Cinelerra-GG MantisBT

Audio tracks align with no video track, and in working with Mixers · Then possibly, rendering the project will have no problem - will render 1:1 scale.

PhyllisSmith wrote:
I need to test more options myself before asking others to do so -- sorry
int Track::is_armed()
{
//	return armed && gang_master()->armed;
    return gang_master()->armed;
}

int Track::is_ganged()
{
//	return ganged && gang_master()->ganged;
    return gang_master()->ganged;
}

int Track::armed_gang(Track *track)
{
//	if( !track->ganged ) return 0;
    if( edl->local_session->gang_tracks == GANG_NONE ) return ganged;
    Track *current = gang_master();
//	if( !current->ganged ) return 0;
    for(;;) {
        if( track == current ) return 1;
        current = current->next;
        if( !current || current->master ) return 0;
    }
    return 1;
}

int Track::plays()
{
    return play && gang_master()->play;
}

I don't use the forum because I tried twice to create an account there and I never received an e-mail of confirmation.
...shortcut. But they doesn't work for the keyframes of curves, mask, mode and pan. It's only for the keyframes of the plugins, right?
But she always says: "You will be allowed to touch me after two years at our wedding."
bad file format: /media/D2/my_vid/batchtest1.mp4

View Issue Details

ID:	Category:	Severity:	Reproducibility:	Date Submitted:	Last Update:
654	[Cinelerra-GG] Bug	minor	have not tried	2024-02-28 17:40	2024-02-28 18:21

Reporter:	PhyllisSmith	Platform:
Assigned To:		OS:
Priority:	normal	OS Version:
Status:	new	Product Version:
Product Build:		Resolution:	open
Projection:	none
ETA:	none	Fixed in Version:
		Target Version:

Summary:	AgingTV plugin allows for picking type of aging, but slider amount seems to be ignored.
Description:	If everything is unchecked and then one of the 3 items of Scratch, Pits, or Dust is checked BUT set to zero, you can still see that item in the compositor. So it looks like the slider number value is ignored. However, it looks like my interpretation of the slider is wrong -- instead of the number of items display, maybe it is the interval between when the item appears but it does not really look like it; in which case the documentation is wrong. Grain is either on or off and works as expected. The good news is that at least you can designate which type of aging you want, be it grain, scratch, pits, dust, or all or none.
Tags:
Steps To Reproduce:	1) load video 2) drag AgingTV plugin from the Resources window to the video track 3) check only 1 of Scratch, Pits, or Dust and try varying the slider value 4) note even when the slider value is 0, there is still that type showing in the Compositor when playing
Additional Information:
Attached Files:

View Issue Details

ID:	Category:	Severity:	Reproducibility:	Date Submitted:	Last Update:
653	[Cinelerra-GG] Bug	minor	always	2024-02-04 22:18	2024-02-18 19:42

Reporter:	PhyllisSmith	Platform:
Assigned To:		OS:
Priority:	normal	OS Version:
Status:	new	Product Version:
Product Build:		Resolution:	open
Projection:	none
ETA:	none	Fixed in Version:
		Target Version:

Summary:	At least 1 LV2 plugin works until click on UI button
Description:	This is for doing you own builds, not AppImage. LV2 plugin works fine as long as you do not click on the UI button (not sure if this particular L2 plugin has a graphical interface or not but it seems to). I think this is from the Guitarix GX plugins set. L2_GxFlanger has this problem but L2_Dj flanger does not. There is a workaround so you can continue using it -- click on the "turn off" button and take down the "show controls" window. Then you can just turn it back on and do NOT hit the UI button any more! It seems like this is a fixable bug. Error message is: void PluginLV2Client::process_buffer(int); timeout: L2_GxFlanger which comes from cinelerra/pluginv2client.C
Tags:
Steps To Reproduce:	1) in Settings->Interface, you have to provide an LV2 Path where these L2 plugins reside in your system 2) load video with audio 3) drag over L2_GxFlanger and play -- all sounds good 4) click on the plugin "Show Controls" button and a window will popup 5) click on the UI button and now it does try to play but jiggles and the error message pops up
Additional Information:
Attached Files:

View Issue Details

ID:	Category:	Severity:	Reproducibility:	Date Submitted:	Last Update:
652	[Cinelerra-GG] Bug	minor	always	2024-02-02 21:38	2024-02-08 19:48

Reporter:	PhyllisSmith	Platform:
Assigned To:	PhyllisSmith	OS:
Priority:	normal	OS Version:
Status:	confirmed	Product Version:
Product Build:		Resolution:	open
Projection:	none
ETA:	none	Fixed in Version:
		Target Version:

Summary:	If using AppImage, using Calf Plugins give error
Description:	When running CinGG from an AppImage, the graphics ard on the computer that created the AppImage is very likely different from what the user has therefore the UI portion of Calf Plugins can not work. (A different LV2 plugin, drowaudio-flanger.lv2, does seem to work and it has a graphical UI.) It does not crash but the playhead just jiggles back and forth and there is the error message in a popup window something like: void PluginLV2Client::process_buffer(int); timeout: L2_Calf Reverb Other LV2 plugins that specifically do NOT have a potential UI interface work OK. Error message comes from the following code in pluginlv2client.C void PluginLV2Client::process_buffer(int size) { PluginLV2ParentUI ui = get_ui(); if( !ui ) return; shm_bfr->done = 0; ui->send_child(LV2_SHMID, &shmid, sizeof(shmid)); // timeout 10duration, min 2sec, max 10sec double sample_rate = get_project_samplerate(); double t = !sample_rate ? 2. : 10. * size / sample_rate; bclamp(t, 2., 10.); ui->output_bfr->timed_lock(t*1e6, "PluginLV2Client::process_buffer"); if( !shm_bfr->done ) eprintf("timeout: %s",server->title); According to what I found on the internet, you can determine if you are running from an AppImage via: "If the app is an AppImage, process.env.APPIMAGE contains the path to the .AppImage file, otherwise it is undefined." And then you need to find a programmatic way to determine if running from an AppImage, and somehow skip running code to bring up a UI so that you will be at least able to use the Calf plugin without the UI interface.
Tags:
Steps To Reproduce:	1) have the Calf LV2 plugins installed on your system 2) execute the CinGG AppImage that was NOT created on your computer 3) drag a Calf plugin to the audio track on the timeline 4) attempt to play the track You will get an error message and the playhead will just jiggle back and forth.
Additional Information:
Attached Files:

View Issue Details

ID:	Category:	Severity:	Reproducibility:	Date Submitted:	Last Update:
651	[Cinelerra-GG] Bug	minor	always	2024-01-06 18:21	2024-01-06 18:28

Reporter:	PhyllisSmith	Platform:
Assigned To:		OS:
Priority:	normal	OS Version:
Status:	new	Product Version:
Product Build:		Resolution:	open
Projection:	none
ETA:	none	Fixed in Version:
		Target Version:

Summary:	Getting error messages using SVT render with svt_av1.webm or svt_av1.mkv
Description:	The output to the terminal window (even though the render seems to succeed) when using svt_av1.webm or svt_av1.mkv is: Svt[info]: ------------------------------------------- Svt[info]: SVT [version]: SVT-AV1 Encoder Lib v1.8.0 Svt[info]: SVT [build] : GCC 10.3.1 20210422 (Red Hat 10.3.1-1) 64 bit Svt[info]: LIB Build date: Jan 5 2024 13:17:05 Svt[info]: ------------------------------------------- Svt[info]: Number of logical cores available: 16 Svt[info]: Number of PPCS 140 Svt[info]: [asm level on system : up to avx2] Svt[info]: [asm level selected : up to avx2] Svt[info]: ------------------------------------------- Svt[info]: SVT [config]: main profile tier (auto) level (auto) Svt[info]: SVT [config]: width / height / fps numerator / fps denominator : 352 / 240 / 30000 / 1001 Svt[info]: SVT [config]: bit-depth / color format : 8 / YUV420 Svt[info]: SVT [config]: preset / tune / pred struct : 10 / PSNR / random access Svt[info]: SVT [config]: gop size / mini-gop size / key-frame type : 161 / 32 / key frame Svt[info]: SVT [config]: BRC mode / rate factor : CRF / 35 Svt[info]: ------------------------------------------- Svt[warn]: Failed to set thread priority: Invalid argument FFStream::encode_frame: encode failed. file: /tmp/aspectno.webm err: Resource temporarily unavailable FFMPEG::mux_video err: Operation not permitted FFStream::encode_frame: encode failed. file: /tmp/aspectno.webm err: Resource temporarily unavailable FFMPEG::mux_video err: Operation not permitted Render::render_single: Session finished. ** rendered 2713 frames in 13.612 secs, 199.309 fps As you can see from the above, there are these 3 concerning error messages: FFStream::encode_frame: encode failed. FFMPEG::mux_video err: Operation not permitted err: Resource temporarily unavailable
Tags:
Steps To Reproduce:	1. load video, no audio needed 2. select File pulldown option Render 3. set an output file name 4. click on the video wrench icon 5. select ffmpeg and then either webm or mkv 6. use the down arrow to select svt_av1 in either case 7. look at the output on the CinGG startup window for error messages
Additional Information:
Attached Files:

View Issue Details

ID:	Category:	Severity:	Reproducibility:	Date Submitted:	Last Update:
649	[Cinelerra-GG] Bug	minor	always	2023-12-21 18:41	2024-01-06 16:52

Reporter:	PhyllisSmith	Platform:
Assigned To:	PhyllisSmith	OS:
Priority:	normal	OS Version:
Status:	confirmed	Product Version:
Product Build:		Resolution:	open
Projection:	none
ETA:	none	Fixed in Version:
		Target Version:

Summary:	lv2 modified configure in the build does not work in Fedora 39
Description:	Because it seems that Python 3.12 will not build the lv2 plugins, Andrew used the following successfully on termux and slackware: ./configure --with-lv2 --disable-lv2 --disable-sord --disable-serd --disable-suil --disable-sratom --disable-lilv However, this does not work as a workaound on Fedora 39, which uses Python 3.12. Errors of can't find headers: checking for lv2 availability... checking for lilv/lilv.h... no checking lilv headers... no checking for serd/serd.h... no checking serd headers... no checking for sord/sord.h... no checking sord headers... no checking for sratom/sratom.h... no checking sratom headers... no checking for suil/suil.h... no checking suil headers... no even though lilv-devel is installed as shown below: Fedora 39 *** root@fedora:/tmp# grep lilv a lilv-libs.x86_64 0.24.20-1.fc39 lilv.x86_64 0.24.20-1.fc39 lilv-devel.i686 0.24.20-1.fc39 lilv-devel.x86_64 0.24.20-1.fc39 lilv-libs.i686 0.24.20-1.fc39 python3-lilv.x86_64 0.24.20-1.fc39
Tags:
Steps To Reproduce:
Additional Information:
Attached Files:

Notes
(0005597) Andrew-R 2023-12-01 18:01	#!/bin/sh vncserver -kill :1 pulseaudio --exit-idle-time=-1 -D& rm $PREFIX/usr/tmp/.X1-lock vncserver :1 -geometry 1280x800 -shmem export DISPLAY=":1" export LANG="en_US.UTF-8" export PULSE_SERVER="127.0.0.1" #openbox& xfwm4& gdb bin/cin ==== actually this. (for debugging, for normal run just leave bin/cin at last line)

(0005596) Andrew-R 2023-12-01 17:59	cin.sh #!/bin/sh vncserver -kill :1 pulseaudio --exit-idle-time=-1 -D& rm $PREFIX/usr/tmp/.X1-lock vncserver :1 -geometry 1280x800 -shmem export DISPLAY=":1" export LANG="en_US.UTF-8" export PULSE_SERVER="127.0.0.1" #openbox& xfwm4&

Notes
(0005595) Andrea_Paz 2023-11-26 21:09	@Andrew-R With a new build now everything works fine and I get the same results as you.

(0005593) Andrea_Paz 2023-11-20 21:21 (Last edited: 2023-11-20 21:22)	I think so; the ./configure command shows no errors. Can you tell me how to do further verification?

(0005592) Andrew-R 2023-11-20 21:12	@Andrea_Paz, can you double-check that openexr actually enabled in your build? For me it works even if I rebuild cingg with new internal ffmpeg 6.1.

(0005591) Andrea_Paz 2023-11-20 21:09	My Cinelerra.rc: ... FILE_PROBE_TOTAL 20 FILE_PROBE0_NAME EXR FILE_PROBE0_ARMED 1 FILE_PROBE1_NAME FFMPEG_Early FILE_PROBE1_ARMED 0 .... It appears to be enabled, but ffmpeg is still active. Could it be the new ffmpeg6.1 giving problems?

(0005590) Andrew-R 2023-11-20 20:28	strange for me in .bcast5/Cinelerra_rc I have ``` FILE_PROBE_TOTAL 20 FILE_PROBE0_NAME FFMPEG_Early FILE_PROBE0_ARMED 0 FILE_PROBE1_NAME Scene FILE_PROBE1_ARMED 1 FILE_PROBE2_NAME DB FILE_PROBE2_ARMED 1 FILE_PROBE3_NAME DV FILE_PROBE3_ARMED 1 FILE_PROBE4_NAME SndFile FILE_PROBE4_ARMED 1 FILE_PROBE5_NAME PNG FILE_PROBE5_ARMED 1 FILE_PROBE6_NAME PPM FILE_PROBE6_ARMED 1 FILE_PROBE7_NAME JPEG FILE_PROBE7_ARMED 1 FILE_PROBE8_NAME GIF FILE_PROBE8_ARMED 1 FILE_PROBE9_NAME EXR FILE_PROBE9_ARMED 1 FILE_PROBE10_NAME FLAC FILE_PROBE10_ARMED 1 FILE_PROBE11_NAME CR2 FILE_PROBE11_ARMED 0 FILE_PROBE12_NAME TGA FILE_PROBE12_ARMED 1 FILE_PROBE13_NAME TIFF FILE_PROBE13_ARMED 1 FILE_PROBE14_NAME OGG FILE_PROBE14_ARMED 1 FILE_PROBE15_NAME Vorbis FILE_PROBE15_ARMED 1 FILE_PROBE16_NAME MPEG FILE_PROBE16_ARMED 1 FILE_PROBE17_NAME EDL FILE_PROBE17_ARMED 1 FILE_PROBE18_NAME FFMPEG_Late FILE_PROBE18_ARMED 1 ``` and it works for me .... try with temporary new profile?

(0005589) Andrea_Paz 2023-11-20 20:08 (Last edited: 2023-11-20 20:12)	It does not maintain the setting. Activating enable and then "Apply," etc. The internal engine still does not work and if you reopen "probe order" you can see that it is disabled again.... Edit Sorry, I forgot to mention that I brought exr up as the first item.

