Mouse latency laced screenmode

[Octopus66]

Hi Toni - hope you are keeping safe!

I was wondering why my perception of the mouse latency is poorer in Hires Laced vs Hires? Is this just me - or is the latency noticeable lower?

I am using Vsync beamrace 4 slice and latest winuae 4.3.0

I don't think this behavior has changed, I just haven't used hires laced much seen beamracing vsync was introduced.

Thanks

[Toni Wilen]

Perhaps you have "frames" interlace mode (Display panel)? Change it to fields or fields+ to get full frame updates.

[Octopus66]

Thanks Toni - this seems to do the trick!

[rutra80]

Here it doesn't make any difference - on laced Workbench mouse pointer is twice as choppy no matter if in Frames or Fields modes.

[Toni Wilen]

Make sure you did adjust interlace mode, not normal mode. (Display panel has 2 sets of similar options)

[rutra80]

Yup, I did. Even if I set both to fields(+), as soon as I switch Workbench to laced, pointer gets half the refresh rate. Tried even with black frame insertion to make it as close to real Amiga as possible.

[Toni Wilen]

Perhaps you have some other options that affect it enabled. Try normal windowed mode without any vsync etc..

Do you see "flicker" effect if you use fields+?

[rutra80]

I do see the flicker effect. Tried it every possible way around including inbuilt A500 & A1200 configs - in laced modes cursor has half the refresh rate. You surely know Blur Busters motion tests. WinUAE behaviour is exactly as in Phantom Array Effect test - in laced modes cursor loses half the positions (there should be all even if with comb artifacts).

[Toni Wilen]

It looks like you don't talk about hardware refresh rate (screen and mouse cursor sprite is updated at 50Hz. Difference is obvious between double=25Hz and fields=50Hz) but how AmigaOS handles mouse cursor in interlace mode. This is normal behavior.

[rutra80]

There's no difference here in Workbench mouse cursor fluidity in laced modes between double & fields (except that in fields there are comb artifacts).
Could you please elaborate a bit on how AmigaOS handles mouse cursor in interlace mode? IIRC on real Amiga it was as fluid as in non-laced mode?

[Toni Wilen]

Originally Posted by rutra80

There's no difference here in Workbench mouse cursor fluidity in laced modes between double & fields (except that in fields there are comb artifacts).

At least opposite field "ghost" image should be nearer the actual mouse image when moving the mouse fast.

Could you please elaborate a bit on how AmigaOS handles mouse cursor in interlace mode? IIRC on real Amiga it was as fluid as in non-laced mode?

Problem is interlace itself.

It can only look correct in "fields+" mode where alternate lines are blanked. Without it comb artifacts are visible (both fields are visible simultanously when in real interlace they have 1/50s difference)

You can see exact same effect when running any demo that has interlace graphics and horizontal scroller. Scroller will look horrible without "fields+".
(for example Shed Tears by Ozone). But it can flicker also badly if refresh rate is not in sync with monitor = vsync or variable sync monitor required.

In WB (also works in some programs that take over the system) you can "fix" it by ticking "Remove interlace artifacts" option (=emulate frame as progressive by basically emulating both fields at the same time)

[rutra80]

Yes. But it doesn't seem to be what happens here.

In PAL non-laced modes, when on WB I move mouse cursor fast, I can see its full image in the next position every 1/50s - correct.

I switch to laced, fields(+) - I should see odd fields of cursor in 1st position on 1/50 frame, then even fields of cursor in 2nd position on 2/50 frame, then odd fields in 3rd position on 3/50 frame, etc. - unfortunately that's not what happens.
What I see is that cursor moves every 1/25s.

But. I do see some combing artifacts. After closer look it seems like in laced modes mouse cursor position isn't sampled evenly.
When I move mouse say 50 pixels/frame, its coordinates along 8 frames should be as follows:
0 50 100 150 200 250 300 350 etc...

While here in laced modes it seems to be something like this:
0 1 99 100 200 202 299 301 etc...

That might be why it feels so choppy...