- Cin - Cinelerra-GG Mailing Lists

Re: [Cin] Issues related to Motion Plugin
by Miroslav Rovis 28 Sep '22

28 Sep '22

Georgy, I suppose it was not intentional, but in case this did not arrive to Cin ML (likely, it's not in the archive), I want to send this email, because I worked a couple of hours on it, and I mentioned it in another email today. I need it to show what issues I had. -- Miroslav Rovis Zagreb, Croatia https://www.CroatiaFidelis.hr my PGP-key: https://www.croatiafidelis.hr/FCF13245ED247DCE443855B7EA9884884FBAF0AE.asc I'm removing one layer of citation, for legibility, as if Georgy sent it, not me. ========================================================== On 220926-10:20+0700, Georgy Salnikov wrote: On Sun, 25 Sep 2022, Miroslav Rovis wrote: > > Phyllis has such a very shaky example. May be, I have yet a cingg project to > > stabilize it using Motion plugin keyframes. Should I search for it? > If you think so. Miroslav, in attachment you find the archive. Unpack it, there are two files: the source video clip, and the cingg project file. This is a rather difficult example not only because of strong shaking. Moreover, the video is taken perhaps from a moving car, so the landscape itself in the viewport is changing. The example project demonstrates the possibility of the Motion plugin to switch from one key part of image to another as the former part goes off the viewport, becomes obscured or changes shape. This is not smoothing video, but freezing it. The result is not completely still, but additionally applying vidstab to it perhaps should improve the result once more. To test, do the following after unpacking: Open shaker2.xml (replace current project) Make sure: Motion and Time Average must be off Rewind to the beginning Uncheck 'Play track' Open Motion controls dialog Action: Do Nothing Calculation: Save coords to tracking file Scroll timeline forward, set cursor after the next keyframe Action: Do Nothing Calculation: Save coords to tracking file Repeat that for all the rest keyframes (total of 4) Close Motion controls dialog Rewind to the beginning Switch Motion on Check 'Play track' Play forward up to the end Video will be still very shaky. The file 'motion-shaker4' will appear in /tmp Uncheck 'Play track' Rewind to the beginning Open Motion controls dialog Action: Stabilize Subpixel Calculation: Load coords from tracking file Scroll timeline forward, set cursor after the next keyframe Action: Stabilize Subpixel Calculation: Load coords from tracking file Repeat that for all the rest keyframes (total of 4) Close Motion controls dialog Rewind to the beginning Switch Motion on Switch Time Average on Check 'Play track' Render project Now compare the result of rendering with the original one. ....... Some of my examples to apply ffmpeg's vid.stab is as follows: rm transforms.trf ffmpeg -i video-src.mp4 -c:v rawvideo -filter:v vidstabdetect -an -f null -y /dev/null ffmpeg -i video-src.mp4 -c:v libx264 -filter:v vidstabtransform,unsharp=5:5:0.8:3:3:0.4 -b:v 12000k -b:a 192k -ac 2 -ar 48000 -y video-dst.m4v Regards Georgy _______________________________________________________________________________ Georgy Salnikov NMR Group Novosibirsk Institute of Organic Chemistry Lavrentjeva, 9, 630090 Novosibirsk, Russia Phone +7-383-3307864 Email sge(a)nmr.nioch.nsc.ru _______________________________________________________________________________