(0005588) Andrew-R 2023-11-20 19:38	@Andrea_Paz be sure that EXR entry is Enabled:

(0005587) Andrea_Paz 2023-11-20 19:26	Yes, I also noticed that "try ffmpeg first" seems to be active. Actually in all the tests I always left the "try ffmpeg last" on. I don't know why it doesn't work...

(0005586) Andrew-R 2023-11-20 19:21	@Andrea_Paz note that default FFmpge exr decoder apparently can't put out big values to cingg's pipeline like native decoder does. So, try to disable ffmpeg first button before testing, so lines like FFMPEG::open_decoder: some stream have bad times: /home/paz/test/9.exr disappear ...

(0005585) Andrea_Paz 2023-11-20 19:15	@Andrew-R More tests. I only used the 2 new patches, omitting the "noclamp..." and also "alt_shortcut..." that I was always putting in lately. I made two customizations: "--disable-doc" to keep the build from crashing is adding "AC_CONFIG_MACRO_DIRS([m4])" in configure.ac otherwise autogen.sh doesn't work. The second one has been years in the making. The results are the usual: all exr images have the clip except the usual one, which works fine instead. The images that have the clip give this message to the terminal: FFMPEG::open_decoder: some stream have bad times: /home/paz/test/9.exr FFMPEG::open_decoder: some stream have bad times: /home/paz/test/Ocean.exr FFMPEG::open_decoder: some stream have bad times: /home/paz/test/blender_test2.exr FFMPEG::open_decoder: some stream have bad times: /home/paz/test/BrightRingsNanInf.exr FFMPEG::open_decoder: some stream have bad times: /home/paz/test/blender_test.exr FFMPEG::open_decoder: some stream have bad times: /home/paz/test/Desk.exr As you can see "small_empty_room.exr" gives no errors and in fact works without any problems. This time the priority of tracks from top to bottom is respected and it works normally.

(0005584) Andrew-R 2023-11-20 17:41	@Andrea_Paz strange, it seems to work for me ...

(0005583) Andrew-R 2023-11-20 17:24	@Andrea_Paz well, there is possibility I messed things up! Be sure you do not have previous patches applied. I'll try to rebuild fully too.

(0005582) Andrea_Paz 2023-11-20 13:01	Every time I do a build I get different behaviors. Now I am back to having clipped values on every exr except "small_empty_room.exr" which always works for me. The clipped value read with color picker is 1.0 for some exr and 0.9990 for other images. Gradient still does not work. What is very strange is that the upper track is not the one with priority, but the lower track is! Appimage on the other hand works normally, with the upper track in order. I honestly don't understand all these behaviors and if I'm doing something wrong...

(0005581) Andrew-R 2023-11-20 10:52	@Andrea_Paz new patches replace earlier ones, you can start with just first (0001-TEST-Simpler-unlimited-float-overlay.patch ) and then see if you have any use for second (saving negative-values exrs ?)

(0005580) Andrea_Paz 2023-11-20 10:45	Could you tell me which patches to apply to test? Opengl3 no; do these two new patches replace others or are they just to be added?

(0005579) Andrew-R 2023-11-20 09:25	And this is second patch allowing cingg behave more like cin-cv with negative numbers as exemplified by AllHalfValues.exr test file

(0005578) Andrew-R 2023-11-19 23:36	Here is simpler non limited float overlayer patch - with it gradient still works, and I can boost brightness above 1.0f and bring it back (with details) with fader OR load big values EXR

(0005577) Andrea_Paz 2023-11-19 20:50	Ah, that's where I was wrong! I was using X11-OpenGL. Now with X11, Gradient does not work and fade/eydropper give results above 1.0 (except for BrightRings.exr; Desk.exr and Ocean.exr)

(0005576) Andrew-R 2023-11-19 19:21	@Andrea_Paz be sure to set output device to x11 and not x11-opengl. OpenGL still clips. Also, gradient with patches presented here also works for me in rgba-8bit, rgb-8bit, yuv modes and rgb-float but NOT in rgbA-float. It just makes black output in compositor.

(0005575) Andrea_Paz 2023-11-19 19:16	If the fade works for you then the color picker also reports values above 1.0. To me they are fixed at 1.0 except for the exr with the window on the outside... The same happens with the exr made with Blender that you provided. I wonder what I may have done; or was it an Arch update. The gradient works for me with RGB8; YUVA8 and RGB-Float, both on track with exr and mov and in an empty track.

(0005574) Andrew-R 2023-11-19 15:53	@Andrea_Paz at least BrightRings seems to fade correctly for me - with rgba-8 bit white square remains on top of rings, with rgba-float it fades into black background leaving only rings ... But gradient still does not work for me :(

(0005573) Andrew-R 2023-11-19 15:21	@Andrea_Paz thanks for testing while I was asleep! May be I should rebuild fully before concluding. Does gradient look the same in all project formats - like yuv8-bit or RGB-8bit in additional to RGB(A)-float ? I just put plugin over empty video track for testing .... will re-test openexr test images to see if they work for me. opengl3 patch is sadly incomplete (I think). It will work for single stage processing but eventually if you add more effects or tracks it will render to 8 bpc buffer and thus clip ... (not tested yet)

(0005572) Andrea_Paz 2023-11-19 13:11	I did some tests. Gradient plugin works without any problems. I did not put the patch "opengl3...." and the fade does not work for me while Histogram does. In Histogram the black remains clipped, but in my opinion that is good. Only "small_empty_room_1_4k.exr" works for me while the others I downloaded from "openexr.com/test_image" all turn out clipped at 1.0. I know they are in 16-bit float and not 32-bit; could this be the cause of the clipping? In Gimp and Krita they are not clipped. Should I apply the patch opengl3... as well?

(0005571) Andrew-R 2023-11-19 06:19	oh, no change in cinelerra/overlay_direct_rgba_float.C break gradient plugin :(

(0005570) Andrew-R 2023-11-19 03:27	ah, I think I found where it clips colors to 0.

(0005569) Andrew-R 2023-11-18 23:09	new patch for overlayer, clamp to FLT_MAX (not found yet where min clamping to 0 happen)

(0005568) Andrew-R 2023-11-18 21:20	according to stackoverflow https://stackoverflow.com/questions/4786663/limits-for-floating-point-types define FLT_MAX 3.402823466E+38F and minimum negative number is https://stackoverflow.com/questions/2528039/why-is-flt-min-equal-to-zero "If you want the minimum floating point number (including negative numbers), use -FLT_MAX"

(0005567) Andrew-R 2023-11-18 21:05	Note that according to EXR test page values in color channels in floating-point EXR can go much above 1000: https://openexr.com/en/latest/_test_images/index.html

(0005566) Andrew-R 2023-11-18 21:03	Andrea tried removing clamp in histogram plugin, and I tried to remove limit in overlayer code.

Notes
(0005503) Andrea_Paz 2023-06-13 21:42	https://dyne.org/software/frei0r/ Section: Acknowledgments [...] Frei0r is the result of a collective effort in coordination with several software developers meeting at Piksel between 2003 and 2005 to find a common standard for video effect plugins to be used among their applications: Andraz Tori (Cinelerra/CVS), ... Perhaps looking at the code of CinCVS you can find the implementation of frei0r?

(0005500) Andrew-R 2023-06-12 18:41	Another non-native idea is to use GIMP + G'MIC filters like described here on one sample image from video http://gimpchat.com/viewtopic.php?f=28&t=11231 (HSV modes included, tested on gmic 2.2.3 + gimp 2.10.32) for making CLUT file, then apply such file as usual :)

(0005499) Andrew-R 2023-06-12 18:02	on RGB curves in general apparently you can use ffmpeg's levels filter with Photoshop' ACV file ... converted from GIMP's curve preset! https://github.com/fehlfarbe/gimp2acv https://video.stackexchange.com/questions/16352/converting-gimp-curves-files-to-photoshop-acv-for-ffmpeg

(0005498) Andrew-R 2023-06-12 17:35	in theory frei0r set of effects provides irs own (8 bpc) version of curves apparently working with hue/saturation values too: https://github.com/balint256/frei0r/blob/master/src/filter/curves/curves.c right now this plugin is disabled, but you can try to enable it if you added frei0r plugins to your ffmpeg compilation, as I tried in 2020: "I tried to recompile CinGG's version of ffmpeg with those plugins enabled: FFMPEG_EXTRA_CFG="--enable-frei0r" (I have frei0r plugins 1.7.0 installed) and after tweaking /usr/share/cin/ffmpeg/plugin.opts (added frei0r filter_name=vignette at very end of it) I can see this plugin and even use some effects. "

(0005461) Andrew-R 2023-02-18 02:17	@Andrea_Paz, @PhyllisSmith - thanks for testing! It seems cinelerra really re-reads whole filter chain on actual rendering even if background render files present for given range. I tried Oil Painting plugin with bigger parameter - so it was clearly much slower than decoding/encoding. I'll try to put my addition inside name-generated function with ext. added as parameter, but I agree this is deviation from this bugreport's goal. Attempt at cheating away from some more serious work not worked, huh :-)

(0005460) PhyllisSmith 2023-02-17 19:29 (Last edited: 2023-02-17 19:30)	@Andrew-R ALMOST, but not quite. Tested Render Farm by itself with Background Rendering turned off - still works great. Background Rendering by itself with the extension added also works great. BUT the problem is that if you turn on both the Render Farm AND Background Rendering, some files are created with the extension and some are not and then when playing, you get many errors of missing brenderxxx.extension files (because the Render Farm created them without the extension; or somebody did). I guess the reason for enabling BOTH background rendering and render farm is for a speedup. I did not even know you could do this, but it has been working as documented. "Frames per background rendering job This only works if a Render Farm is being used; otherwise, background rendering creates a single job for the entire timeline." Although, not impacted, still tested rendering to a JPEG sequence list because wasn't sure about it and it works as expected.

(0005459) Andrew-R 2023-02-17 03:53	It seems Natron can re-read its own output, but you need to link writer and reader nodes? https://github.com/NatronGitHub/Natron/issues/545 I hope it also works with sequences ...

(0005458) PhyllisSmith 2023-02-17 02:51	@Andrew-R I will test tomorrow.

(0005456) Andrew-R 2023-02-17 01:46 (Last edited: 2023-02-17 18:32)	@Andrea_Paz, my idea behind using background rendered set of images was to modify them "behind cinelerra's back", so we can play and render from them as usual, without manually creating fileluit index (there is script for this, as far as I remember). Not sure if Natron can just overwrite all images in sequence? You need to skip very first 0 sized pseudo-index file, but apart from this it should work? Note, if you change timeline in cingg she will overwrite part you worked on, so Natron actions (is there way to make macros in there?) must be redone (at least on image range modded by cingg). Hopefully renderfarm + bg render also still work ( Phyllis, can you test at least on your big machine?) so, cinelerra-specific operations still benefit from all connected computes with cinelerra render (but you need to update cins on render nodes, so they agree on image format extension we just added).

(0005455) Andrea_Paz 2023-02-16 21:37	Thank you Andrew. With your patch the background render produces the images with extension, so you can import them and work in Natron (and I guess also in Blender or DaVinci Resolve, etc). So we can apply Hue curves, but also Rotoscoping and tracking. Then we can render a video or again a new sequence of images to import into CinGG. If we choose a new sequence of images we have to create an index (filelist) that is recognized by CinGG; how can we do that? Natron does not create a filelist and therefore CinGG cannot import it. Or do you think it is possible to import the edited images from Natron without having to do a rendering?

(0005454) Andrew-R 2023-02-16 19:46	Oh, I was blind, there already was filename extension variable! use on top of 0001 or just add this line to original file manually.

(0005453) Andrew-R 2023-02-16 19:26	I tried to make cingg add filelist extension automatically for brender files. It seems to work but obv. need more testing with renderfarm and usual lists ...

(0005452) Andrea_Paz 2023-02-16 15:16	I also think the idea is great. I tried a background render (exr sequences) and noticed that it doesn't put the extension to the images created. Natron doesn't recognize them and can't import them. Reading Natron's documentation you can see that it wants a simple list of images, without the need for a filelist ("name.exrs") as an index. So the background render is perfectly suitable if you can assign the extension to the various files you create. However, you don't get out of the usual mode of making exchange renderings between programs...

(0005451) PhyllisSmith 2023-02-16 13:57	@Andrew-R That is an interesting possibility and would be useful for other potential plugins too.

(0005450) Andrew-R 2023-02-16 02:14	Thinking about possible workaround - may be Natron can work over our background render files, and then if not owerwriting them but at least create new set, and then bash script just renames natron-modified files back into format cinelerra accepts?

(0005447) Andrea_Paz 2023-02-15 09:37 (Last edited: 2023-02-15 09:39)	Hue curves are complementary to the histogram (also called RGB curves). https://helpx.adobe.com/my_en/premiere-pro/using/adjust-color-rgb-hsl-curves.html It would be nice to be able to implement ofx in CinGG, it would fill the few shortcomings of CinGG in CC and Compositing. As Phyllis said, one workaround is the Banana Blue plugin, but it is complex to use and inaccurate as results.

(0005446) Andrew-R 2023-02-15 05:39	also, just looking at GUI may be Histogram Bezier can be modified for HSV in addition to RGB? https://cinelerra-gg.org/download/CinelerraGG_Manual/Histogram_Bezier_Curves.html

(0005445) Andrew-R 2023-02-15 04:21	code for this effect in Natron seems to be here: https://github.com/NatronGitHub/openfx-misc/blob/master/HueCorrect/HueCorrect.cpp but may be whole ofx host work will be better idea?

(0005444) PhyllisSmith 2023-02-14 18:05	@goat We definitely need this effect. I checked to see if this could be done with Blue Banana, but it does not appear to be able to.

Notes
(0005497) Andrew-R 2023-06-06 17:37	I guess we can at least put some comment in code saying "if you want bigger node point try to multiply HANDLE_W/HANDLE_H" (otherwise we forgot how to do it!)

(0005463) PhyllisSmith 2023-02-28 01:20	@Andrew-R This solution worked very well. AVLinux does not do builds though so now that I know where to look, maybe I can come up with something for his use.

