- Cin - Cinelerra-GG Mailing Lists

Discovered problem with future ffmpeg 8.0
by Phyllis Smith 13 Jun '25

13 Jun '25

After applying current ffmpeg.git as of May 25: ae0f71a387f05be7ac90d545af4d3aeb8e1cca89 and applying Andrew's patches: 0001-Add-check-for-nb_inputs-before-calling-avfilter_get_.patch 0002-Same-possible-crash-fix-in-ffstreamaudio-for-dynamic.patch 0003-Do-not-filter-out-dynamic_input-filters.patch 0004-HACK-ffmpeg-8.0-build.patch CinGG builds but none of the ffmpeg plugins work. The Video plugins crash (did not try them all, but a few like F_boxblur) when added to the timeline and then click on "Show Controls. The Audio plugins do not crash but you get the warning tone only when you play them (again, I did not try them all, but F_chorus was one of them). I have not looked at the code yet, but most likely it is just an oversight in trying to get it to compile. Also, line 77 in 0004-HACK-ffmpeg-8.0-build.patch has "#esle" instead of "else".

2 4

mpeg2hd422p.m2ts - groundwork for a high-end preset
by Terje J. Hanssen 13 Jun '25

13 Jun '25

While the newer h.264 and hevc are known for best compression, mpeg2 may have better quality at video bitrate > 30Mbps This example capture a raw video stream (optional tape playback) from the camera sensor over HDMI out and transcode audio to PCM into a high-end mpeg2hd422p.m2ts -input_format yuyv422 -video_size hd1080 possibly may be commented out, as the encoder default uses the maximum capability it get from the capture card ms2130 and driver v4l2 mpeg2video is difficult to get quit perfect with ffmpeg; muxing overhead: 6.206455% is the minimum I have achieved as in this case. At this high bitrate, 39Mbps here, a combination with segmented recording is suggested and preferred to control the file sizes (segments) and eventually corrupted parts better. ffmpeg -hide_banner -f v4l2 -framerate 50 -i /dev/video0 -f alsa -ac 2 -i hw:CARD=UHD,DEV=0 -c:v mpeg2video -refs 1 -bf 2 -b:v 35M -maxrate 35M -minrate 35M -bufsize 38M -muxrate 38M -dc 10 -c:a pcm_bluray -mpegts_m2ts_mode 1 -t 30 -f mpegts mpeg2hd422p_35M.m2ts Input #0, video4linux2,v4l2, from '/dev/video0': Duration: N/A, start: 79.422429, bitrate: 1658880 kb/s Stream #0:0: Video: rawvideo (YUY2 / 0x32595559), yuyv422, 1920x1080, 1658880 kb/s, 50 fps, 50 tbr, 1000k tbn [aist#1:0/pcm_s16le @ 0x55e2664e95c0] Guessed Channel Layout: stereo Input #1, alsa, from 'hw:CARD=UHD,DEV=0': Duration: N/A, start: 1749208971.359170, bitrate: 1536 kb/s Stream #1:0: Audio: pcm_s16le, 48000 Hz, stereo, s16, 1536 kb/s Stream mapping: Stream #0:0 -> #0:0 (rawvideo (native) -> mpeg2video (native)) Stream #1:0 -> #0:1 (pcm_s16le (native) -> pcm_bluray (native)) Press [q] to stop, [?] for help [mpeg2video @ 0x55e2664d0500] Warning vbv_delay will be set to 0xFFFF (=VBR) as the specified vbv buffer is too large for the given bitrate! Output #0, mpegts, to 'mpeg2hd422p_35M.m2ts': Metadata: encoder : Lavf61.7.100 Stream #0:0: Video: mpeg2video (4:2:2), yuv422p(tv, progressive), 1920x1080, q=2-31, 35000 kb/s, 50 fps, 90k tbn Metadata: encoder : Lavc61.19.101 mpeg2video Side data: cpb: bitrate max/min/avg: 35000000/35000000/35000000 buffer size: 38000000 vbv_delay: N/A Stream #0:1: Audio: pcm_bluray, 48000 Hz, stereo, s16, 128 kb/s Metadata: encoder : Lavc61.19.101 pcm_bluray [out#0/mpegts @ 0x55e2664e9740] video:128445KiB audio:5648KiB subtitle:0KiB other streams:0KiB global headers:0KiB muxing overhead: 6.206455% frame= 1500 fps= 50 q=3.5 Lsize= 142416KiB time=00:00:29.98 bitrate=38915.0kbits/s speed=0.998x