4 5

Issues related to Motion Plugin
by Miroslav Rovis 27 Sep '22

27 Sep '22

Reusing what I wrote in late last night's thread. On 220924-23:31+0200, Miroslav Rovis wrote: > I've been using VP9 (previously VP8) webm a lot. E.g. anywhere > you go from: > https://www.croatiafidelis.hr/2022/ > or, to take just one sample (4 minutes): > https://www.croatiafidelis.hr/2022/Jazovka/VID-2022-06-22-1034-Jazovka-proc… > It's nothing but motion plugin set up, blur plugin in the > bottom video at 100%, and go render. Then work the audio > separately, mux and post. Of course, the stabilizing is slow, > motion plugin takes long to calculate. > > But all those people moving can be scrutinized in a lot of > detail, as if I filmed with a mobile phone attached to a good > stabilizer, and instead, the original was a shaky video as > ever, because I held the mobile with just my hand. > > Cinelerra can stabilize you videos well. It's worth the > time-expensive learning curve to get to know how to do it. > > Here are some comparison videos with both the shaky and > stabilized video (however I've gotten better since then, these > are all old, but I have no other comparisons) > > e.g. from: > https://www.croatiafidelis.hr/2019/ > the videos e.g.: > https://www.croatiafidelis.hr/2019/Jaska/190824-Jaska-Vukovic-demo-1-stabil… > or from: > https://www.croatiafidelis.hr/2015/Macelj/Macelj-2-cmp.php > the video being: > https://www.croatiafidelis.hr/2015/Macelj/150607-Macelj-HebrangCOMP.webm But the above do not sufficienlty well demonstrate how well Cinelerra can stabilize shaky videos. I think I have a better example that I hope I can offer. But I plan to make it both a demonstration of this capability and a tutorial so other users can benefit from it, on the one hand. And on the other, I want to show the issues that come across in Motion Plugin usage. I can say what it is right now, but that's dry and no way immediately plausible. Here: The fist issue: When using Motion Plugin, a last-used setting (previusly used, even in a different video) is written in the new XML along with the newly setup numbers for a particular part of a video to which Motion Plugin Effect is attached. And it sometimes messes up the XML, and rendering does not get what the user put in. The second issue: And, sometimes, there appear to be problems with some Cinelerra cache somewhere so bad that tweaking the Motion Plugin settings, while written correctly in the new XML, do not actually apply. The rendering gets the user the earlier tweaked result no matter the new editing. While I am not certain I can, within the time that I can set appart for this, show how to reproduce the second issue, I believe I can show the first issue within, hopefully this day today. I want to do it along with demonstrating a very good motion stabilization on not-too difficult parts of a 1m 30s video that otherwise would be almost bad for viewing. I plan to post it, God permittign, at: https://croatiafidelis.hr/foss/cap/cap-220925-cinelerra/ where, at the time of writing this, there is only the short video: VID_20220910_133415_5.mp4 [*] that I want to use for this demonstration/tutorial/bug-report all-in-one. Regards! ---- [*] The timestamp in VID_20220910_133415_5.mp4 is for 2022-09-10 13:34:15, the start of my filming, but that's part 5, which I obtained with cutting the VID_20220910_133415.mp4 with mkvmerge. It's and event that the media exanded for a few days on, an, thankfully innocuous standoff with Croatian police... I want to show how good reporting can be done with simple cheap mobile phone camera, made good only after motion stabilization. I mean, look at that video, it's so shaky because I walked and even talked and waved (well, that's not seen) at people... Stabilization is absolutely necessary. This is what you should get (SHA256 of the video): 8ae65c8b17b3d607020c392323fa5ab8569bd40083648d808602a0dfffc8f1f9 VID_20220910_133415_5.mp4 -- Miroslav Rovis Zagreb, Croatia https://www.CroatiaFidelis.hr my PGP-key: https://www.croatiafidelis.hr/FCF13245ED247DCE443855B7EA9884884FBAF0AE.asc

5 11

Mesa3d va-api encoding MR
by Andrew Randrianasulu 27 Sep '22

27 Sep '22

So, few hours ago this merge request for mesa3d surfaced (I think it affects radeonsi AMD cards) --- why: to enhance va video encoding quality how: use va encoding quality_level interface, and provide default value and encoding quality adjustment options, so that users can finetune encoding quality and performance from va quality interface. There are 3 settings added: -- preset modes: speed, balance, quality they are using different encoding strategies -- vbaq modes: vbaq mode is using variance based strategy to improve the subjective image quality -- pre-encoding modes: Using scaled down input image for pre-encoding to have better rate-control reaction and consume more memory in the same time. Only preencoding-4x mode is enabled. --- https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17994

1 1

Suggestion: improve setting for VP9 webm
by Miroslav Rovis 27 Sep '22

27 Sep '22

It's simple, and my wording may not be quite precise, but you guys and gals will understand. So, when using the setting in the render window, I figured out that there are better settings to use than the presets that are offered. And they are simpler too. In essence, the Wiki page for VP9 can tell you all: https://trac.ffmpeg.org/wiki/Encode/VP9 [*] I have no rendering, actually no Cinelerra opened right now, so, from memory: It's just b=0, then add crf=31 (or other number), and then row-mt=1, and remove the min and max settings from the preset. I did use b=0 (which would correspond to -b 0 in the VP9 wiki page (but I see only -b:v 0 in the page). And it did work I can tell you. However, see there around: "more contemporary CRF-style approach for the final pass to achieve a certain perceptual quality" and there is -b:v 0 in the second pass. and around: "must use a combination of -crf and -b:v 0. Note that -b:v MUST be 0." Also find: "FFmpeg added support for row based multithreading in version 3.4, released on January 25th, 2018. As of libvpx version 1.7.0 this multithreading enhancement is not enabled by default and needs be manually activated with the -row-mt 1 switch." which I find works great as well on my systems. They are relatively new: AMD Ryzen 7 PRO 4750G and also one: AMD Ryzen 7 5700G so if the above maybe would not work well on older machines, then, don't know... So, this is a suggestion to improve the VP9 webm presets in Cinelerra FFmpeg. While I am getting better at understanding the source code, I would still probably be to slow to find it and change it and send a patch myself. Too slow. And that would take me more time than I can set apart at this time. Regards! --- I've been using VP9 (previously VP8) webm a lot. E.g. anywhere you go from: https://www.croatiafidelis.hr/2022/ or, to take just one sample (4 minutes): https://www.croatiafidelis.hr/2022/Jazovka/VID-2022-06-22-1034-Jazovka-proc… It's nothing but motion plugin set up, blur plugin in the bottom video at 100%, and go render. Then work the audio separately, mux and post. Of course, the stabilizing is slow, motion plugin takes long to calculate. But all those people moving can be scrutinized in a lot of detail, as if I filmed with a mobile phone attached to a good stabilizer, and instead, the original was a shaky video as ever, because I held the mobile with just my hand. Cinelerra can stabilize you videos well. It's worth the time-expensive learning curve to get to know how to do it. Here are some comparison videos with both the shaky and stabilized video (however I've gotten better since then, these are all old, but I have no other comparisons) e.g. from: https://www.croatiafidelis.hr/2019/ the videos e.g.: https://www.croatiafidelis.hr/2019/Jaska/190824-Jaska-Vukovic-demo-1-stabil… or from: https://www.croatiafidelis.hr/2015/Macelj/Macelj-2-cmp.php the video being: https://www.croatiafidelis.hr/2015/Macelj/150607-Macelj-HebrangCOMP.webm -- Miroslav Rovis Zagreb, Croatia https://www.CroatiaFidelis.hr my PGP-key: https://www.croatiafidelis.hr/FCF13245ED247DCE443855B7EA9884884FBAF0AE.asc