(0005462) Andrew-R 2023-02-27 21:29	So, I looked into code.... cinelerra-5.1/cinelerra/trackcanvas.C function void TrackCanvas::draw_auto(Auto current, int x, int y, int center_pixel, int data_h) { int x1, y1, x2, y2; x1 = x - HANDLE_W / 2; x2 = x + HANDLE_W / 2; y1 = center_pixel + y - HANDLE_W / 2; y2 = center_pixel + y + HANDLE_W / 2; if( y1 < center_pixel - data_h/2 ) y1 = center_pixel - data_h/2; if( y2 > center_pixel + data_h/2 ) y2 = center_pixel + data_h/2; draw_box(x1, y1, x2 - x1, y2 - y1); } HANDLE_W defined in trackcanvas.inc as #define HANDLE_W xS(8) #define HANDLE_H yS(8) so I guess it already scale-able with gui scale setting, just not independently? Anyway, you can try to (multiply) this specific use oh HANDLE_W (*2 for example) in this specific function and see if it works as intended? May be other mouseclicking functions in same file will need same enlargement, too

Notes
(0005495) Andrew-R 2023-06-06 17:29	update: I found external compiz 0.8.x plugin doing ICC transform. https://github.com/compiz-reloaded/compicc https://sourceforge.net/p/compicc/wiki/Home/ "The Compiz ICC colour server, or short compicc, lets you colour manage your whole desktop at once and in hardware. Play movies, watch images on wide or narrow gamut displays. Each connected monitor is colour corrected for its own. CompICC is written to help the many colour management unaware applications. It takes them off any decision, what to do around monitor profiles. To do so CompICC does its work unquestioned. The result is a more consistent desktop. CompICC follows the opt out colour correction paradigm. Colour management unaware applications can continue to assume they are drawing in sRGB. They do not need any ICC profile nor any new library to do so. sRGB is the standard colour space for the internet. There are some applications, which want to do own colour corrections for displaying large gamut photos, proofing or monitor calibration. Those need to tell CompICC not to colour correct certain regions of their windows. This means slightly more work for aware applications to do advanced colour management other than sRGB." idea is to keep Compositor fullscreen, and teach this plugin about selectable input colorspace, so it will do both ICC profile transform and display colorspace transform from linear/user defined at timeline.

(0005299) Andrew-R 2022-05-06 07:58	ffmpeg git (on road to 5.1) has some work on ICC profile support in filters https://git.ffmpeg.org/gitweb/ffmpeg.git?a=search&h=HEAD&st=commit&s=icc

(0005189) Andrew-R 2021-11-16 16:59	there is code from mlayer2, but I sadly have no skills for transplanting it.. https://github.com/astiob/mplayer2/blob/all/libvo/vo_gl3.c

(0005188) Andrew-R 2021-11-16 16:38	well, sadly implementing this a bit more than I can take. I heard GSOC now is not limited to students - but for practicating as project we need programmer who understand Cin well enough, and those ppl a bit rare..

(0005187) Andrea_Paz 2021-11-16 12:30	Looking at Olive 0.2 for 10-bit support, I saw their embedded OpenColorIO. They integrated Blender's ocio.conf file (with associated LUTs) to get that support. It can be found in: /home/$USER/.cache/olivevideoeditor.org/Olive/ocioconf/ocio.conf Do you think it is possible to use it in CinGG as well? https://opencolorio.readthedocs.io/en/latest/quick_start/for_devs.html

(0002200) Olaf 2019-10-01 10:25	OpenColorIO makes sense, is currently also built into Olive. But it doesn't use anything as long as nobody can integrate it.

(0002185) Andrew-R 2019-09-27 06:38	Probably not very related to Compositor, but there are some patches for ffmpeg adding ICC profile reading/writing (against ffmpeg 4.3-dev): http://ffmpeg.org/pipermail/ffmpeg-devel/2019-September/250398.html [FFmpeg-devel] [PATCH 0/5] adding ICC profile support to MOV decode/encode Also potentially interesting patchset about dv encode/decode: http://ffmpeg.org/pipermail/ffmpeg-devel/2019-September/249794.html [FFmpeg-devel] [PATCH] avcodec/dvdec: correctly decode bottom mb row in 1080i field mode and next few patches ( [PATCH] avcodec/dvenc: support encoding dvcprohd , [PATCH] avcodec/dvdec: correctly set interlaced and tff)

(0001663) Andrew-R 2019-06-05 19:53	yeah, OpenColorIO seems to be more complete (capable?) solution. In c++. With GPU path (hopefully GLSL, and not just Cg from NVIDIA proprietary past) , too. http://opencolorio.org/developers/usage_examples.html But probably complete support will require also saving this 'color profile' info into image formats, and reading it back? Still, display seems to be first logical (most visibly) step.

(0001660) PhyllisSmith 2019-06-05 15:11	I will close 64 since this refers to it and since I can still look at it for future information when/if we ever get to this ticket. Other than bugs and looking into CUDA, issue 0000002 of MASKING and all of its side issues is the oldest so takes the highest precedence.

(0001659) Andrea_Paz 2019-06-05 14:14	A subject I'm very interested in. On the color management there is also an old ticket (0000064); maybe you can close it and leave only this.

(0001657) Andrew-R 2019-06-05 11:09	and discussion on CPU-only threaded use: https://sourceforge.net/p/lcms/mailman/message/30848823/ (relative to lcms2-2.4) --------copy------ Hughes wrote: > Hi all, > > I'm trying to make my transform go fast. I've got a 1920x1080 RGB > image being transformed from sRGB to the display profile. I've got a > quad core processor on my development box, no shaders or GPU, and I'm > trying to do the transform as quickly as possible. Re: [Lcms-user] Threading performance in LCMS2 From: Sebastien Leon <sleon@po...> - 2013-05-30 20:32:04 Hi Richard, [....] - littleCMS legacy code : 92.12 CPU Cycle per pixel transformed - littleCMS hacked : 27.75 CPU Cycle per pixel transformed --------copy end----- so, for 2Mpx image (~fullHD) it will be ... 93 * 1920 * 1080 = 194,273,280 cycles. *30 fps = 5,828,198,400 = 6Ghz core ... Note at the same thread it was said "Note that I also rewrote the critic loop with SSE4 assembly code (just for fun). I found no real improvement because most of the work is about memory exchange... So I kept the basic C code..." . Other links also alive (about measuring time spend in those transforms). Not sure if those transforms will work directly on yuv data ... so, probably initial implementation can only work with some project colorspaces (RGBA8/float?) and some video outputs (GL/X11) ?

(0001653) Andrew-R 2019-06-04 18:36	Not sure if this will be of any use, but there was little discussion about OpenGL and iccpprofiles: https://sourceforge.net/p/lcms/mailman/message/33575634/ -----------------copy-------------------- Re: [Lcms-user] Adobe's Color Engine, GPUs, And Other Rambling From: Gerhard Fuernkranz <nospam456@gm...> - 2015-03-10 19:25:03 Attachments: Message as HTML Create an RGB -> RGB transform and re-sample it on a regular grid, for instance like below (code is a bit older, still using lcms1, though). Regards, Gerhard #define GRIDPOINTS 64 static GLushort clut[GRIDPOINTS][GRIDPOINTS][GRIDPOINTS][3]; static GLuint img_texture; static GLuint clut_texture; static GLfloat clut_scale; static GLfloat clut_offset; void make_clut_texture(const char profile_name) { int r, g, b, n = GRIDPOINTS; cmsHPROFILE sRGB, monitor; cmsHTRANSFORM xform; sRGB = cmsCreate_sRGBProfile (); // image profile monitor = cmsOpenProfileFromFile (profile_name, "r"); if (monitor == NULL) { fprintf(stderr, "Cannot open display profile %s\n", profile_name); exit(1); } xform = cmsCreateTransform (sRGB, TYPE_RGB_16, monitor, TYPE_RGB_16, INTENT_PERCEPTUAL, cmsFLAGS_NOTPRECALC); if (xform == NULL) { fprintf(stderr, "Failed to create transformation\n"); exit(1); } clut_scale = (double) (n - 1) / n; clut_offset = 1.0 / (2 n); for (r = 0; r < n; r++) { for (g = 0; g < n; g++) { for (b = 0; b < n; b++) { unsigned short in[3]; in[0] = floor ((double) r / (n - 1) * 65535.0 + 0.5); in[1] = floor ((double) g / (n - 1) * 65535.0 + 0.5); in[2] = floor ((double) b / (n - 1) * 65535.0 + 0.5); cmsDoTransform (xform, in, clut[b][g][r], 1); } } } cmsDeleteTransform (xform); cmsCloseProfile (monitor); cmsCloseProfile (sRGB); glGenTextures (1, &clut_texture); /* texture 1 = clut / glActiveTextureARB (GL_TEXTURE0_ARB + 1); glBindTexture (GL_TEXTURE_3D, clut_texture); glTexParameteri (GL_TEXTURE_3D, GL_TEXTURE_WRAP_S, GL_CLAMP); glTexParameteri (GL_TEXTURE_3D, GL_TEXTURE_WRAP_T, GL_CLAMP); glTexParameteri (GL_TEXTURE_3D, GL_TEXTURE_WRAP_R, GL_CLAMP); glTexParameteri (GL_TEXTURE_3D, GL_TEXTURE_MAG_FILTER, GL_LINEAR); glTexParameteri (GL_TEXTURE_3D, GL_TEXTURE_MIN_FILTER, GL_LINEAR); glTexImage3D (GL_TEXTURE_3D, 0, GL_RGB16, n, n, n, 0, GL_RGB, GL_UNSIGNED_SHORT, clut); / back to texture 0 = image */ glActiveTextureARB (GL_TEXTURE0_ARB); } -------------------copy end------ additionally there was this comment at the end: ----copy------------ Re: [Lcms-user] Adobe's Color Engine, GPUs, And Other Rambling From: Gerhard Fuernkranz <nospam456@gm...> - 2015-03-10 23:54:12 Am 11.03.2015 um 00:36 schrieb marti.maria@...: > > Quoting Gerhard Fuernkranz <nospam456@...>: > >> Create an RGB -> RGB transform and re-sample it on a regular grid, >> for instance like below (code is a bit older, still using lcms1, though). > > Great idea. lcms2 works fine in this way. Please note in lcms2 you can get float, double or "half float" values directly, so you need not to downscale to 16 bits anymore. For instance, in half float type, use TYPE_RGB_HALF_FLT as formatter. The reason for the integer LUT texture was actually the inability of my GPU to interpolate 3D float textures with GL_LINEAR filter, when I wrote this stuff quite some years ago. Likely, recent GPUs don't suffer from such a limitation any more. Regards, Gerhard --------------------copy end------------------- so, this might help with X11-OpenGL. For plain X11 - just same thing via CPU? (GIMP 2.10.x probably uses software/CPU method)

Notes
(0005494) Andrew-R 2023-06-06 17:22	no progress on git log, but somewhat interesting articleon programking with xlib: https://handmade.network/forums/articles/t/2834-tutorial_a_tour_through_xlib_and_related_technologies https://www.tonyobryan.com/index.php?article=9 ("Setting An X11 Window to Full Screen") I think I saw some strange behavior running compiz + TDE + cingg when it comes to recording virtual desktops, so worth investigating how those attribs works.

(0005378) PhyllisSmith 2023-01-11 16:51	Waiting to hear possible results. Meanwhile want to log how Rob did this: "implemented such at my last job by saving git-commit-string as a macro that could be read by the compiler and expanded during compiler preprocessing. the slight complication with your build system is that you don't always build from a git repo so in lieu of being able to call "git log" you would need an authoritative fall-back text file .version??? as part of the build." True, I never build from the git repo.

(0005376) Andrew-R 2023-01-11 03:33	for git hash google pops up this idea: https://gist.github.com/lindenb/3c7f0cc391749802a120 in makefile githash.h: echo -n '#ifndef GIT_HASH\n#define GIT_HASH "' > $@ && \ git rev-parse HEAD \| tr -d "\n" >> $@ && \ echo '"\n#endif' >> $@ then one can include this githash.h in our version.h (and not forgot to rm it on make clean. I'll try this one

Notes
(0005488) Andrew-R 2023-06-01 15:18	There are gtk and qt bugs about 30bit support (qt5 supposed to be fixed) https://bugzilla.gnome.org/show_bug.cgi?id=625202 (not sure if patch was ever officially applied?) https://codereview.qt-project.org/c/qt/qtbase/+/86610

(0002151) Andrew-R 2019-09-16 02:31	glxinfo just in case ...

(0002150) Andrew-R 2019-09-16 02:24	second gdb log, this time cin was build with clang, yet ...weird colors still around! and yes, I can see this line: BC_DisplayInfo::gl_fb_config failed

(0002149) Andrew-R 2019-09-16 02:15	gdb log (without patch from https://git.cinelerra-gg.org/git/?p=goodguy/cinelerra.git;a=commit;h=a2ebbc5e88800bd740652a32285f8145a1acfdc6 , it doesn't compile for me with gcc 5.5.0) as far as I can see $1 = {ext_data = 0x0, visualid = 0x2a9, c_class = 0x4, red_mask = 0x3ff, green_mask = 0xffc00, blue_mask = 0x3ff00000, bits_per_rgb = 0xa, map_entries = 0x400} actually chooses visual: visual id: 0x2a9 class: TrueColor depth: 30 planes available colormap entries: 1024 per subfield red, green, blue masks: 0x3ff, 0xffc00, 0x3ff00000 significant bits in color specification: 10 bits will try to rebuild with clang ... (after patching again)

(0002147) Andrew-R 2019-09-15 22:10	Yeah, will follow instructions ... Right now xdpyinfo looks like this (on 30bpp screen)

(0002145) PhyllisSmith 2019-09-15 20:36	From GoodGuy: See the attached text.

(0002091) Andrew-R 2019-09-08 18:30	https://wiki.gentoo.org/wiki/30bpp - Gentoo wiki. It references forum thread, where you can see some more detailed observations about what works and what not yet .....

Notes
(0005474) PhyllisSmith 2023-05-16 18:08	@Andrea_Paz update: " the ability to insert subfolders nested inside the folders we create in the Resources window. . As a non-programmer I would have said it was a relatively simple request to implement, since there is already the ability to create folders, but I'm probably wrong. The ability to create subfolders nested within each other allows for optimal source management, much more powerful than what we can do now by creating folders without subfolders. This is a pro feature (called Logging) because it is indispensable for very large projects with dozens or hundreds of files that need to be searched and accessed quickly." I would love it if CinGG could have this functionality.

(0004651) fary54 2021-02-17 07:33	@Andrea_Paz Excellent proposal. It would be very useful for large projects. It would make it easier to find, classify and much clearer.

(0004649) PhyllisSmith 2021-02-16 18:25	Oh, now I understand and this would be a nice addition. You can create a folder within a folder already but all of the files have to be in the preceding level also so is not like a directory structure at all. It would be great to have it look like the expanders of the Audio Effects.

(0004648) Andrea_Paz 2021-02-16 17:21	Sorry, I explained it wrong. I meant to create a personal folder and within that create several sub(sub)folders: -... -Media -MyFolder1 -Subfolder11 -subfolder12 -Myfolder2 -...

