/content/images decode in cingg as "just" 8 or so streams/tracks, you can disable them one by one and del low res versions, and combine hires images on single track manually) So much new technology :}

1 0

OpenColorIO
by Andrea paz 09 May '26

09 May '26

I'm starting a new post because the one about Lcms2 is already too long. Now ./configure works for me with --with-opencolorio. However, when I start CinGG for the first time, I get these lines: [OpenColorIO Info]: Color management disabled. (Specify the $OCIO environment variable to enable.) [ocio_616 @ 0x562a521bd680] Error: Config or arguments invalid. [ocio_616 @ 0x562a521bd680] Input color space is null [ocio_616 @ 0x562a521bd680] Display is null [ocio_616 @ 0x562a521bd680] View is null [ocio_616 @ 0x562a521bd680] Failed to create OCIO processor. PluginFFilter::new_ffilter(ocio) And the opencolorio filter doesn't show up. I'm not sure which environment variable to use. Trying to use $OCIO prevents CinGG from starting, saying that this variable doesn't exist.

2 4

New patch to alter file loading from command line
by Andrew Randrianasulu 09 May '26

09 May '26

I thought it may be useful to alter default "Add filenames from command line to new track" behavior, and made new cmd line switch "-l", that can be set to anything, but useful values I tested are: 2 - load concatenated on one track (for jpegs) 6 - load as resources only Default is 1 as before Usage: bin/cin -l 6 ~/*.jpg or bin/cin -l 2 ~/*.jpg Please test if possible, and even merge if it doesn't add regressions

3 4

alpha-hevc decoding workaround from ffmpeg.git
by Andrew Randrianasulu 09 May '26

09 May '26

https://source.ffmpeg.org/gitweb/ffmpeg.git/commit/eedf8f0165fe6523be692c5a… I saved this patch and applied to our ffmpeg. Now ffmpeg itself correctly decodes puppets_with_alpha_hevc.mov file from test project at https://developer.apple.com/documentation/avfoundation/using-hevc-video-wit… thirdparty/ffmpeg-8.1/ffmpeg -i /dev/shm/HEVC-Videos-With-Alpha-Exports/Resources/puppets_with_alpha_hevc.mov -vframes 1 1.png but in cinelerra-gg I still have fully-black background (prores sample from same project works ok)

1 0

Patch: also accept 'q' in tipwindow as close event
by Andrew Randrianasulu 08 May '26

08 May '26

just for consistency - I have tipwindow set to 1 in my 64-bit chroot, so I often q-q all windows by automatic muscle memory

2 2

lcms2 usage
by Andrew Randrianasulu 06 May '26

06 May '26

https://stackoverflow.com/questions/22561176/lcms2-convert-cmyk-to-rgb-thro… this was answered like this === copy === hTransform = cmsCreateTransform(hInProfile, TYPE_CMYK_8, hOutProfile, TYPE_ARGB_8, // ARGB is most likely for QImage INTENT_PERCEPTUAL, 0); uchar *in = outImage->bits(); // bits() is equivalent to scanLine(0) uchar *out = outImage->bits(); cmsDoTransform(hTransform, in, out, outImage->byteCount() / sizeof(QRgb)); === official article show bigger switch https://littlecms.com/blog/2020/12/09/using-lcms2-on-qt/ /** * Convert from Qt format to lcms2 format */ static cmsUInt32Number toLcmsFormat(QImage::Format fmt) { switch (fmt) { case QImage::Format_ARGB32: // (0xAARRGGBB) case QImage::Format_RGB32: // (0xffRRGGBB) return TYPE_BGRA_8; case QImage::Format_RGB888: return TYPE_RGB_8; // 24-bit RGB format (8-8-8). case QImage::Format_RGBX8888: case QImage::Format_RGBA8888: return TYPE_RGBA_8; case QImage::Format_Grayscale8: return TYPE_GRAY_8; case QImage::Format_Grayscale16: return TYPE_GRAY_16; case QImage::Format_RGBA64: case QImage::Format_RGBX64: return TYPE_RGBA_16; case QImage::Format_BGR888: return TYPE_BGR_8; default: return 0; } } so I think we can add support for rgb8 timeline format at least, in same place where swscale call done in our ffmpeg.C in ffmpeg we have libavfilter/vf_iccdetect.c #include <lcms2.h> const AVFrameSideData *sd; cmsHPROFILE profile; sd = av_frame_get_side_data(frame, AV_FRAME_DATA_ICC_PROFILE); profile = cmsOpenProfileFromMemTHR(s->icc.ctx, sd->data, sd->size); cmsCloseProfile(profile); Full documentation in pdf form https://www.littlecms.com/LittleCMS2.18%20tutorial.pdf TYPE_RGB_FLT/TYPE_RGBA_FLT exist in /usr/include/lcms2.h so rgba-float should be supportable too. I'll try to write easy detect part for configure first.

3 25