Fullscreen graphics distortion

All about ST/STE demos

Moderators: Mug UK, lotek_style, Moderator Team

defjam
Retro freak
Retro freak
Posts: 11
Joined: Mon Sep 25, 2006 10:50 am
Location: Dresden/Germany
Contact:

Fullscreen graphics distortion

Postby defjam » Mon Sep 25, 2006 11:40 am

Did anyone noticed some fullscreen graphics distortion on some TV sets ?
(for example the Shadow of the beast main menu from the Phaleon demo) ...
it seems that having data in the right border causes this problem.

Any ideas ?


btw.. I figured out that there's no distortion of the graphics if you use the ULM - like way of "closing the right border" (switch to medres instead of hires)

User avatar
keops
Atari Super Hero
Atari Super Hero
Posts: 590
Joined: Mon Jul 26, 2004 3:39 pm
Location: Canada
Contact:

Postby keops » Mon Sep 25, 2006 6:00 pm

What do you mean exactly by graphics distortion?

defjam
Retro freak
Retro freak
Posts: 11
Joined: Mon Sep 25, 2006 10:50 am
Location: Dresden/Germany
Contact:

Postby defjam » Mon Sep 25, 2006 6:33 pm

Graphics distortion - the graphics are distorted (like some sinewave) - but it occured only the TV sets here at home. It's NOT sync errors!!

ijor
Hardware Guru
Hardware Guru
Posts: 3815
Joined: Sat May 29, 2004 7:52 pm
Contact:

Postby ijor » Mon Sep 25, 2006 6:58 pm

Yes, of course, stabilization is done well before the line signal ends. It usually doesn't matter because most monitors or TVs won't display that far. But some do.

User avatar
keops
Atari Super Hero
Atari Super Hero
Posts: 590
Joined: Mon Jul 26, 2004 3:39 pm
Location: Canada
Contact:

Postby keops » Mon Sep 25, 2006 8:42 pm

I haven't coded a fullscreen since 1993 or something but I remember that big parts of the fullscreen were not displayed on my TV, unlike what others displayed. I probably never encountered those distortions with my TV (afair).

User avatar
leonard
Moderator
Moderator
Posts: 658
Joined: Thu May 23, 2002 10:48 pm
Contact:

Postby leonard » Tue Sep 26, 2006 9:44 am

Well don't know exactly what's the effect. You speak about Phaleon mainmenu or did you see that distrotion with every fullscreen on your specific TV?

If it's only some demos, maybe they use a bad "top border removal". Some people set 60hz mode in VBL, and keep at 60hz till the shifter start to decode, and set set 50hz. So the top border is removed, but about 33 lines are parsed at 60hz, which do a sin distortion on some monitors/TV. (cf level16 fullscreen in Union demo use that technic, I noticed that few years ago when trying to run it under SainT :-))

But you may speak about something else.
Leonard/OXYGENE.

defjam
Retro freak
Retro freak
Posts: 11
Joined: Mon Sep 25, 2006 10:50 am
Location: Dresden/Germany
Contact:

Postby defjam » Tue Sep 26, 2006 12:37 pm

Leonard: no - not all fullscreens. f.e. the ULM ones are working fine on the TV sets (and they have graphics and rasters in the borders)... I think it's due to the switch to medrez instead of hirez to "close" the right border. (The first one - the medrez method, is used by ULM)

ijor
Hardware Guru
Hardware Guru
Posts: 3815
Joined: Sat May 29, 2004 7:52 pm
Contact:

Postby ijor » Tue Sep 26, 2006 6:30 pm

Hi Leonard,

leonard wrote:If it's only some demos, maybe they use a bad "top border removal". Some people set 60hz mode in VBL, and keep at 60hz till the shifter start to decode, and set set 50hz. So the top border is removed, but about 33 lines are parsed at 60hz, which do a sin distortion on some monitors/TV. (cf level16 fullscreen in Union demo use that technic, I noticed that few years ago when trying to run it under SainT :-))


Are you sure the L16 screen in the Union Demo does that?