(0004647) PhyllisSmith 2021-02-16 15:52	Although it may be different than what is expected and CinelerraGG philosophy has always been to not copy other people's work (to quote another person, if you like the way it works somewhere else, use that software instead!!) subfolders are available for usage. In Resources window, right click on the Media folder and there is an option to create a folder. It is in section 2.5.2. But this BT is a good notice to alert other people of subfolders existence.

Notes
(0005441) Andrew-R 2023-02-12 01:50	sorry, not looked into this yet

(0005423) Andrea_Paz 2023-02-05 20:45	I confirm that the test result is the new project with 2 audio tracks but with only the first channel repeated 2 times. However, the first channel is 100% on the left channel and the second channel is 100% on the right channel. So it is not a true mono, but a stereo with the 2 equal channels (same as the first channel of the original project).

Notes
(0005380) IgorBeg 2023-01-14 10:50	In the Manual FAQ, Known Problems and Workarounds -> Audio tracks align with no video track, and in working with Mixers you can find the Workaround: https://cinelerra-gg.org/download/CinelerraGG_Manual/FAQ_Known_Problems_Workarou.html I put here the section about: ========================= The use of Open Mixers always requires the existence of a video track. See But, I want to use only the first set of audio tracks... We cannot act on audio tracks only, for example to align them. A workaround to get the audio tracks aligned would be: 1. You create/open the Mixer Viewer with Open Mixers in Resources window for the Videos with Audio. It will create, in the timeline, one Video track and two Audio tracks for each Media (Video/Audio), and its Mixer Viewer. 2. Add a new Video track (Shift+T shortcut) and two Audio tracks (t shortcut) for your audio (Left+Right channels?) without video: you are adding an empty Video track because Align Mixers feature needed a video track. 3. Move the Video track down until it is on the top of the two Audio tracks. 4. Insert your Audio into the two Audio tracks in the usual way. 5. Open/Create a Mixer Viewer from Menu → Window → Mixer... → Mixer Viewer (Shift+M). 6. Assign that Mixer Viewer (a white frame is shown when selected) to the one Video tracks (empty) and two Audio tracks, of the 0000002 point, pressing the arrow icons (Mixer) you find on the Patchbay on the left of the Main window. When assigned, you must to see the 3 arrow (for 1video + 2Audio) are pointing up. 7. Now, you can use Align Mixers feature as usual. See Audio/Video sync using Waveforms/Timecodes =========================

(0005379) PhyllisSmith 2023-01-11 17:45	@Andrea_Paz I fixed the problem with the zip file -- awhile back I could not remember what that file was, saw it was big, so deleted it. My mistake. It is now back with a name that I should be able to recognize and I checked into GIT the corrected name. Thanks so much for your feedback on this issue.

(0005377) Andrea_Paz 2023-01-11 13:17	I confirm that the audio only track (the external and main one) is not opened with OpenMixers; only those "video + embedded audio" tracks are opened and loaded into timeline. I don't know if it is possible to implent audio-only tracks in OpenMixers as well. One workaround is to manually insert the external audio track, then open an empty mixer and connect the external track to that mixer. The other tracks with embedded audio are opened normally with OpenMixers. Try following the example found in the manual: https://cinelerra-gg.org/download/CinelerraGG_Manual/Workflow_with_multi_cam_ext.html (Note: if you want to download the files in the example the link on the page is no longer valid; you can use the link to my google drive if you want: https://drive.google.com/file/d/1XfIlyF60NWGlcZsL1efgfkx2fGria3RL/view?usp=share_link) Another link that explains well how to associate a track (in the example a video track, but the same applies to audio tracks) with a mixer is the following: https://www.youtube.com/watch?v=49EExsRsjlY @Phyllis No point in leaving a link in the manual that leads to nothing. You can remove it entirely or you can add my link to Google Drive; keep in mind, though, that dealing with Google is not the best in terms of privacy.

(0005375) PhyllisSmith 2023-01-11 01:24	@webmadman : @Andrea_Paz created that video so hopefully he will have some advice. Maybe this can be of some help: https://cinelerra-gg.org/download/CinelerraGG_Manual/Audio_Video_sync_via_Wavefo.html

View Issue Details

ID:	Category:	Severity:	Reproducibility:	Date Submitted:	Last Update:
596	[Cinelerra-GG] Bug	crash	always	2021-11-02 12:42	2023-11-26 01:58

Reporter:	sparkillay	Platform:	Linux
Assigned To:	PhyllisSmith	OS:	Bodhi Linux 18.04
Priority:	normal	OS Version:	5.1 32bits i386
Status:	acknowledged	Product Version:
Product Build:		Resolution:	open
Projection:	none
ETA:	none	Fixed in Version:
		Target Version:

Summary:	Crash when uploading files-video cin gg 32bits
Description:	error while loading shared libraries: libva.so.1 Rename: libva.so.2 to libva.so.1 libva-x11.so.2 to libva-x11.so.1 libva-drm.so.2 to libva-drm.so.1 I got into the interface but I can't upload files. I can render video, upload images and upload audio. I can't upload video. codecs tried: H264 VP8 MPEG2video I use the November and October version appimage i386.
Tags:	Load Files
Steps To Reproduce:	Upload any video.
Additional Information:	My computer is old but I only use it to support 32-bit distributions.
Attached Files:	cinelerra_5125.dmp (51,912 bytes) 2021-11-02 12:42 https://www.cinelerra-gg.org/bugtracker/file_download.php?file_id=864&type=bug gdb Cini386 default (4,529 bytes) 2021-11-02 17:30 https://www.cinelerra-gg.org/bugtracker/file_download.php?file_id=865&type=bug gdb (5,516 bytes) 2021-11-05 02:02 https://www.cinelerra-gg.org/bugtracker/file_download.php?file_id=866&type=bug cinelerra_12989.dmp (58,314 bytes) 2021-11-05 02:02 https://www.cinelerra-gg.org/bugtracker/file_download.php?file_id=867&type=bug LD_DEBUG (35,648 bytes) 2021-11-05 02:02 https://www.cinelerra-gg.org/bugtracker/file_download.php?file_id=868&type=bug Video_work (7,751 bytes) 2021-11-05 18:48 https://www.cinelerra-gg.org/bugtracker/file_download.php?file_id=869&type=bug Videos fail crash cinelerra (9,113 bytes) 2021-11-05 18:48 https://www.cinelerra-gg.org/bugtracker/file_download.php?file_id=870&type=bug Screen_video_fail.mp4 (322,370 bytes) 2021-11-06 03:14 https://www.cinelerra-gg.org/bugtracker/file_download.php?file_id=871&type=bug simplescreenrecorder-2021-11-05_22.10.34.mp4 (630,702 bytes) 2021-11-06 03:14 https://www.cinelerra-gg.org/bugtracker/file_download.php?file_id=872&type=bug gdb bt full (6,919 bytes) 2021-11-08 21:25 https://www.cinelerra-gg.org/bugtracker/file_download.php?file_id=873&type=bug Screenshot_20231126-045354.png (218,683 bytes) 2023-11-26 01:58 https://www.cinelerra-gg.org/bugtracker/file_download.php?file_id=934&type=bug

Notes
(0005374) Andrew-R 2023-01-10 09:14	i think this might be a bit more involved than just appending directory to filepath.. Because user might have two different but same-named files in different locations. And just turning full path into filename might fail easily because full filename length limited to 128-255 chars.

Notes
(0005373) IgorBeg 2022-12-16 13:21	You are not alone. ;-) It doesn't depend by Browser, unfortuantely. You can see "0000619: Forum annoying issues, but not detrimental": https://www.cinelerra-gg.org/bugtracker/view.php?id=619 and "Recent Posts no updated": https://www.cinelerra-gg.org/forum/announcements-and-updates/recent-posts-no-updated/ Sometimes I have to swith to another Language, from English to Deutsch (or other), to reply and I don't know/understand German.

(0005372) PhyllisSmith 2022-12-13 16:37	@blueskies3 Oh... part of the problem of what you see in your case is that when you make an entry in the Forum, I am getting an email asking me to "Please Moderate" so I have to login to the website and Approve your Forum entry. With the time differences, I do not get this done until several hours later after you post. It takes several entries by a user before they automatically get approved. Again, this is because we get weekly Spam about completely offtopic product endorsements, gambling, etc. However, I will update the manual to make notes of these issues.

(0005371) PhyllisSmith 2022-12-13 15:20	@blueskies3 "not sure if this is the right place to report such issues" -- you can report issues anywhere that works for you as we want to know about potential problems. I think the issue with the Forum could be due to the way CloudFlare works to propagate it throughout the entire world on some time basis. Several users have reported that sometimes it does not show items for several hours after they have been posted. "Forum not accessible from Firefox" - I always use Firefox and almost daily access the forum (using Fedora) and have not seen any problem viewing it - except for the aforementioned issue. "'Add comment' button does not respond" - I have seen this problem at least twice and have found that if I logout and login again a few minutes later, it then responds. I will put in more effort the next time I see this to see if I can find the root cause. "started has no buttons to reply to my post" - I think this is a result of not actually being logged in. Just an Important additional note -- we get quite a bit of Spam on our Forum and have had to do a lot of blocking of ip addresses to help to alleviate this, which potentially could impact the routes taken.

(0005370) blueskies3 2022-12-13 05:28	I would like to update that I seem to be able to access the forum once again using the Chromium browser. Another note is that I had difficulties with both Ubuntu 20 and Ubuntu 22, Chrome Version 108.0.5359.98 (Official Build) (64-bit) & Firefox version 107.01 from a snap build. Thankyou.

Notes
(0005364) PhyllisSmith 2022-11-20 21:27	Updating this based on IgorBeghetto last entry in mailing list where you can read the full details: https://www.mail-archive.com/[email protected]/msg05173.html Fix as Andrew provided in line 104 of "virtualvconsole.C" from 0xff to 0x00 With this fix: - old projects work - tested Open EDL feature using EDL_Files.zip by RafaMar as a sequence on the top of project works - test with Mixer Viewer as shown in Tutorial "Cinelerra GG: Multi Cam editing (with Proxy)" https://www.youtube.com/watch?v=49EExsRsjlY works - STILL A PROBLEM when the Nested clip with fade In is pasted in the Master project

(0004692) IgorBeg 2021-03-14 08:27	I understand you, Phyllis. It is not easy to find a right way. Thank you!

(0004691) PhyllisSmith 2021-03-14 00:50	I deleted the porterduff-srcover and made the suggested wording changes/additions. As Andrea can attest, it is quite difficult to keep the documentation updated and that is why it is better to have all of the documentation in only one place in the Manual. We already have trouble keeping the Shortcuts updated since they are in 2 places and are becoming out of synch. OK, I will work on adding a section in the chapter "Troubleshooting and Help" and will call it "Workarounds for Known Problems". So far I have 3 items to go in as follows: 1) this OpenEDL alpha problem 2) Fade/flash bug 3) clips with transitions and Motion 51 plugin 4) others?

(0004690) IgorBeg 2021-03-13 10:28	Thanks Phyllis! I read the section about the workaround. It seems good to me. You know, English/US is not my first language. If possible I would change... - From "... using the middle mouse button on ..." To "... using the MMB (Middle Mouse Button) on ..." - In the sentence "Change the value of the Alpha to 0.00." I would add "The Alpha Plugin may be disabled". I agree with RafaMar about "the one of using PorterDuff-> SrcOver I would not include". It can be used for other reasons and in other section of the Manual. What do you think about create a new "FAQ and Known Problems" section in the Manual? I don't know if it is complicated and it has sense for you. So, an user, like me, may search if someone found a trick/workaround for a problem. Maybe this a question to report in Mailing List.

(0004689) RafaMar 2021-03-13 07:48	Thank you very much @IgorBeg My idea was simply to put a help button after the "EDL Strategy" buttons that would link to a short manual that explains to a user the correct way to use these options and what each of them does. @PhyllisSmith the solution of using Alpha's effect is perfect for both situations, the one of using PorterDuff-> SrcOver I would not include, since this option sometimes causes an undesired result.

(0004687) PhyllisSmith 2021-03-12 23:32 (Last edited: 2021-03-12 23:38)	@IgorBeg , please check the wording in the following (the section beginning with the word Workaround) https://cinelerra-gg.org/download/CinelerraGG_Manual/Nested_Clips.html#sub:nested_clips @RafaMar , about "doing something similar to what I did with the Batchrender Assistant? A help button opening a PDF that explains these things very abbreviated." It is a good concept but this was does not have a wizard (menu) like Batch Render does. With the advent of the HTML version of the Manual, it is easier to narrow down which section to look for workarounds. BT 0000538 is still open and possibly the "h" key will be hooked up in the future to make using the HTML lookup really easy. @All , BUT we still need a fix for this error and if you carefully read the manual on the nested section, you see the following statement: "creating nested clips is a complicated internal process for CinelerraGG, especially for managing audio tracks". And I believe there is another issue with nesting sizes that was presented by @fary54 . It is BT 0000440 so that if a programmer comes along to work on this bug, it can all be addressed at the same time. Let us see if anyone responds to the other 2 bugs that need fixes and have a PayPal donation associated with them.

(0004686) IgorBeg 2021-03-12 09:34	Thanks to both, RafaMar and Phyllis. Mine is just a trick to (temporarily?) bypass the issue. <code>RafaMar wrote: "... with the alpha at 0 and without the need to disable the effect.."</code> In my #4 point, I know that it works even with the Plugin enabled. I disable the Plugin because, I think, it consumes less cpu, or at least I hope. Maybe Phyllis know if it is true or not. <code>RafaMar wrote: "what do you think ... A help button opening a PDF that explains these things very abbreviated."</code> Sorry, I don't like the Help button inside Plugins to link an external PDF. For me it has no sense. It is true, there is a checkbox Help in Mask tool and Sketcher plugin but it is a different thing (... and they should be "harmonized"). I think the Manual is there for this. An User who wants to learn some things should download the Manual and use the "Find a word or phrase in the document" function to search for what He needed there. He could learn more from the Manual. If something is unclear , and it may be, there is the support from other Users in the Forum, Mailing List, and MantisBT. I am sorry but I think so.

(0004682) RafaMar 2021-03-11 20:26	The normal thing would be that by nesting the EDL, the transparency operated without problems, but as not being the case, the first solution that Igor proposes is very good, then now it is simply that people know. And in this way it is not necessary to touch the code. I know that the ideal would be that Cinelerra did this internally and automatically, but when you know the way to solve it is also just as valid. Phyllis, what do you think of doing something similar to what I did with the Batchrender Assistant? A help button opening a PDF that explains these things very abbreviated.

(0004678) PhyllisSmith 2021-03-11 19:50	Yes, I will add this to the manual as a solution after having @IgorBeg review my wording to make sure I say it correctly. He is quite the expert when it comes to using nesting.

