Thanks, just wanted to double check

@Harley Acheson (harley) I see that current design is a bit confusing, do you think this should be confirmed?

Sorry, missed that file is included here. I can confirm that this file uses huge amount of memory, so I will close this report. @Alaska (Alaska) provided good analysis here.

I was able to reproduce with 2.93.2, no issue in master.

First I thought this is trolling, then I laughed, finally watching video without sound, I guess this is real issue.

I have confirmed this, because we have function BLI_filename_make_safe so it can be used perhaps with RNA set function. Path would be sanitized if you would enter it through file browser. In this case, path is also ammended, but there is report for that too - T88193

I think this is same issue as T69762 as far as shading glitches goes, mesh has some faces with degenerate geometry.

Can you upload sample .blend file with image so we can reproduce here?

I watched my GPU utlization and it never came above 2%.

This was intentional change in 3.0, see rB3b6ee8cee70. You should see message that says, you may expect loss of data due to opening newer file. It is specifically due to vertex groups refactor.

This will probably be issue in more places than render output.

Like this?

@jordi torres (jordi_torres) does Ctrl+Shift+Alt+A work without factory settings?
Do you use default keymap? If so is it modified by you? What @Eugene Du (APEC) suggested is good idea to check, but I would expect you to notice that it does something different.

In T90357#1201553, @michael campbell (3di) wrote:

Hi @Richard Antalik (ISS), oh, I must have forgot to set it back to the full sequence when I swapped the images before uploading (so they both used the same image). Making it a full sequuence or movie will still not update, but I suppose your first point of them not being able to run simultaneously is the issue. I don't suppose sinking in the timeline will help if the VSE having strips stops redraw from working in the compositor.

I missed the file, so I was trying to reproduce with default cube.

I think it would be nicer design, if there would be only sequence_editor.display_stack() function and you can provide either sequence or None. So you can do sequence_editor.display_stack(sequence_editor.meta_stack[-1]) to go up one level for example.

There are 2 issues, but no bug as far as I can see:

• VSE is on "same level" as compositor, so if both have content, you have to choose one or another in Post Processing panel in Output Properties. See https://docs.blender.org/manual/en/dev/render/output/properties/post_processing.html#post-processing
• Image input node is set to 1 frame, so it doesn't play whole sequence

I can't reproduce this issue.

I am not sure if this is bug, because this seems to be caused by overlapping faces between original geometry and hull. So you can add transform node with scale of 0.999999 to resolve this overlap.

All these issues are resolved, so closing the report

Thanks for nice list.

Can't reproduce this issue. I assigned Ctrl+Shift+Alt+A and it works.

This will be fixed in 2.93.2 which is release candidate now, but you can download it from https://builder.blender.org/download/daily/

Hmmm not sure if I would be happy with transform widget that UV editor uses, I would expect it to follow image orientation, but perhaps this is just matter of implementation.
Cropping could perhaps be done with modifier key.

In D11933#313158, @Peter Fog (tintwotin) wrote:

I'm not sure why you guys want to keep exposing a value which doesn't make any sense to users, and then have the degrees as a label?

Technically this never worked properly, but arguably rBd373b43f07f5 made it a bit worse. Got some initial fix ready, will continue tomorrow to cover more edge cases.

VSE should be able to use DNA defaults, but this wouldn't resolve this issue for some properties like blend mode mentioned here.

Currently strips don't snap to themself, because this is bit more tricky to handle when more strips are selected.

Yes this indeed looks like a mistake, I missed the point that in light theme this color is already different.

Just don't define arguments, you don't need. Like this:

I asked because of statement:

I also observe that it does not detect any graphics card

Can confirm, I think this would be nice to resolve, even though it may not be bug technically. So will check how this can be resolved and if this proves to be difficult I may classify this as known issue

I can reproduce the issue, but strangely it dissappears when showing frame numbers instead of time. This is the case in all editors not only VSE.