I remember checking that screen not too long ago and I couldn't see that. Actually, what it seems it does, is to sync only once (not every VBL) and then it toggles frequency every single scan line (displayed or not).

User avatar
leonard
Moderator
Moderator
Posts: 658
Joined: Thu May 23, 2002 10:48 pm
Contact:

Postby leonard » Tue Sep 26, 2006 9:04 pm

oh ijor I just checked it again with SainT debugger and you're right... they do overscan lines on the complete screen. Very strange, coz I really fix a bug in SainT when screen is at 60hz at the beginning. It's not easy to trace initialisation code but maybe they use that trick to do the first sync, at init time.

defjam: never heard any "distortion" difference between ULM (med res) and classic (high res) shifter-stabilization technique. There is a difference because it does not display exactly the same bitmap at the top right, but never heard about distortion.

But maybe Alien/StCnx could answer such question, he's a fullscreen-technic expert.
Leonard/OXYGENE.

ijor
Hardware Guru
Hardware Guru
Posts: 3815
Joined: Sat May 29, 2004 7:52 pm
Contact:

Postby ijor » Tue Sep 26, 2006 10:04 pm

leonard wrote:It's not easy to trace initialisation code but maybe they use that trick to do the first sync, at init time.


Ah, yes, you are right. They use something like that on the setup/calibration, just for the purpose of finding out if the machine's top border is 47 or 63 lines:

Set 60Hz at VBL.
Switch to 50Hz after the first line.
Count the number of timer B interrupts until next VBL.

There is a difference because it does not display exactly the same bitmap at the top right


Interesting. Can you elaborate on this? What do mean by "top right"?

User avatar
thothy
Hatari Developer
Hatari Developer
Posts: 427
Joined: Fri Jul 25, 2003 9:36 am
Location: Germany
Contact:

Re: Fullscreen graphics distortion

Postby thothy » Wed Sep 27, 2006 10:24 am

Defjam__ wrote:Did anyone noticed some fullscreen graphics distortion on some TV sets ?
(for example the Shadow of the beast main menu from the Phaleon demo) ...
it seems that having data in the right border causes this problem.


Seems like this already happened in the "good old days"... cite from http://www.tcs.informatik.uni-muenchen. ... screen.txt :

A line consists of 25.5 instead of 20 words without the right border,
but should use only 23 of these 25.5 words, because the picture
is distorted on some STs (We made this mistake in the "Musical
Wonder - 1991") if you use too much words. Anyway, there are
hardly any monitors that have such a huge visible right border. .

User avatar
leonard
Moderator
Moderator
Posts: 658
Joined: Thu May 23, 2002 10:48 pm
Contact:

Postby leonard » Wed Sep 27, 2006 7:49 pm

Interesting. Can you elaborate on this? What do mean by "top right"?


hard to explain without a monitor. classical fullscreen (high res switch) clear the last 12 or 16 pixels of a scanline (what I call "top right"). If you run ULM demo, if you can "scroll" the picture with your monitor to display the top-rigth, you will see the last 16pixels column is very strange: not black, and no real graphics, a kind of pixel-soup :-) This is noticeable with ULM fullscreen , and I suppose it comes from the medres switch (but I really have no explain).

I'm pretty sure Alien could explain this coz he has studyed it I guess (there is a very good article on St-Magazine, a french magazine)
Leonard/OXYGENE.

ijor
Hardware Guru
Hardware Guru
Posts: 3815
Joined: Sat May 29, 2004 7:52 pm
Contact:

Postby ijor » Wed Sep 27, 2006 8:59 pm

leonard wrote:hard to explain without a monitor. classical fullscreen (high res switch) clear the last 12 or 16 pixels of a scanline (what I call "top right"). If you run ULM demo, if you can "scroll" the picture with your monitor to display the top-rigth, you will see the last 16pixels column is very strange: not black, and no real graphics, a kind of pixel-soup :-)


Thanks Leonard. But why the term "top right"? This effect happens only on the top scan(s) or on the last pixels of all scans?