(0004676) RafaMar 2021-03-11 18:36	Hi @IgorBeg, thank you very much, the trick of setting the alpha effect has worked perfectly in both cases, with the alpha at 0 and without the need to disable the effect, both the one that enters by fading and the one that enters with the projector. Thanks to you I can continue with the tutorial on the EDL strategy with a very good solution for the alpha issue when working nested or as a reference. The solution seems to me so good that even @PhyllisSmith could indicate this process, in the Cinelerra GG manual, when we make a clip that we want to insert as a reference or nested. And with your first solution this issue is solved. Thank you very much for your help.

(0004672) IgorBeg 2021-03-11 12:47	In the meantime a programmer came out,... this is my workaround... The trick for the Projector problem on the Sequence when it is used in a Master Project. 1. Open your EDL (the Sequence) with "Open EDL" by MMB (Middle Mouse Button) on the Nested Edit on the track (or as you prefer). 2. In the Sequence, insert a new Video track to the bottom of the last Video track (in your video between Video_2 and Audio_1) 3. Added the Alpha plugin (Effect) to the new Video track for all the length of the Sequence. Probably you can use another Plugin but I use that because it remember me why I am using it. 4. Change the value of the Alpha to 0.00 and Turn OFF the Alpha Plugin. Yes it is strange you turn OFF the plugin but this Video Track is a kind of fake. 5. Close EDL. 6. Now the Master Project works right. The trick for the FadeIn/Out with autos. 7. In the Master Project, in the Video Track where you putted the Nested EDL, in Patchbay change the "Overlay Mode" from Normal to PorterDuff->SrcOver. I know it is not the better but... better that nothing. ;-) Please, tell me if it is clear or you needed a demo video. Thanks!

(0004669) RafaMar 2021-03-10 07:39	This is an alpha channel error when using an EDL inserted as a reference or nested. This EDL acts as if it is set to RGB or YUV with no alpha channel and should work as RGBA or YUVA. The problem is that I don't know how to program in C to be able to fix this, I'm sorry, that's why I'm reporting it, in case some programmer can fix it.

(0004668) PhyllisSmith 2021-03-09 22:30	I have only run the demo one time but it is going to take some real study on my part to understand what is going on.

(0004667) RafaMar 2021-03-09 14:40	Files I have used for the video example.

Notes
(0005363) PhyllisSmith 2022-11-19 21:52 (Last edited: 2022-11-19 22:48)	Both Andrea and Andrew have confirmed on several different build/computer types. Workaround for now is to not change Color Model Format in session just before loading backup.

Notes
(0005362) terje 2022-10-05 16:38	@Phyllis Yes, let me try to explain a bit more and hopefully better .... As I see it, Cin-CV-GG has common roots and background from Cin-HV. What I suggest is to 1) if possible extend the time span to search the missing part of the current CV and GG email Archives. As now, the time span for Cin-CV email archive is 2014 - 2020, Cin-GG archive 2018-now 2022, while the common GG/CV seach cover only 2020-2022. That is we lack search for 2014-2019. 2) add or enable "linked ponters" to the somewhat unknown and old C(VS) Archive and Search. This just my suggested sample: The GG web page https://www.cinelerra-gg.org/search-mail-archives/ contains 2 magnifier (search) A minor bug here is that both currently points to email threads. My point here is that sorting by date is available at the top second left tab of the first magnifier's thread page. And edit the text below the magnifier correspondingly and inform the time span is for 2020-2022, hopefully from 2014 The second magnifier icon can then be available for the old Cin-CV(S) search as mentioned 2010-2019 2 envelopes (archives) The first i.e for a common Cin-CV+GG archives 2014-now 2022 and the second for the mentioned old Cin-CVS arhive 2001-2016

(0005361) PhyllisSmith 2022-10-05 01:07	@terje For 2 days now I have been trying to understand your suggestion in your latest note update. Being able to search Cin-CV is extremely useful but as you stated, you can already do that using the link you provided in the last line of your note. As far as I can tell, there is very little mail in Cin-CV past the current searchable post.

(0005359) terje 2022-10-02 11:23	Hello again. I have reopened this previous eMail Archive and Search topic, and want to suggest a completion of the old Cin-CV(S) email Archive. The reason is that there are valuable Cinelerra history and background information in this older email archive. Because this archive and searchability already exist, I think it will not require much work, if they just can be linked and used as is. Currently it looks like we have available email Archives and searchability for the following time spans on the Cin-GG website https://www.cinelerra-gg.org/search-mail-archives/ Cin-GG email archive: Nov 2018 - now Oct 2022 Cin-CV email archive: 2014 - 2020 Combined search GG+CV: 2020/10/26 - now Oct 2022 Possible to extend the search time span here? As menitioned in the last section of my initial post: "As a related sample, the (Old) [email protected] list is searchable, and many interesting Cinelerra subjects from the past can be looked up here in a quick and easy way." At that time the old Cinelerra email list [email protected] was hosted by Skolelinux ("school linux" project based on Debian Edu) https://en.wikipedia.org/wiki/Skolelinux old Cin-CV(S) email Archive: 2001 (mainly 2003) - 2016 https://lists.cinelerra-cv.org/pipermail/cinelerra-skolelinux/ Searchable old Cin-CV(S): 2010 - 2019 (mainly 2015) https://www.mail-archive.com/[email protected]/index.html

(0002028) PhyllisSmith 2019-08-18 16:44	Works well - and I meant to close this (going through issues this week to re-assess).

(0001401) Sam 2019-04-21 11:34	The desired feature has been integrated so that this ticket can be closed.

(0001390) Sam 2019-04-20 19:54	I changed the mail archive search again. https://www.cinelerra-gg.org/search-mail-archives/ The search via DuckDuckGo was really bad, because the content was never indexed correctly. I activated the service via Mail Archive https://www.mail-archive.com/[email protected]/index.html . All new emails from April 2019 onwards will be indexed and searchable. The emails before this date are not indexed, but at least the new messages will be archived. On the support page (https://www.cinelerra-gg.org/support/) I have put a link. The search page can now be found in the menu. I wouldn't like to change our Archive HMTL page, there is no option in Mailman, at least I haven't found any how to change it on the backend page. I would have to change them directly on the server. I'm not sure if it will be reset at the next update. I also leave such things on default to avoid accidentally creating security holes. I think who really wants to search has enough options to do it now. I have put a link to the old CV archive on the search page.

(0001232) Sam 2019-03-17 13:25	I just added the mail archives to the Search Console yesterday, that usually triggers the indexing of the content after a few days, so it takes some time until all posts are visible. At the moment there are very few posts available. Thanks for your suggestions, I'm happy to try them out. But I won't be able to implement it until the end of the week.

(0001231) terje 2019-03-17 10:37	@ Sam Thank you for doing that! I did a search test a got one hit, but not on a couple of others. The archive indexing is a time consuming process as mentioned. Add-on suggestions/questions: 1. Publish this search feature on the GG site (Support>Mailing list ...Cin Info page) and add it on top of the Mail archives site, optionally also add it in the foot note on received Mails https://lists.cinelerra-gg.org/pipermail/cin/ 2. When this GG mail archive is ready, would it be possible (an idea) to link also the Cinelerra cv mailing list archive, as it contains several cingg related topics and a lot of generic Cinelerra topics https://lists.cinelerra-cv.org/pipermail/cinelerra/

(0001225) Sam 2019-03-17 00:27	@terje @PhyllisSmith I have linked the mail archive with the DuckDuckGo search engine, i.e. we can now also search the mail archive. However, I have to emphasize that many posts have not yet been indexed. Indexing takes a long time, but once it's done, you can use this feature. An own search script which searches this archive was out of the question for security and performance reasons. By the way, the Google search is even worse when indexing such mailing archives, so I included DuckDuckGo. https://www.cinelerra-gg.org/search-mail-archives/ https://lists.cinelerra-gg.org/mailman/listinfo/cin

Notes
(0005360) PhyllisSmith 2022-10-04 16:41 (Last edited: 2022-10-04 16:45)	@deim I hate to say that this may be a "computer / operating system problem" with specific videos, BUT I have an AMD gaming laptop that will frequently crash when rendering the cows 4k webm video. This same rendering on an older, much slower HP computer never crashes. In the past, I have attempted to analyze it without having the skills to do so and therefore, I have no recommendations of how to go about debugging this. My only guess is that it has something to do with SMP (symmetric multiprocessing - computer processing done by multiple processors that share a common operating system and memory). Maybe a "batch" job uses the system resources in a different manner and that could be why that worked for you. When I render, I almost always use FFmpeg/mp4 without making any changes to the defaults and I have my Performance Preferences Project SMP cpus set to 16 which is the number on this particular laptop. In addition, since the render is slow, I am almost always doing some small task at the same time. Hoping someone else may have some suggestions. The only suggestions I have, which would be quite time-consuming because of crashes, is to vary the number of CPUs being used or change the render parameters. I am assuming that the only work you lose is the render time? because there are many backups of your actual editing work. Late EDIT: of course, now while testing this same render twice, it does not crash. Maybe FFmpeg 5.1 is a factor as this is the first time I have tried this render since the latest GIT.

Notes
(0005358) PhyllisSmith 2022-09-15 19:50	@goat Thanks so much for explaining your usage -- now it does make for a good potential solution. Unfortunately right now we are in need of a developer/programmer who can work on some of the more complicated / in-depth solutions. We will leave this Feature Request open so that eventually someone will come along and volunteer their time to help out. There is a small dollar amount available via PayPal if a volunteer wishes to tackle this.

(0005357) goat 2022-09-15 04:07	In Raspberry Pi 3 model B when i open videos in low resolution with turned on "Use thumbnails in resource window" that works smooth and i can see what which video is. But when i use fullHD videos i turns off this option (Use thumbnails in resource window). That's works pretty stable in proxy mode (I used 1/6). But sometimes i cant understand which video is and i should watch videos in Viewer window to understand which video is. And i think that it will be good to add option "Use proxy thumbnails only". I think that this option sometimes can be helpful on slow computers.

(0005356) PhyllisSmith 2022-09-14 20:29 (Last edited: 2022-09-14 20:30)	@goat Here is the sample menu. Please let us know if this is a sufficient solution already.

(0005355) PhyllisSmith 2022-09-14 20:29	@goat I may be misunderstanding, but for slower or less ram computers, you can already solve the problem in a different way. In the main Program window, there is a pulldown called "Settings". If you click on this and choose the option "Preferences", a menu comes up and if you highlight the "Appearance" tab you will see something similar to the attached file. Towards the bottom on the right hand side you see "use thumbnails in resource window" and if you uncheck this (like shown) no thumbnails will be generated but you will just have a stand-in image that is the same for all files of that type. In addition in that same "Appearance" tab in the upper right size you can use a smaller number for "View thumbnail size" and "Vicon quality" with "Vicon color mode" set to low. P.S. attachment will be added in the next note

Notes
(0005354) IgorBeg 2022-08-17 07:55	Mmh, I don't know. I don't know anything of WordPress. Is there any possibility to set, temporarily (1 or 2 weeks), the Forum only in English language to see how it works? All users write the posts in English so, I think, it wouldn't be a problem for 2 weeks, if possible.

(0005353) PhyllisSmith 2022-08-16 15:05	It seems like Problem 2 is easily solves according to the attached gif file, but I do not feel comfortable making the change.

(0005345) IgorBeg 2022-08-04 07:29	Thank you, @PhyllisSmith, to open the ticket. I think the problem is due to the multilanguages. The language used In MantisBt is only English and it works really good.

Notes
(0005331) PhyllisSmith 2022-06-13 21:14	@quintao Good suggestion - I just do not know how hard this would be to program but it would definitely have to be a choosable option so it would look the same for those who wanted it to be the same. ALTERNATIVE - instead of scrolling to the right to look for the icon controls to use them, just Right Mouse Click on the desired Title Bar and you see the icon equivalents without having to move all the way to the right to get to the icons.

Notes
(0005327) PhyllisSmith 2022-06-11 17:47 (Last edited: 2022-06-11 17:58)	I looked at the dump again more thoroughly and have still not found any "smoking gun". The same code (cinelerra and libraries) is executed whether it has a bitrate of 4 MB or 8 MB. That is why I still think some resource is being exhausted, but this usually shows itself in the dump with weird numerical/nonsensical values and I see none of those. I did look at the disassembly of the executed code as provided in the dump with line 20 below being the code executing when it crashed - this code could be anywhere! in the program so I am clueless. Disassembly: 0: 00 ba 20 00 00 00 add BYTE PTR [rdx+0x20],bh 6: 4c 63 83 24 93 00 00 movsxd r8,DWORD PTR [rbx+0x9324] d: ff 93 c8 bb 00 00 call QWORD PTR [rbx+0xbbc8] 13: 48 8b 44 24 60 mov rax,QWORD PTR [rsp+0x60] 18: 48 83 c4 40 add rsp,0x40 1c: 48 c1 e0 06 shl rax,0x6 20: 4c 8b 94 03 10 18 00 mov r10,QWORD PTR [rbx+rax*1+0x1810] 27: 00 28: 4d 85 d2 test r10,r10 2b: 74 2d je 0x5a 2d: 44 89 e8 mov eax,r13d 30: 45 89 e1 mov r9d,r12d I have completely run out of ideas so it looks like the only workaround is to use a different render format or smaller bitrate. Not sure if it is the O/S and I could not get CinGG from Elive to work on my 32-bit system (it installs and comes up but then just hangs and it is probably due to my total lack of experience). Oh, in case anyone else missed this the dump shows that the CinGG program was built EXE: built: Feb 3 2021 17:00:52 -- that is 2021, not 2022 so there is more changes than what I previously stated. P.S. we are aware that "vaapi" renders will crash in the AppImage unless the level of library code is an exact match of the computer on which Cinelerra was built.

(0005326) Andrea_Paz 2022-06-11 14:23	Yes, the difference between intraframe and interframe codecs is always very big. With intraframe the frames are all retained, more or less compressed. With interframe, a lot of frames are removed, which are then more or less reconstructed. I took your example clip and duplicated it to 1440 frames (58 sec). Rendering in mp4 @9Mbit results in a 68 MB file. DNxHR_proxy produces a 271 MB file (but of insufficient quality, it serves only as a temporary proxy). DNxHR_SQ --> 844 MB and DNxHR_HQ --> 1.2 GB. I guess you used the higher quality DNxHR (444 or HQX), they are useless starting with a 420p interframe. You could try avcintra_100.mxf (130 MB), which is more compressed (and lossy) but still intraframe.

(0005325) quintao 2022-06-11 12:57	OK, through the manual I found mxf, which I hadn't noticed before. The DNxHR options produce some pure quality but massive files... 1 min @6Gigbytes! But there is also mpeg2 as well, and h264, which I can use with high bitrate. No crash just now, but I only tested 1min with each. Yes crashes earlier were with the Appimage.