3 8

DV50 - the forgotten optimal codec for digitizing analog Hi8/S-video ?
by Terje J. Hanssen 10 Jun '25

10 Jun '25

In case this also may be something for a CinGG "dv50.mov" preset: Long time ago, Renomath "Video with Linux, Workflow Resources" by Eric Olson, catched my interest (then unobtainable) DV50 when I read Hi8 Tape to DVD Video Workflow with Linux <https://renomath.org/video/linux/hi8/> /If you have a faster system you may want to consider capturing using the DV50 codec using a 4:2:2 color space. In order to do this, libng from the xawtv project needs to be patched to recognize the dv50 codec. Using dv50 as the capture codec really allows the Linux solution to surpass a stand-along DVD recorder. It is also possible to capture the video using ffmpeg or lavrec. / Cinelerra Version 4.x yuv4mpeg Patch <https://renomath.org/video/linux/cinelerra/> /Patch for dnxhd and dv50 Quicktime Input, which add dnxhd and dv50 input to Cinelerra. The advantage of dnxhd and dv50 over mjpeg is native support of 1080i, 4:2:2 chroma and greater encoding quality due to more flexible quantization and compression methods. ------- /While ffv1.mkv often is suggested as preservation master file for migrating analog VHS materials, a mezzanin file has also been suggested using File Wrapper: QuickTime (.mov), Video Stream: DV50, Audio Stream: 48Khz/16-bit PCM https://memoriav.ch/wp-content/uploads/2014/07/VARRFP.pdf ---- FFmpeg documentation has little documentation about dv50, which seemingly is less used: https://ffmpeg.org/ffmpeg.html#Main-options https://ffmpeg.org/pipermail/ffmpeg-user/2023-June/056548.html -target type (output) Specify target file type (vcd, svcd, dvd, dv, *dv50*). type may be prefixed with *pal-*, ntsc- or film- to use the corresponding standard. All the format options (bitrate, codecs, buffer sizes) are then set automatically. The dv50 target is identical to the dv target except that the pixel format set is *yuv422p* for all three standards. ---------- I have tried to experiment with dv50 encoding via my ms2130 capture and v4l2 driver. First with just "c:v dvvideo -target dv50" which resulted in "25fps" and 50Mbps, but this caused jumping video images. I expect that issue was due to a unsupported frame rate. Some AI search later told me that DVCPRO50 (aka DV50) possibly optional could shoot in 50fps, so I tried to add "-r 50" in the output, which resulted in twice as