User avatar
leonard
Moderator
Moderator
Posts: 658
Joined: Thu May 23, 2002 10:48 pm
Contact:

Postby leonard » Thu Sep 28, 2006 6:46 pm

But why the term "top right"? This effect happens only on the top scan(s) or on the last pixels of all scans?


oh sorry ! I mean every scanlines. "top" is not the right word, I mean "the most right", or "the last pixels of each scanline" :-)

If you have a monitor you could try a oxygene fullscreen and an ULM fullscreen, it's quite visible.
Leonard/OXYGENE.

User avatar
Zorro 2
Administrator
Administrator
Posts: 2209
Joined: Tue May 21, 2002 12:44 pm
Location: Saint Cloud (France)
Contact:

Postby Zorro 2 » Tue Oct 24, 2006 10:29 am

Do you know this program : Un-Overscan ???
You do not have the required permissions to view the files attached to this post.
Member of NoExtra Team

User avatar
alien
Atari maniac
Atari maniac
Posts: 95
Joined: Sat May 01, 2004 4:01 am
Location: USA
Contact:

Re: Fullscreen graphics distortion

Postby alien » Tue Oct 24, 2006 5:19 pm

thothy wrote:
Defjam__ wrote:Did anyone noticed some fullscreen graphics distortion on some TV sets ?
(for example the Shadow of the beast main menu from the Phaleon demo) ...
it seems that having data in the right border causes this problem.


Seems like this already happened in the "good old days"... cite from http://www.tcs.informatik.uni-muenchen. ... screen.txt :

A line consists of 25.5 instead of 20 words without the right border,
but should use only 23 of these 25.5 words, because the picture
is distorted on some STs (We made this mistake in the "Musical
Wonder - 1991") if you use too much words. Anyway, there are
hardly any monitors that have such a huge visible right border. .


Ok, I'm awake now... Four suggestions

1/ Small distortions of a few pixels on ULM screens on some STs in some wakeup states is due to the midrez stabilizer working less well than the hirez one... Usually the screen shimmers

2/ If the timing of the stabilizer wrong/opening the left border too long the sync signals are messed up. My TV would distort massively in such cases.

3/ If opening the right border is done by switching to 60Hz and switching back during or after HSYNC nasty things would happen.

4/ XXX international and others set 60Hz at the top of the screen and switched back to 50 very late in the game, causing the top of the screen to distort (and sometimes the whole screen)

I never ever saw what Flix said about the Musical Wonder Demo. He may be right, or he may have had bad overscan timings, and clearing the left side helped. I have my doubts because if you look at the RGB signal output SYNC is supposed to override any data, so the pins can do whatever they want. A UHF encoder should do the same. Also, most demos don't keep those words clean... Few monitors even show that far right.

Please bare in mind I haven't code a full screen for 14 years now...
Alien / ST-Connexion

Gunstick
Captain Atari
Captain Atari
Posts: 263
Joined: Thu Jun 20, 2002 6:49 pm
Location: Luxembourg
Contact:

Postby Gunstick » Thu Dec 07, 2006 3:22 pm

leonard wrote: If you run ULM demo, if you can "scroll" the picture with your monitor to display the top-rigth, you will see the last 16pixels column is very strange: not black, and no real graphics, a kind of pixel-soup :-) This is noticeable with ULM fullscreen , and I suppose it comes from the medres switch (but I really have no explain).


I can't explain it either. I was just proud that I could diplay some pixels more than any other fullscreen... followed by 8 pixels of background color and then some other pixels I don't know where they come from.

Georges


Social Media

     

Return to “Demos - General”

Who is online

Users browsing this forum: No registered users and 4 guests