(0005324) Andrea_Paz 2022-06-11 06:52	If you don't have space/transfer rate problems, I recommend DNxHR which always guarantees quality and efficiency during editing. You can see an overview of codecs and formats in the manual: https://cinelerra-gg.org/download/CinelerraGG_Manual/Overview_on_Formats_Codecs.html However you have too many problems with rendering, you must have something wrong with your system (maybe some missing dependencies?).

(0005323) quintao 2022-06-11 04:39	The MPEG render option doesn't export the audio. What output format would you recommend for a high quality file than can be edited later on? Where is MPEG2? I set output to MPEG-2_vaapi and the prog crashed on hitting OK to render.

(0005322) Andrea_Paz 2022-06-10 15:40	Unfortunately, I cannot reproduce the crash. I have tried h264 with 1 - 3 - 6 and 9 Mbit. @Andrew, can you you examine the dump to figure out the cause of the crash?

(0005321) PhyllisSmith 2022-06-10 15:02 (Last edited: 2022-06-11 17:29)	@quintao And today's crashes are with AppImage?? instead of Elive package? I will test with a much bigger file here and use bitrate @8Mbit -- my current "guess" is that some resource is being overrun. Without being able to create the problem here, there is no way to fix it.

(0005320) quintao 2022-06-10 06:28	More crashes today. Any bitrate @3Mbit or higher, both h264 and h265. Encoding @1Mbit succeeded. Two effects, colour balance and contrast applied throughout. Encode to MPEG @6Mbit succeeded.

(0005319) Andrea_Paz 2022-06-09 21:43	It is difficult to compile CinGG with system ffmpeg; you have to use the --without-thirdparty option and also disable dv, dpx and something else I don't remember. I succeeded only because of Andrew's help. There is a thread on the mailing list but I can't find it. Better to use CinGG's internal one which, I believe, will soon be upgrading from 4.4 to ffmpeg 5.1. It is not directly mentioned in the manual. You can see chapter 1 "installation" and the appendix "Developer Section".

(0005318) quintao 2022-06-09 01:44	Re. Ffmpeg, I really don't know which one would be used by cinelerra, the system ffmpeg or the inbuilt one. Does the manual explain about it? In any case, I just downloaded ffmpeg 5.0

(0005317) PhyllisSmith 2022-06-08 15:09	@Andrea_Paz Elive has both 64 and 32 bit. But anyway I do have a 32-bit 11 Debian partition and will see if I can install Elive Debian there and test.

(0005316) Andrea_Paz 2022-06-08 14:34	Am I wrong or is Elive 32 bit? Recently there have been problems with 32-bit Debian 10-11, although they should be fixed now. Do you use system ffmpeg instead of CinGG's internal ffmpeg? That could give problems as well.

(0005315) quintao 2022-06-08 13:19	Ok, first segfaults happened with cinelerra-gg_5.1.202103+git Updated recently to the cinelerra-gg_5.1.2022+03+git package from elive repo and still got segfaults. Format: ffmpeg mp4 h264, bitrate 4000000, 1920*1080 I have ffmpeg 4.2 built from source on my system, could that interfere?

(0005314) PhyllisSmith 2022-06-08 01:55	@quintao The one major new thing in the code starting in March was the addition of Alt-h for Help and that would take more memory but it sounds like you are watching the memory usage from your initial description.

(0005312) PhyllisSmith 2022-06-08 01:50	@quintao You are right, the indexes would have been rebuilt when you ran as root. I wish it were that simple that it is just the package install, but that would be unusual. Which version did you have installed before March and which version is now giving the render issues? AppImage?? your own build ?? which render options are you using ?? The dump is not leading me to any insight, but I can only do the simplest of dump analysis. Will continue to see if I can reproduce the crash.

(0005311) quintao 2022-06-08 01:29	When I ran as root for the debugging, surely the indexes were rebuilt? Must have been something with the package install messed up somewhere??

(0005310) quintao 2022-06-08 01:26	I think we are sorted. No segfault with the Appimage. I will always use that from now on. Was using the elive package before. Thanks for helping out.

(0005309) PhyllisSmith 2022-06-07 18:12	@quintao Have you done a "rebuild index"? This has been known to resolve this type of problem when other people (Andrea and myself) can not reproduce the error. This can be done in the Resources Window after the Media is loaded. You can rebuild all indexes in the Settings, Preferences, Interface, Index area - it will slow down as every loaded file has to rebuild indexes the first time (indexes help to play faster). I will continue to edit your test file a lot and then try rendering some more. I will also load it in an earlier version to see if the updated version is causing the problem, but I have looked at the March changes in the releasenotes.pdf and see nothing that should affect rendering at all. Or perhaps you updated from an earlier version and I should be checking for changes since version dated ???

(0005306) Andrea_Paz 2022-06-07 12:08	I tried a few renders of your example and other sources as well. All OK. I used both CinGG compiled by me and appimage. I use arch linux. Unfortunately, I cannot install devuan on vm so I could not test it. Try CinGG appimage.

(0005305) quintao 2022-06-07 08:30	It's also not immediate crash, could reach 80%, sometimes less. I keep an eye on memory, I have 16Gig. Normally it's around 11 or 12 Gig used during a render.

(0005304) quintao 2022-06-07 08:22	No batch render. No farm. Just direct from render dialog

(0005303) quintao 2022-06-07 08:20	Sample mp4 file here https://mega.nz/file/srY3VTiA#tZ4qb0kk0mbAwqE6yYiNUz6xLvf_GgJ7r5vVsDTx694

(0005302) Andrea_Paz 2022-06-07 07:03	Are you using batch render? Render farm? Are your projects all 10-bit mp4? I can try to test if you tell me the steps you followed to render.

Notes
(0005301) PhyllisSmith 2022-05-23 12:46	@deim A "guess" based on the message; "VIconThread::draw_lock, VIconThread::run" is a lock problem in the Resources window drawing the thumbnails. A suggestion, which is not a fix, just so you can continue is to turn off these thumbnails and use the words instead: - use the single character v on an empty spot in the resources window until you see no more thumbnails, just letters - also, because $HOME/.bcast5/ saves the thumbnails so it does not have to recreate them sometimes, you could temporarily use a new/empty startup via at a terminal window, keyin: CIN_CONFIG=/tmp/bcast5 bin/cin to run Cinelerra and see if the recreated thumbnails solve the problem. This does not have any of your configuration values you have set so you may want to only test.

Notes
(0005300) Andrew-R 2022-05-13 04:48	there is also 'colorspace' plugin, hopefully it also can convert part of timeline between those colorspaces

(0005294) glitterball 2022-04-23 21:03	@Andrew-R You are quite right about the colorspace filter: I had not seen the 'irange' option and had only played with 'range'. When YUV color range is set to MPEG (in Cin Preferences>Appearance), setting irange to 2 in the colorspace filter does change the contrast, however the data outside the 16-235 range is already gone (clipped). Interestingly, when YUV color range is set to JPEG, using the colorspace plugin to set the range to MPEG does not hard clip, and values outside the 16-235 range can be be brought back into range using another plugin such as Blue Banana.

(0005293) Andrew-R 2022-04-23 18:49	so, this hack shows color range for asset in it info window.. how just need to learn how to make this dropdown list, and pass user selection back to program..

(0005292) Andrew-R 2022-04-23 18:08	so, cin already carries this info in ff_color_range variable in asset.c $ grep ff_color_range cinelerra/* grep: cinelerra/armv8l: Is a directory cinelerra/asset.C: ff_color_range = -1; cinelerra/asset.C: ff_color_range = asset->ff_color_range; cinelerra/asset.C: ff_color_range = GET_DEFAULT("FF_COLOR_RANGE", ff_color_range); cinelerra/asset.C: UPDATE_DEFAULT("FF_COLOR_RANGE", ff_color_range); cinelerra/asset.C: fprintf(fp," ff_color_range=%d\n", ff_color_range); cinelerra/asset.h: int ff_color_space, ff_color_range; grep: cinelerra/ci: binary file matches grep: cinelerra/data: Is a directory cinelerra/ffmpeg.C: if( (vid->color_range = asset->ff_color_range) < 0 ) cinelerra/ffmpeg.C:int FFMPEG::ff_color_range(int stream) cinelerra/ffmpeg.h: int ff_color_range(int stream); cinelerra/fileffmpeg.C: if( asset->ff_color_range < 0 ) cinelerra/fileffmpeg.C: asset->ff_color_range = ff->ff_color_range(0); grep: cinelerra/randrik: Is a directory grep: cinelerra/tests: Is a directory we just need to expose it as changeable param in assetedit.c\|h or so I think...

(0005291) Andrew-R 2022-04-23 17:44	@glitterball this filter have two ranges - range for output and irange for input.. setting range to mpeg while irange at jpeg (with jpeg in prefs) definitely changes output picture...

(0005290) glitterball 2022-04-23 16:46	@Andrew-R Yeah, I tried that filter, and changing the range there doesn't seem to make any difference either.

(0005289) Andrew-R 2022-04-23 16:31	@glitterball have you tried ffmpeg plugin I mentioned (with cin preference set to full)?

(0005288) glitterball 2022-04-23 16:27	@Andrew-R I tried changing the x11 settings as suggested and none of the changes made any difference. Rendering a video produces a file where the those values are clipped. The only way to keep the full range is to set the YUV color range to 'JPEG' in appearance. However this means that all clips must be transcoded to the same range before being imported into Cinelerra, as Cinelerra cannot handle video files with differing ranges within the same project.

(0005287) Andrew-R 2022-04-22 14:57	@glitterball, as far as I can see this preferences setting control ffmpeg's swscaler, before it converted to rgba-float. May be x11 output clip them? try x11-opengl, try uncheck direct x11 path, also try to render into some image/video file and see if those values still around.... if they are - then we need to look at video output specifically...

(0005286) glitterball 2022-04-22 14:39	@Andrew-R Sorry I wasn't clearer: it makes no visible difference to the clip (I.e. everything outside 16-35 is still clipped even with those settings). I do get some extra error messages when adding the .opts file mostly "[swscaler @ 0x7efd9c20d900] deprecated pixel format used, make sure you did set range correctly"

(0005285) Andrew-R 2022-04-22 12:39	@glitterball $ cat ~/matrixbench_highdivx_ac3.opts #lowres=1 #fast=1 #video_filter=format=yuv422p video_filter=scale=in_range=limited:out_range=full sorry, try optS? in my case it prints something: [swscaler @ 0xe0d24000] No accelerated colorspace conversion found from yuv420p to rgba.

(0005284) glitterball 2022-04-22 12:17	@Andrew-R Thanks for the response, I gave your suggestion a go, but unfortunately adding the .opt file didn't appear to make any difference.

(0005283) Andrew-R 2022-04-22 09:48	@glitterball, yeah, from user perspective this is good idea, nit sure if I am up to task about coding it (soecifically, not sure if ffmpeg>cin transfer has access to specifics of assets/clips) I found another answer, you can try to add this into opt files, one per asset/file https://superuser.com/questions/1405177/ffmpeg-how-to-handle-color-range-conversions "Input is limited range, I want to output full range Use the scale filter, -vf scale=in_range=limited:out_range=full" try to put this line (without leading "-") in file named exactly like your input file just with opt extension...? sure, this is workaround (if it works as I think it works!) and not resolution..

(0005282) glitterball 2022-04-22 08:31	The way Davinci Resolve handles this, is to allow the user to set whether a clip's data levels are 16-235 or 0-255 under 'Clip Attributes'. Specifically there is a radio button under 'Clip Attributes' (right-click on a clip, then Clip Attributes) called 'Data Levels' with the following options: Auto, Video or Full. In Cinelerra, the same options could be provided in the 'Info...' dialog box (found by right-clicking on a file in Resources>Media). Thinking about it, this is the logical solution, as many video projects will use video files with differing frame rates, data levels etc, so it should be possible to set this on a per-file basis rather than project-wide. I also note that the 'auto' option for data levels did not work in Davinci Resolve, and I had to set this manually.

(0005246) Andrew-R 2022-01-29 04:48	also, may be ffmpeg's filter colorspace can help here? seems to be enabled in Cingg and suppirts >8 bit depth https://github.com/FFmpeg/FFmpeg/blob/master/libavfilter/vf_colorspace.c

(0005244) Andrew-R 2022-01-29 04:11	does same problem exist in ffmpeg? I think until recently ffmpeg was internally 8bit, so clipping may happen there...

Notes
(0005296) rs68 2022-04-25 03:42	Hi @fary54, Yes, linux is really cool. I hope too. This tool should be helpful for the selection step, while editing. Have a nice day.

(0005295) fary54 2022-04-24 09:57	Hi @rs68, Good suggestions Personally, I only use linux. I'm also a french cin-enthusiast. Making long films (more than one hour), I find myself managing a large number of rushes (more than 500). Your suggestions would give me a quick visual feedback of the situation. I hope that these won't require too much work because it would be nice to have them.

(0005265) rs68 2022-04-19 06:52	@PhyllisSmith Hi Phyllis, I already know the 'info' feature, but what I'm looking for is some kind of feature people could enable/disable on right clicking on a media icon with something like a submenu 'overlays' > 'occurrences', 'duration' in the context menu. That way, each one could choose to show up overlayed informations over the media icon... or not (at their choice). All in the pop up menu, I think this could work well. @Pierre Bonjour Pierre, I also know the inter-view mode, but this not what I'm looking for. When you have about 200 rushes to examine before editing, it is much much better to have all occurrences at a glance into the media list. @Andrea_Paz, Hi Andrea, I didn't know this feature. By now, I've made my own filterings and it works great. Thanks a lot. Thank you all for your effort, guys, great work! Cheers.

(0005264) Andrea_Paz 2022-04-18 20:45	Great ideas the first two points; I hope someone knows how to implement them. The third point, if I understand correctly, should already be available in CinGG: https://www.youtube.com/watch?v=_diECpahbq0&list=PLOnHGAy9DoU1mBA55D9SoQOb-V721345u For an example (in Spanish) on how to filter media in the various "user bins" see: https://www.youtube.com/watch?v=hjrSTjl1-M0&list=PLOnHGAy9DoU1Xaj9HVhBwyov-daKFZmIE&index=21 In the CinGG manual, these parts can be found here: https://cinelerra-gg.org/download/CinelerraGG_Manual/User_Folders_Bins.html

(0005263) PhyllisSmith 2022-04-18 19:08 (Last edited: 2022-04-18 21:40)	@Pierre -- oh, that is much better for the first one -- thanks!