In D11995#311153, @Félix (Miadim) wrote:

Re-reading it, I think it would be better to move that to rna_sequencer.c -> rna_def_meta to have the function accessible only for meta sequences.

Thanks I almost forgot to close this

So far I have checkedperformance and don't see any regression here

Can you upload movie file here as well? I suspect this would be seek bug

In D6110#312859, @Peter Fog (tintwotin) wrote:

@Richard Antalik (ISS) The horizontal line is drawn in consistency with the horizontal line in the graph editor(for quick check see the gif above).

I think this is intentional, but looks to be incorrect anyway.

Sorry, I must have been doing something wrong, I have re-checked and I see curves now. I see they are normalized too. I would probably suggest using different color for 0 line. I find it distracting when it is same as current frame. Personally I would go for 1px thin darker line in same hue as strip body. Or not draw it at all if it is still distracting.

In T90280#1199602, @Germano Cavalcante (mano-wii) wrote:

I'm still not sure what this "Hold Offset" is.

It's when you drag handle beyond strip content so it outputs static image. Kinda sorta, but in context of snapping that boundary is the hold offset.

Patch is correct, so I accept.

Can you link where this was reported? This probably fixes the issue of availability, but either I don't know how to use use_stamp_sequencer_strip, or it is broken. So I just want to have more context before I check more in detail.

I have already applied this patch so it's probably best to respond (already...).

@Peter Fog (tintwotin) I re-checked the file and still can't reproduce bug, but I think it is related to prefetch being enabled. There is task to address this issue T90041.

In D11790#311772, @Peter Fog (tintwotin) wrote:

What is the point of having the horizontal and vertical lines the same theme color?

They have two different purposes. And why do this theme color not have alpha?

Can't really say why. I guess this could be changed

In D11939#311766, @Peter Fog (tintwotin) wrote:

I'm using this clip: https://www.pexels.com/video/aerial-shot-over-a-white-sand-beach-4135116/
In this .blend file:

speed4.blend603 KBDownload

• Nor seq_speed_effect_target_frame_get shoul be static :/

• Oops, seq_effect_speed_rebuild_map is not static

In D11933#311161, @Peter Fog (tintwotin) wrote:

@Richard Antalik (ISS) So, how to proceed?

In T34064#1197997, @ok what (ok_what) wrote:

Hi, sorry for the late reply.
If this feature is something that's quickly needed, it might be a better idea for someone more experienced to work on it. My experience in general and especially with the Blender codebase is low. I originally took this up since it seemed like there was little activity and no-one was working on it.
I currently don't have much time to work on this, but I could try to help if I find time for it later.

As for the previous comment, yes I was using the transform operator to implement it.

In D11939#311698, @Peter Fog (tintwotin) wrote:

Actually I first applied Germanos patch and did a gif(with no flicker): https://developer.blender.org/D6110#311650

Then I applied your patch on top of it, and then there was flicker. As mentioned it's just animating any value, it shouldn't be related to the source material.

• Use static qualifier

• Clarify seq_effect_speed_rebuild_map return value

In D11939#311651, @Peter Fog (tintwotin) wrote:

When not using interpolation I'm getting flicker and render errors with this patch:
(Using Frame Number going from a key with the full length to zero)

In D11939#310961, @Germano Cavalcante (mano-wii) wrote:

With that maybe we could avoid checking if there is an f-curve with id_data_find_fcurve. But a 7ms overhead doesn't seem to be a problem.

• Fix memleak

In T90271#1198582, @Kevin Kelly (kelly) wrote:

Same issue persists with experimental build

Do you use Pulseaudio as output sound device? Check in User preferences > System > Sound.

• remove unrelated change
• Add only overwrite mode to header.

In D11790#311319, @Julian Eisel (Severin) wrote:

These are nice improvements in my opinion. Good to see the VSE getting a bit more attention on the aesthetics, where it's clearly lacking behind (like in many other parts).
But I'll let @Francesco Siddi (fsiddi) give his feedback first.

• Add a description for show_grid property