big output file with 50fps as follows, but steady video images. ffmpeg -hide_banner -f v4l2 -i /dev/video0 -f alsa -ch_layout stereo -i hw:CARD=UHD,DEV=0 -codec:v dvvideo -b:v 50M -minrate 50M -maxrate 50M -target pal-dv50 -r 50 -c:a copy -t 10 -f mov DV50p+pcm.mov Input #0, video4linux2,v4l2, from '/dev/video0': Duration: N/A, start: 174.116728, bitrate: 1990656 kb/s Stream #0:0: Video: rawvideo (YUY2 / 0x32595559), yuyv422, 1920x1080, 1990656 kb/s, 60 fps, 60 tbr, 1000k tbn Input #1, alsa, from 'hw:CARD=UHD,DEV=0': Duration: N/A, start: 1749543254.028464, bitrate: 1536 kb/s Stream #1:0: Audio: pcm_s16le, 48000 Hz, stereo, s16, 1536 kb/s [vost#0:0/dvvideo @ 0x55ecffd02f80] Multiple -r options specified for stream 0, only the last option '-r 50' will be used. Stream mapping: Stream #0:0 -> #0:0 (rawvideo (native) -> dvvideo (native)) Stream #1:0 -> #0:1 (copy) Press [q] to stop, [?] for help [dvvideo @ 0x7f5c30706300] Only top-left chroma location is supported in DV, input value is: unspecified ...... [dvvideo @ 0x55ecffce9440] Only top-left chroma location is supported in DV, input value is: unspecified Output #0, mov, to 'DV50p+pcm.mov': Metadata: encoder : Lavf61.7.100 Stream #0:0: Video: dvvideo (dvcp / 0x70637664), yuv422p(tv, progressive), 720x576, q=2-31, 50000 kb/s, 50 fps, 12800 tbn Metadata: encoder : Lavc61.19.101 dvvideo Stream #0:1: Audio: pcm_s16le (sowt / 0x74776F73), 48000 Hz, stereo, s16, 1536 kb/s [out#0/mov @ 0x55ecffd02700] video:140625KiB audio:1875KiB subtitle:0KiB other streams:0KiB global headers:0KiB muxing overhead: 0.026991% frame= 500 fps= 50 q=-0.0 Lsize= 142539KiB time=00:00:10.00 bitrate=116767.7kbits/s dup=0 drop=98 speed=0.998x ------------ du -sh DV* 69M DV50i.mov 138M DV50p.mov 140M DV50p+pcm.mov --------- ffprobe -hide_banner DV50p+pcm.mov Input #0, mov,mp4,m4a,3gp,3g2,mj2, from 'DV50p+pcm.mov': Metadata: major_brand : qt minor_version : 512 compatible_brands: qt encoder : Lavf61.7.100 Duration: 00:00:10.00, start: 0.000000, bitrate: 116767 kb/s Stream #0:0[0x1]: Video: dvvideo (dv5p / 0x70357664), yuv422p(progressive), 720x576 [SAR 16:15 DAR 4:3], 57600 kb/s, 50 fps, 50 tbr, 12800 tbn (default) Metadata: handler_name : VideoHandler vendor_id : FFMP encoder : Lavc61.19.101 dvvideo Stream #0:1[0x2]: Audio: pcm_s16le (sowt / 0x74776F73), 48000 Hz, stereo, s16, 1536 kb/s (default) Metadata: handler_name : SoundHandler vendor_id : [0][0][0][0]