(0005262) Pierre 2022-04-18 18:50	Bonjour et bienvenue (je suis québécois), I think you will find even better than what you ask for to identify the use already made of a media file in an EDL. Just click with the middle buttons of your mouse on the different media in sources and a window will appear with a detailed representation of the use already made. See section 5.6 of the manual concerning the "Inter-View" mode. https://cinelerra-gg.org/download/CinelerraGG_Manual.pdf

(0005261) PhyllisSmith 2022-04-18 18:29	Good suggestions. For # 2. -- media icons ... media's duration in HH:MM:SS format -- you can get this information by right mouse clicking on the highlighted media and then left mouse click on "info" and then "details". Although is would be very convenient to see this without taking any action, the Resources Window is not that big and some people have a lot of media there so I do not think just putting it there in the corner would look very nice. Maybe it could be added the the "comment" portion when in "list" mode instead of "icon" mode.

Notes
(0005260) Andrew-R 2022-04-11 09:13	here is my test hdd image for qemu-system-ppc (1.6 gb) https://cloud.mail.ru/public/3FFj/p2Zn2DhGt sha256sum 5gb_ppc_debian8.qcow.xz 90af1c6479375fcdc3da8962a507d170283d9ad86cf0cf9d47ab27b3ad716c81 5gb_ppc_debian8.qcow.xz

(0005259) Andrew-R 2022-04-07 09:43	ppc hacks relative to git state just before giflib disabling went in

(0005258) Andrew-R 2022-04-07 09:35	configure switches I used

Notes
(0005249) PhyllisSmith 2022-02-10 19:30	Just some notes. After experimenting with different AppImages created on different computers, different Operating Systems with different graphics cards and different vaapi/vdpau libraries installed, or none of these libraries installed, I think I can safely conclude the following: 1) Hardware acceleration, vaapi or vdpau, works the same running either the created binary or the created AppImage from that binary on that same computer. 2) Hardware acceleration using an AppImage created on another computer running a different Operating System or even just a different level of Operating System, most likely will not work unless the user's computer just happens to have the same characteristics with respect to vaapi/vdpau as the computer where the AppImage was created. ************************************* Some experimental results **************************** For example using Test1 (hockey.ts television recording) and Test2 (Life Untouched 4K HDR10 Demo.mp4 downloadable file): CinGG-20220131-x86_64.AppImage -on AMD laptop with Radeon RX-580 graphics card where AppImage was created (Fedora 32): Test1 hardware acceleration vaapi and vdpau both work as expected. Test2 hardware acceleration vaapi works, but vdpau does not - this is as expected. CinGG-20220131-x86_64-older_distros.AppImage - on AMD laptop with Radeon RX-580 using AppImage created on AMD Opteron Ubuntu 16.04: Test1 and Test2 hardware acceleration do not work due to missing Radeon so file. CinGG-20220131-x86-64-older_distros-multibit.AppImage - on AMD laptop with Radeon RX-580 using AppImage created on Intel 4-core with Broadwell graphics: Test1 hardware acceleration only vdpau works and that could be due to a "wrapper". Test2 hardware acceleration does not work. CinGG-20220131-x86_64.AppImage (same results as if built on this computer/this O/S) - on Intel 4-core laptop with Broadwell using AppImage created on AMD laptop with different O/S of Fedora 32: Test1 hardware acceleration only vaapi works, but vdpau does not. Test2 hardware acceleration does not work.. CinGG-20220131-x86_64-older_distros.AppImage - on Intel 4-core laptop with Broadwell using AppImage created on AMD Opteron Ubuntu 16.04: Test1 and Test2 hardware acceleration do not work. CinGG-20220131-x86-64-older_distros-multibit.AppImage - on Intel 4-core laptop with Broadwell using AppImage created on same computer but older O/S of Fedora 29: Test1 and Test2 hardware acceleration do not work.

Notes
(0005248) Andrew-R 2022-01-30 15:08	sorry, no idea why both prores encoders lose some info :(

(0005247) PhyllisSmith 2022-01-29 22:56 (Last edited: 2022-01-29 23:08)	@Andrew_R According to the forum post, he has tried all options of both ProRes and ProRes_ks. But as far as I can tell he did not add any other parameters that might solve the problem. Can you recommend any? I think we are missing something that is necessary and it needs to be fixed for the good of CinGG. He started rendering using DNxHR HQX which keeps the highlight detail as a workaround.

(0005245) Andrew-R 2022-01-29 04:17	I think there are two prores enciders in ffmpeg/libavcodec, what one was used here?

Notes
(0005241) cine22linx 2022-01-22 00:48	after logging out, tried again and the error returned, [my previous 'success' was using a 'change passw' link rhat was in my mail] 'i'm in no need of using the forum now, so i'm stopping, at least for a good while, checking here, hope gets solved, bye. '

(0005240) cine22linx 2022-01-22 00:37	not Allowing javascript, now i can enter, and changed the passw, not many config/preferences... anyway, please do notdelete my account [the Forum one] lol Thx

(0005238) Andrew-R 2022-01-21 15:45	@administrator quick search for this error lead to https://wordpress.org/support/topic/error-2252-please-contact-to-forum-admin/ might be something with URLs as seen by software (if our forum really uses Wordpress)

(0005237) IgorBeg 2022-01-21 08:46	@cine22linx I have some problems with forum. To login and see the recent posts sometimes I switch to a different language from English to Deutsch, Francais, Espanol, Italiano. I know it is not a good way but may work. If you have a bit of time, can you try?

(0005236) cine22linx 2022-01-21 00:33	tryed again. same; since i cant see the config/preferences for my account, please delete it [the Forum one] Thx

(0005233) cine22linx 2022-01-20 00:55	sorry, i entered the sent by you passw, and 'Error 2252 \| Please contact the forum administrator'. 'i'm in no need of using the forum now, so i'm stopping, at least for a good while, checking here, hope gets solved, bye. '

(0005232) administrator 2022-01-19 12:26	I have set you up for the forum and forwarded you an email. Hope you can log in now without any problems.

(0005231) cine22linx 2022-01-18 23:27	Yes, i won't use it right now, but please add me manually, i confirm that i agree with the terms of use of this website and the forum. Thx

(0005230) administrator 2022-01-18 05:28	Sorry for the inconvenience. Sometimes the protection measures against spammers also take effect with real users, so sometimes the website blocks the registration for the forum. This often affects email addresses of well-known email providers. I can add you to the forum manually if you want, then you can use the forum from now on. You just have to confirm that you agree with the terms of use of this website and the forum. Cheers Sam

(0005229) cine22linx 2022-01-18 00:48	FFox. sorry i'm stopping, at least for a good while, checking here, hope gets solved, bye.

(0005227) IgorBeg 2022-01-17 08:55	What browser are you using?

(0005226) PhyllisSmith 2022-01-17 00:35	@cine22linx OK, will leave open -- in case someone comes along who knows what the problem is.

(0005225) cine22linx 2022-01-17 00:09	i'm in no need of using the forum now, but the bug is still there, so it may be better to leave it open, so someone that knows what's Error 2252 could see the post and offer advice, others may fail to register and just give up... [for 'administrator' reference, i did same user name and email as here].

(0005220) cine22linx 2022-01-16 01:41	same error message. thx, let's leave it at that, i have no urgent need to use the forum.

(0005218) PhyllisSmith 2022-01-15 05:23 (Last edited: 2022-01-15 17:37)	@cine22linx I think this error message has something to do with WordPress, but am not sure as there are various reasons provided on the internet search. Tested with Username of "testing22" and it added the account with no error message and then I deleted it with no problem. It looks like there is nothing unusual about the Username you used or the steps you followed. If you have not already tried again today, could you do so because it just worked for me? It is OK to put problems you need help with in the Bug Tracker if the forum is not responding.

Notes
(0005217) PhyllisSmith 2022-01-12 14:31	For future reference -- Phyllis can test on the computer "Storm" (.17, 12 CPU, AMD) as it is non-ssse3 and has the problem).

(0005216) Andrew-R 2022-01-11 18:06	also see 0000596

(0005215) Andrew-R 2022-01-11 18:05	https://github.com/vanakala/cinelerra-cve/commit/88ea4f6428e046d4340d1bfa0d6da96a1f05cbe7 I think cingg added similar modification earlier, but may be we missed few places where it matters?

(0004972) Andrew-R 2021-09-19 02:49	this is exactly my problem - I do not understand why I suddently got this result! May be one thread feed uninitialized value down to this function? or some weird asm-level interaction...

(0004971) PhyllisSmith 2021-09-19 01:27	@Andrew-R This is too hard for me, but it is so strange that hscale is 1.000000 before and after the -nan (not a number, but we know it is negative!). If it is going outside the address space (address sanitizer), how can it come back so easily?

(0004968) Andrew-R 2021-09-17 00:51	checking for nan for hscale eliminates crash, but then processing (for example via fade automation) completely broken :/ cat cdrom_root/tmp/guicast-2.diff diff --git a/cinelerra-5.1/guicast/bcxfer.C b/cinelerra-5.1/guicast/bcxfer.C index 8eb6533b..da437114 100644 --- a/cinelerra-5.1/guicast/bcxfer.C +++ b/cinelerra-5.1/guicast/bcxfer.C @@ -1,3 +1,4 @@ +#include <cmath> void BC_Xfer::init( uint8_t *output_rows, int out_colormodel, int out_x, int out_y, int out_w, int out_h, @@ -101,11 +102,19 @@ void BC_Xfer::init( this->total_out_w = out_rowspan; this->in_pixelsize = BC_CModels::calculate_pixelsize(in_colormodel); this->out_pixelsize = BC_CModels::calculate_pixelsize(out_colormodel); + + if( in_w ) { this->scale = (out_w != in_w) \|\| (in_x != 0); / + 1 so we don't overflow when calculating in advance / column_table = new int[out_w+1]; double hscale = (double)in_w/out_w; + if (isnan(hscale)) + hscale = (double)in_w/out_w; + //printf("hscale = %f \n", hscale); + //printf("in_w: %i \n", in_w); + //printf("out_w: %i \n", out_w); + //printf("in_x: %i \n", in_x); for( int i=0; i<out_w; ++i ) { column_table[i] = ((int)(hscale i) + in_x) * in_pixelsize; if( in_colormodel == BC_YUV422 \|\| in_colormodel == BC_UVY422 ) diff --git a/cinelerra-5.1/guicast/xfer/Makefile b/cinelerra-5.1/guicast/xfer/Makefile index ff12aa48..39557ac7 100644 --- a/cinelerra-5.1/guicast/xfer/Makefile +++ b/cinelerra-5.1/guicast/xfer/Makefile @@ -11,7 +11,7 @@ all: $(OUTPUT) clean: rm -rf $(OBJDIR) xfer.[Ch] -BFLAGS:=-Ofast +BFLAGS:=-O0 -g -fsanitize=address -fsanitize=undefined $(OBJDIR)/%.o: %.C $(CXX) -I.. `cat $(OBJDIR)/c_flags` $(BFLAGS) -DMSGQUAL=$ -c $< -o $@ root@slax:/mnt/sdb1/admin#

(0004967) Andrew-R 2021-09-16 23:01	so I added debug like this cat cdrom_root/tmp/guicast.diff diff --git a/cinelerra-5.1/guicast/bcxfer.C b/cinelerra-5.1/guicast/bcxfer.C index 8eb6533b..d0a8aafe 100644 --- a/cinelerra-5.1/guicast/bcxfer.C +++ b/cinelerra-5.1/guicast/bcxfer.C @@ -106,8 +106,12 @@ void BC_Xfer::init( /* + 1 so we don't overflow when calculating in advance / column_table = new int[out_w+1]; double hscale = (double)in_w/out_w; + printf("hscale = %f \n", hscale); + printf("in_w: %i \n", in_w); + printf("out_w: %i \n", out_w); + printf("in_x: %i \n", in_x); for( int i=0; i<out_w; ++i ) { - column_table[i] = ((int)(hscale i) + in_x) * in_pixelsize; + column_table[i] = (int)(((hscale * i) + in_x) * in_pixelsize); if( in_colormodel == BC_YUV422 \|\| in_colormodel == BC_UVY422 ) column_table[i] &= 0000003:0000003; } diff --git a/cinelerra-5.1/guicast/xfer/Makefile b/cinelerra-5.1/guicast/xfer/Makefile index ff12aa48..efb2e628 100644 --- a/cinelerra-5.1/guicast/xfer/Makefile +++ b/cinelerra-5.1/guicast/xfer/Makefile @@ -11,7 +11,7 @@ all: $(OUTPUT) clean: rm -rf $(OBJDIR) xfer.[Ch] -BFLAGS:=-Ofast +BFLAGS:=-Ofast -fsanitize=address -fsanitize=undefined $(OBJDIR)/%.o: %.C $(CXX) -I.. `cat $(OBJDIR)/c_flags` $(BFLAGS) -DMSGQUAL=$ -c $< -o $@ -- and this resulted in debug output file like this ..strangely, at one time before crash hscale was '-nan' ?!