4 10

CinGG's Youtube channel
by Andrea paz 25 Sep '22

25 Sep '22

Just a quick note to report that the CinGG Youtube channel has reached 401 subscribers. It must be said, however, that lately the enrollments have slowed down a lot compared to the early days. In recent years I have convinced only one person to switch to CinGG; moreover, lately the arrival of DaVinci Resolve on linux constitutes a formidable competitor even if not open source. ;-)

3 10

Compile Cin from source, but make errors out
by Miroslav Rovis 24 Sep '22

24 Sep '22

Hi! I'm compiling Cinelerra from git source (master, up to date, pulled this morning). Here's the script that I use (attached): cin-build.sh I just completed the first attempt, which failed. I guess the make says the most (but I have other logs, as per the script, if needed). So this is the make (that xz'es into just 232K, but is 7922K unxz'ed, attached): cinelerra_make_220914_1319.xz I don't know if this is only here or is some generic that will happen to other possible builders. I think (not sure) it may be something to do with tiff... So I'll try ./configure ... --disable-tiff . Have no other ideas at the moment. Any help welcome. (And if I figure out myself why the errors, I'll report.) Regards! -- Miroslav Rovis Zagreb, Croatia https://www.CroatiaFidelis.hr my PGP-key: https://www.croatiafidelis.hr/FCF13245ED247DCE443855B7EA9884884FBAF0AE.asc

3 14

flac 1.4.0
by Andrew Randrianasulu 24 Sep '22

24 Sep '22

https://www.phoronix.com/news/FLAC-1.4-Released there ARE breaking changes in API so it may not work with our code yet ... One thing NOT mentioned in Phoronix article but spelled out on github is removal of some assembly optimizations, because they "provide only small benefits" and make fuzzing harder... so speed might be slower, not sure how much

4 69

"Stable diffusion" AI
by Andrew Randrianasulu 22 Sep '22

22 Sep '22

for some not completely outdated event (just two weeks ago) ... https://m.youtube.com/watch?v=nVhmFski3vg

1 0

libtiff 4.4.0
by Andrew Randrianasulu 21 Sep '22

21 Sep '22

http://www.simplesystems.org/libtiff/releases/v4.4.0.html -----Major changes None Software configuration changes <http://www.simplesystems.org/libtiff/releases/v4.4.0.html#software-configur…> - Handle absolute paths in pkg-config file (issue #333 <https://gitlab.com/libtiff/libtiff/-/issues/333>) - Correct fix for the pkgconf file relative paths. - cmake: allow running the tests with a read-only source directory. - cmake: Fix STRIPCHOP_DEFAULT value in CMake builds. - build: Fix static library imports in mingw related to LERC - Fix version in libtiff-4.pc.in, and CMake build: Add requirements to pc file - cmake: Fix build with CMake 3.10. - cmake: Export tiff targets. - Make LERC_SUPPORT conditional on ZLIB_SUPPORT ---- not sure if last item will fix reported compilation problem, but we always can update both library and its configure line (or even patch them)

2 1

places with CinGG
by Andrew Randrianasulu 21 Sep '22

21 Sep '22

Skackbuilds.org obvious for Slackware and Slackware-based distros, like SalixOS [0] https://github.com/SlackBuildsOrg/slackbuilds/tree/master/multimedia/cinele… contain a bit older version of CinGG gitlab mirror, empty (?) https://gitlab.com/cinelerra-gg/cinelerra-gg ubuntubuzz micro reviews https://www.ubuntubuzz.com/2022/06/video-editors-on-ubuntu-floss-alternativ… https://www.ubuntubuzz.com/2022/01/install-cinelerra-and-natron-professiona… [0] - https://forum.salixos.org/viewtopic.php?f=17&t=8409&sid=94d2477c8afcb9b0e15…

3 6