1 0

OpenGL Texture Limit
by Andrea paz 07 Jun '25

07 Jun '25

I have this OpenGL warning with X11-OpenGL driver: void AWindowGUI::update_asset_list(): Warning: /home/paz/Video/hdr/hdr_02.mp4 dimensions 3840x2160 exceed OpenGL texture limit 0 In this case I used a 4k source, but the same happens with any resolution, since the “texture limit” is 0. I found this answer: "If you want to know dynamically the texture size limit of your device (because it's change depending on the device), you have to call this method: int[] maxTextureSize = new int[1]; gl.glGetIntegerv(GL10.GL_MAX_TEXTURE_SIZE, maxTextureSize, 0); And don't forget that for some device (the Nexus One for example), the texture size must be a power of 2 !" And also: "One note: It's a static method, so you can simply call GLES10.glGetIntegerv(...)" See: https://registry.khronos.org/OpenGL-Refpages/es2.0/xhtml/glGet.xml But I have absolutely no idea where and how to use that call code. Suggestions? Note: the limits of my device (on the specification sheet) should be: 16384x16384

3 6

ffv1hd422p.mkv - groundwork for a preservation purpose
by Terje J. Hanssen 06 Jun '25

06 Jun '25

This example capture a raw video stream (optional tape playback) from the camera sensor over HDMI out and transcode audio to PCM into ffv1hd422p.mkv for preservation purpose. -input_format yuyv422 -video_size hd1080 possibly may be commented out, as the encoder default uses the maximum capability it get from the capture card ms2130 and driver v4l2 The video part is used from this source https://avpres.net/FFmpeg/im_FFV1.html At this high bitrate 320Mbps+ here, a combination with segmented recording is recommended and neccesary to control the file sizes (segments) and eventually corrupted parts better. ffmpeg -hide_banner -f v4l2 -framerate 50 -i /dev/video0 -f alsa -ac 2 -i hw:CARD=UHD,DEV=0 -c:v ffv1 -level 3 -threads 8 -coder 1 -context 1 -g 1 -slices 24 -slicecrc 1 -c:a pcm_s16le -t 10 -f matroska ffv1hd422p_preserv.mkv Input #0, video4linux2,v4l2, from '/dev/video0': Duration: N/A, start: 89.753684, bitrate: 1658880 kb/s Stream #0:0: Video: rawvideo (YUY2 / 0x32595559), yuyv422, 1920x1080, 1658880 kb/s, 50 fps, 50 tbr, 1000k tbn [aist#1:0/pcm_s16le @ 0x560bdabb0480] Guessed Channel Layout: stereo Input #1, alsa, from 'hw:CARD=UHD,DEV=0': Duration: N/A, start: 1749208074.680286, bitrate: 1536 kb/s Stream #1:0: Audio: pcm_s16le, 48000 Hz, stereo, s16, 1536 kb/s Stream mapping: Stream #0:0 -> #0:0 (rawvideo (native) -> ffv1 (native)) Stream #1:0 -> #0:1 (pcm_s16le (native) -> pcm_s16le (native)) Press [q] to stop, [?] for help Output #0, matroska, to 'ffv1hd422p_preserv.mkv': Metadata: encoder : Lavf61.7.100 Stream #0:0: Video: ffv1 (FFV1 / 0x31564646), yuv422p(tv, progressive), 1920x1080, q=2-31, 200 kb/s, 50 fps, 1k tbn Metadata: encoder : Lavc61.19.101 ffv1 Stream #0:1: Audio: pcm_s16le ([1][0][0][0] / 0x0001), 48000 Hz, stereo, s16, 1536 kb/s Metadata: encoder : Lavc61.19.101 pcm_s16le [out#0/matroska @ 0x560bdabb0600] video:388733KiB audio:1875KiB subtitle:0KiB other streams:0KiB global headers:0KiB muxing overhead: 0.014887% frame= 500 fps= 50 q=-0.0 Lsize= 390666KiB time=00:00:10.00 bitrate=320033.5kbits/s speed=0.999x

1 0

Apple's EDR technology
by Andrew Randrianasulu 02 Jun '25

02 Jun '25

https://m.youtube.com/watch?v=COxAt8pl_Xw "WWDC21: Explore HDR rendering with EDR | Apple" 34 min runtime What I found interesting that Apple used internal floating-point representation with some values ABOVE 1.0f but those values probably can't leak into final video, at best they might remain inside OpenEXR image format captured at some point "as is" (?) There is textual (but more iOS /mobile oriented) description with pictures: https://medium.com/@maxwellyuchenlong/wwdc22-10113-10114-110565-explore-edr… ===== *Reference Mode* Reference mode is a new display mode for color-intensive workflows that pin settings and block out distractions to provide more objective and reliable reference results for a variety of common videos, such as color grading, editing, and content mode, similar to the reference presets on macOS. When you enable the reference mode, you will have the following features: - The SDK peak brightness is fixed at 100 nits and the HDR peak brightness is fixed at 1000 nits, so there is a 10x EDR headroom. - Disables HDR tone mapping to provide one-to-one media display mapping. - Disables all display dynamic adjustments that occur to adapt to the environment, such as True Tone, Auto Brightness, and Night Shift modes, and instead allows the user to fine-calibrate the white point manually. === end of quotation ==== At least back in 2021 Apple was still using Metal (their Vulkan-like graphics API) and OpenGL for system-wide compositing, not sure if it changed or not lately? Still I am not sure how this provides HDR-like experience on SDR display (may be by using dynamic mapping?) Of course this is just tip of the iceberg I found in few minutes at the morning.