(0004966) Andrew-R 2021-09-16 21:57	so if I change line 110 in http://git.cinelerra-gg.org/git/?p=goodguy/cinelerra.git;a=blob;f=cinelerra-5.1/guicast/bcxfer.C;h=8eb6533baa68b7cc8b3fe1d763f03d65a3d28b4e;hb=HEAD column_table[i] = ((int)(hscale * i) + in_x) * in_pixelsize; to column_table[i] = ((long long)(hscale * i) + in_x) * in_pixelsize; it stopped crashing, yet miniatures on timeline are garbage :(

(0004965) Andrew-R 2021-09-16 21:22	So, I compiled Cin with address sanitizer and result was: cat cin_addr_sanitizer.log cin Cinelerra Infinity - built: Sep 16 2021 21:38:07 git://git.cinelerra-gg.org/goodguy/cinelerra.git (c) 2006-2019 Heroine Virtual Ltd. by Adam Williams 2007-2020 mods for Cinelerra-GG by W.P.Morrow aka goodguy Cinelerra is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. There is absolutely no warranty for Cinelerra. init plugin index: /usr/lib/cin/plugins init lv2 index: /usr/share/cin/lv2 init ladspa index: /usr/lib/cin/ladspa ../bcxfer.C:110:49: runtime error: signed integer overflow: -2147483648 * 3 cannot be represented in type 'int' ASAN:SIGSEGV ================================================================= ==2556==ERROR: AddressSanitizer: SEGV on unknown address 0x26427e00 (pc 0x08a18b4e bp 0xa8efd728 sp 0xa8efd6d0 T71) #0 0x8a18b4d in BC_Xfer::xfer_rgb888_to_bgr8888(unsigned int, unsigned int) (/usr/bin/cin+0x8a18b4d) 0000001 0x88e47fa in BC_Xfer::xfer_slices(int) (/usr/bin/cin+0x88e47fa) 0000002 0x88e725d in BC_CModels::transfer(unsigned char, unsigned char, unsigned char, unsigned char, unsigned char, unsigned char, unsigned char, unsigned char, int, int, int, int, int, int, int, int, int, int, int, int, int) (/usr/bin/cin+0x88e725d) 0000003 0x882b788 in BC_Bitmap::read_frame(VFrame, int, int, int, int, int, int, int, int, int) (/usr/bin/cin+0x882b788) #4 0x88add6d in BC_WindowBase::draw_vframe(VFrame, int, int, int, int, int, int, int, int, BC_Pixmap) (/usr/bin/cin+0x88add6d) 0000005 0x885fd33 in BC_Pixmap::draw_vframe(VFrame, int, int, int, int, int, int) (/usr/bin/cin+0x885fd33) 0000006 0x84ab6f0 in AssetPicon::create_objects() (/usr/bin/cin+0x84ab6f0) 0000007 0x84afaf1 in AWindowGUI::update_asset_list() (/usr/bin/cin+0x84afaf1) 0000008 0x84b00b5 in AWindowGUI::update_assets() (/usr/bin/cin+0x84b00b5) #9 0x84b0417 in AWindowGUI::receive_custom_xatoms(XClientMessageEvent) (/usr/bin/cin+0x84b0417) 0000010 0x88a5c39 in BC_WindowBase::dispatch_event() (/usr/bin/cin+0x88a5c39) 0000011 0x88a63c7 in BC_WindowBase::run_window() (/usr/bin/cin+0x88a63c7) #12 0x84b2544 in AWindow::run() (/usr/bin/cin+0x84b2544) #13 0x88bd22e in Thread::entrypoint(void) (/usr/bin/cin+0x88bd22e) 0000014 0xb7a330c4 (/usr/lib/libasan.so.2+0xb10c4) 0000015 0xb79b5362 (/usr/lib/libasan.so.2+0x33362) #16 0xb77ef6f1 in start_thread (/lib/libpthread.so.0+0x66f1) #17 0xb684923d in __clone (/lib/libc.so.6+0xfa23d) AddressSanitizer can not provide additional info. SUMMARY: AddressSanitizer: SEGV ??:0 BC_Xfer::xfer_rgb888_to_bgr8888(unsigned int, unsigned int) Thread T71 created by T0 here: #0 0xb79b5171 in pthread_create (/usr/lib/libasan.so.2+0x33171) 0000001 0x88bd51f in Thread::start() (/usr/bin/cin+0x88bd51f) 0000002 0x86630c7 in MWindow::start() (/usr/bin/cin+0x86630c7) 0000003 0x8349430 in main (/usr/bin/cin+0x8349430) #4 0xb6767696 in __libc_start_main (/lib/libc.so.6+0x18696) ==2556==ABORTING bash-4.2$ this is even without loading any avi, just jpeg from /usr/share/wallpapers. hm.

(0004959) Andrew-R 2021-09-13 16:05	@PhyllisSmith, you can try and install qemu (user-space part of kvm-based virtualization, but also usable as independent emulator) on Fedora fom repository? https://docs.fedoraproject.org/en-US/quick-docs/qemu/ $ su -c "dnf install qemu" and then try line like qemu-system-i386 -cdrom /path_to_iso/CDROM-4.14.246-big-7.iso -enable-kvm -m 512 -cpu pentium3,+ssse3 yes, I just tested, it works with this fictional cpu :} you can add '?' as parameter to '- cpu' and qemu will give you big list of possible cpus... If you omit 'enable-kvm' it will boot in pure emulation mode (even if kvm/virtualization tot enabled ...something you often need to check in BIOS setup ...) - quite slow (talk about minutes until even text login)!

(0004958) PhyllisSmith 2021-09-13 15:55	@Andrew-R I have downloaded the iso but not sure how I can even try it. Also downloaded the Space Cowboys and of course, had to play that on my laptop -- that was and still is a good movie. In looking at the dmp file, I have no clue and am not adept at reading much there -- it just looks like it crashes on the "Load" just like you stated. Although it sort of looks like it got started because the size is 3,559,424 but the position is 1,386,508 which I am only guessing means that it got to that point before crashing.

(0004957) Andrew-R 2021-09-13 14:57	here is 'dmp' log from VM ....

Notes
(0005214) PhyllisSmith 2022-01-08 18:17	@glitterball Thanks for the followup (and it looks like "Add note" worked OK). If this can not be fixed before the next end of month release, I will back out the fade in/fade out mod.

(0005213) glitterball 2022-01-08 10:42	@Andrea_Paz I started CIN from the terminal, and there are no errors. Another observation is that if I edit the Keyframe parameter ALPHA value, this works as expected in the CinGG-20210930-x86_64.AppImage version (i.e. the opacity changes). In the newer versions, values between 1 and 255 are fully opaque, and 0 is transparent. In other words, ALPHA has become a binary on/off option.

(0005212) IgorBeg 2022-01-08 10:12	Unfortunately I confirm the issue as written in @glitterball's Description (but for older_distros, for me because my OS is UbuntuStudio_16_04). To test, open Cinelerra-GG and put the Title effect on the video track without any Video/image on it or other video track below it. When it works: 1. if there is a video/image above the Title effect in the same video track (as the Phyllis' screencast show); 2. if the Title effect is on a video track alone and there is an other video track below with a video/image in it. It seems an alpha problem due to https://www.cinelerra-gg.org/bugtracker/view.php?id=559

(0005211) PhyllisSmith 2022-01-07 23:20	@glitterball @IgorBeg There must be something that I am not doing to create the problem. Here is a demo of Titler "not" anti-aliasing and then with it both with fade in/out. https://streamable.com/n6mgvw

(0005210) PhyllisSmith 2022-01-07 20:24 (Last edited: 2022-01-07 20:29)	@glitterball (So I added 2 notes here as an ordinary user and had no problems) Not sure why you could not "Add Note" but in the past 2 issues have been reported: 1) If you type something in and there is any kind of wait or interruption, you can lose what you typed in. This used to happen to me as I get interrupted or take too long so what I learned to do is if I have a lot to type in or have to look something up to finish, I would put what I wrote so far into a file so as not to lose it. Or I would just but it in the buffer using Ctrl_c for a backup and then Ctrl_v to get it back. 2) This has been noted but I do not think this was the case for you: "Sometimes there is a problem creating a new report issue in the website’s Mantis Bugtracker using chrome and you lose what you just typed in. Gen-erally when logging into the bugtracker, the option "only allow with this IP address" needs to be disabled, then the bug tracker will work ﬁne." Did you get any error message or did the screen just come back? Anyway, I suspect you were going to "add note" saying that the anti-alising problem does not occur until you add "fade in and/or fade out".

(0005208) PhyllisSmith 2022-01-06 17:25	@glitterball Probably the Freelancer "fix" added in 10/31 appimage impacted this. I will try to duplicate yours and Andrea's tests to verify and will back out the mod accordingly. I know that his fix did impact the PatchBay Overlay Modes and was waiting for the Freelancer to address that but he got Covid and has not been able to do anymore work. Luckily we now have the previous AppImages to use as a backout.

(0005207) Andrea_Paz 2022-01-06 15:26 (Last edited: 2022-01-06 15:27)	For me, with the latest multibit appimage (2021-12-31), both antialiasing and Fade In/Out works. Antialiasing has never been perfect, but the difference is visible. Can you try starting from the terminal to see if there are any error messages?

Notes
(0005206) Andrew-R 2021-12-26 23:44	note, ffmpeg 4.3+ probably can simulate mjpeg cam via v4l2 loopback device: https://forums.developer.nvidia.com/t/solved-gstreamer-help-mpeg2-source-file-to-mjpeg-at-v4l2-loopback-device/178853

(0004981) Andrew-R 2021-09-21 04:12	I tried to add mjpeg decoding, but it still reads (?) from wrong buffer? WARNING - this is just debug sketch - there is no deleting/stopping mjpeg engine ...it hangs eventually ....

(0004980) Andrew-R 2021-09-21 02:03	also, without some check in FFVideoConvert::convert_pixfmt attempt to use ffmpeg encoder with compressed (jpeg) v4l2 resulted in SIGFPE, this patch has unrelated lines, main one is + if(BC_CModels::components(cmodel)) (so we not divide by zero here) Sadly, this does not result in picture_with_sound recording, screen still empty/black (not green). Output with patch: ----[snip]--- BC_CModels::transfer failed:208 1(1280x720) to 9(1280x720) cmodel from convert_pixfmts 1 max_bits 1: 8 BC_components: 0 max_bits 8 imodel: -1 imodel_is_yuv: 0 cmodel_is_yuv 0 BC_CModels::transfer failed:208 1(1280x720) to 9(1280x720) cmodel from convert_pixfmts 1 max_bits 1: 8 BC_components: 0 max_bits 8 imodel: -1 imodel_is_yuv: 0 cmodel_is_yuv 0 ---[snip]--- so, apparently encoding from compressed source still require more work, but at least with this check Cin does not crash .... diff --git a/cinelerra-5.1/cinelerra/ffmpeg.C b/cinelerra-5.1/cinelerra/ffmpeg.C index b05a455d..a2d59244 100644 --- a/cinelerra-5.1/cinelerra/ffmpeg.C +++ b/cinelerra-5.1/cinelerra/ffmpeg.C @@ -1304,10 +1304,19 @@ int FFVideoStream::video_seek(int64_t pos) if( pos == curr_pos-1 && !seeked ) return 0; // if close enough, just read up to current int gop = avctx->gop_size; - if( gop < 4 ) gop = 4; + if( gop < 4 && gop != 0) gop = 4; if( gop > 64 ) gop = 64; + if( gop == 0 ) gop = 1; + //fprintf(stderr, "GOP: %i \n", gop); int read_limit = curr_pos + 3gop; if( pos >= curr_pos && pos <= read_limit ) return 0; + // fixup for I-only codecs, not sure why it needed with ffmpeg 4.2 + if (st->codecpar->codec_id == AV_CODEC_ID_DVVIDEO \|\| st->codecpar->codec_id == AV_CODEC_ID_MJPEG \|\| + st->codecpar->codec_id == AV_CODEC_ID_FFVHUFF \|\| st->codecpar->codec_id == AV_CODEC_ID_DNXHD \|\| + st->codecpar->codec_id ==AV_CODEC_ID_HUFFYUV \|\| st->codecpar->codec_id == AV_CODEC_ID_PRORES) { + gop = 1; + if( seek(pos - 2gop, frame_rate) < 0 ) return -1; + } // guarentee preload more than 2gop frames if( seek(pos - 3gop, frame_rate) < 0 ) return -1; return 1; @@ -1755,12 +1764,20 @@ int FFVideoConvert::convert_pixfmt(VFrame frame, AVFrame op) if( !convert_vframe_picture(frame, op) ) return 1; // use indirect transfer int cmodel = frame->get_color_model(); + printf("cmodel from convert_pixfmts %i \n", cmodel); int max_bits = BC_CModels::calculate_pixelsize(cmodel) * 8; + printf("max_bits 1: %i \n", max_bits); + printf("BC_components: %i \n", BC_CModels::components(cmodel)); + if(BC_CModels::components(cmodel)) max_bits /= BC_CModels::components(cmodel); + printf("max_bits %i \n", max_bits); AVPixelFormat ofmt = (AVPixelFormat)op->format; int imodel = pix_fmt_to_color_model(ofmt); + printf("imodel: %i \n", imodel); int imodel_is_yuv = BC_CModels::is_yuv(imodel); + printf("imodel_is_yuv: %i \n", imodel_is_yuv); int cmodel_is_yuv = BC_CModels::is_yuv(cmodel); + printf("cmodel_is_yuv %i \n", cmodel_is_yuv); if( imodel < 0 \|\| imodel_is_yuv != cmodel_is_yuv ) { imodel = cmodel_is_yuv ? (BC_CModels::has_alpha(cmodel) ? @@ -2717,6 +2734,7 @@ int FFMPEG::init_encoder(const char *filename) ret = get_file_format(); if( ret > 0 ) { eprintf(_("bad file format: %s\n"), filename); + eprintf( "ret, %i \n", ret); return 1; } if( ret < 0 ) {

(0004976) Andrew-R 2021-09-21 01:05	v4l2 info: v4l2-ctl --all Driver Info: Driver name : uvcvideo Card type : Android: UVC Camera Bus info : usb-0000:00:13.2-3 Driver version : 5.6.14 Capabilities : 0x84a00001 Video Capture Metadata Capture Streaming Extended Pix Format Device Capabilities Device Caps : 0x04200001 Video Capture Streaming Extended Pix Format Priority: 2 Video input : 0 (Camera 1: ok) Format Video Capture: Width/Height : 1280/720 Pixel Format : 'MJPG' (Motion-JPEG) Field : None Bytes per Line : 0 Size Image : 1843200 Colorspace : sRGB Transfer Function : Default (maps to sRGB) YCbCr/HSV Encoding: Default (maps to ITU-R 601) Quantization : Default (maps to Full Range) Flags : Crop Capability Video Capture: Bounds : Left 0, Top 0, Width 1280, Height 720 Default : Left 0, Top 0, Width 1280, Height 720 Pixel Aspect: 1/1 Selection Video Capture: crop_default, Left 0, Top 0, Width 1280, Height 720, Flags: Selection Video Capture: crop_bounds, Left 0, Top 0, Width 1280, Height 720, Flags: Streaming Parameters Video Capture: Capabilities : timeperframe Frames per second: 30.000 (30/1) Read buffers : 0

(0004975) PhyllisSmith 2021-09-21 00:49	@Andrew-R Unfortunately, I did not get time to look at this today but did download the VIDEODEVICE_FIX.diff file and will try again tomorrow.

(0004974) Andrew-R 2021-09-21 00:29	so, something like this fix allowed me to record jpeg_list out of mjpeg webcam ..... Warning, also designed to affect raw_DV (from DV input), untested due to lack of device ...

(0004973) Andrew-R 2021-09-20 23:13	It seems cinelerra/recordmonitor.C implements its own decoding for mjpeg/dv (see for example RecVideoMJPGThread::render_frame) so probably RecordVideo::decompress_buffer(VFrame *frame) from cinelerra/recordvideo.C need to grow some of this decoding too ... but then it should somewhat replace original buffer (vframe) with decompressed copy? Because apparently transfer from/to BC_COMPRESSED not implemented in guicast/ code ? I was able to get mjpeg sequence out of device by commenting out 'asset->format = FILE_FFMPEG;' in void VideoDevice::fix_asset in cinelerra/videodevice.C - this allowed selection of jpeg/jpeg sequence but sadly it makes nearly all other choices crashy :( so may be special-case it some more, so only jpeg sequence will be allowed (no audio but slightly better than nothing)....