3 3

Record camcorder video stream via v4l2 and hdmi-usb3 capture as input device
by Terje J. Hanssen 29 May '25

29 May '25

I tried to Record camcorder video stream via v4l2 and hdmi-usb3 capture as input device. ms2130 is more powerful but else similar to the usb2 EasyCap device example described in the CinGG manual https://cinelerra-gg.org/download/CinelerraGG_Manual/Use_Case_1_a_EasyCap.h… My Sony FX7-E 1080i50 HDV camcorder was connected via HDMI to a ms2130 usb3 capture card. My procedure description here on openSUSE Tumbleweed-Slowroll, kernel 6.12.30-1.0.4.sr20250501-longterm ~/Applications> ./CinGG-20250430-x86_64.AppImage Cinelerra Infinity - built: Apr 30 2025 07:18:32 1) Settings > Preferences > Record * File format: FFMPEG mp4 Video In: * Record driver: Video4Linux2 /dev(video0 The first thing I noticed was that default Record video setting was SD NTSC, not PAL as else according to my time zone. Could this possibly be changed as else? * Frames to record on disk: 30 (default)? Changed frame to a camcorder resolution and fps supported by v4l2 and ms2130: * Frame: 1920 x 1080 (was default NTSC SD) * fps: 50 (progressive) File > Record r Two windows opened, one showing the running record time etc. and the other the camera window image. My simple question are: 1. Should the video stream be "recorded" and stored somewhere (according to some settings)? 2. Can it be saved as a video file type on disk? I am also interested to see if yuv422p color sampling is retained. A copy from the running terminal: BC_DisplayInfo::gl_fb_config failed DeviceV4L2Base::v4l2_open pixels="YUYV"; res="1920x1080 1600x1200 1360x768 1280x1024 1280x960 1280x720 1024x768 800x600 720x576 720x480 640x480" v4l2 s_fmt 1920x1080 YUYV DeviceV4L2Base::v4l2_open VIDIOC_S_STD: Uegnet «ioctl» for enhet Terje J. Hanssen '

2 9

New forum question at:
by Phyllis Smith 27 May '25

27 May '25

https://www.cinelerra-gg.org/forum/forum-hardware/record-from-minidv-hangs-…

3 3

vaapi speedup patch (for ffmpeg.git)
by Andrew Randrianasulu 27 May '25

27 May '25

At some point it hopefully landed in, well, ffmpeg main repo, but for now I manually downloaded mbox and applied with git am to ffmpeg tree. https://patchwork.ffmpeg.org/project/ffmpeg/patch/20250507120326.156062-3-n… makes vaapi 4k hevc hdr decode + opencl filtering go from 6.0 fps to 6.3 fps. ./ffmpeg -hwaccel vaapi -init_hw_device opencl=ocl -filter_hw_device ocl -i ~/K38_sdcard1/Documents/iPhone11_4K-recorder_59.940HDR10.mov -vf "format=p010,hwupload,tonemap_opencl=tonemap=mobius:param=0.01:desat=0:r=tv:p=bt709:t=bt709:m=bt709:format=nv12,hwdownload,format=nv12" -c:a copy -c:s copy -c:v rawvideo -an -f nut /dev/null

1 0

Hwaccel FFV1 decoding and encoding
by Terje J. Hanssen 26 May '25

26 May '25

According to Phoronix 22 May 2025 : * FFmpeg FFV1 Vulkan Encoder Lands +35% Improvement For AMD, +50% For NVIDIA * No word on Intel performance benefits from this Vulkan encode optimization. AFAIK Intel has not yet QSV support for FFV1 either. Any idea which ffmpeg version release ffv1_vulkan will be available ? And which color bit depths will be supported: 8-bit yuv422p and 10-bit yuv422p10le/ y210 ? Speed tests? Terje J. H

2 2