800&XL/XE graphic modes ANTIC&CTIA/GTIA

[Bryan]

Yes, but bit 0 of these registers is ignored. More than likely this “bit” doesn’t exist in the hardware.

 

Right, you can look at the GTIA datasheet and see that bit 0 doesn't exist.

 

-Bry

[analmux]

Again, this way of phrasing is a bit misleading. The difference is not exactly between GTIA modes and CTIA modes. What makes possible the 256 colors on gr.9 (and again, only in gr.9) is the fact that the “luminance” registers are not used.

878234[/snapback]

 

Sounds like a good point. Yes indeed gr.9 doesn't use luminance of setcolor-registers, but we still have the question: why (even when GTIA is able to show 16 shades in gr. 9) does it still have the 128 color palette?:

 

So: Why are only 8 luminance steps possible without the GTIA modes?

Is there a design bug in the hardware that I do not know?

 

Maybe GTIA had to be downwards compatible (in normal modes) with CTIA. Imagine that all in a sudden GTIA had a 2 times bigger palette than the CTIA, and then people make graphics with the extra shades, then these wouldn't be displayable on the CTIA (however the same would hold for the extra GTIA modes )

 

The only practical way to display 256 colors is to have 16 or more bands of graphics 9, each of them with a different HUE set by a DLI. It is probably feasible to use cycle counting and change the HUE color in the middle of a scan line. That would work instead of DLIs. But it still must be done with gr.9.

 

Remember, whe can always use missiles (in 5th player mode) to mix with gr. 9. There are 32 colors possible on 1 line (16 shades of 2 colors)

[ijor]

but we still have the question: why (even when GTIA is able to show 16 shades in gr. 9) does it still have the 128 color palette?:

...

Maybe GTIA had to be downwards compatible (in normal modes) with CTIA. Imagine that all in a sudden GTIA had a 2 times bigger palette than the CTIA, and then people make graphics with the extra shades, then these wouldn't be displayable on the CTIA (however the same would hold for the extra GTIA modes  )

 

It is a very good question, that probably we can only guess the answer.

 

One possible explanation is yours, actually it could be even worse. If GTIA would use bit 0 of the luminance registers, this might break compatibility with older software. It is possible that same old games "poked" different values to that bit for some reason, while the intention was to get the same color. That could be intentional or not. The point is that if GTIA would use that bit, then the display won't be the intended one.

 

A more likely explanation is a limitation of chip resources. Adding the extra flip-flops (for the unused bits), plus all the necessary support to make them work needs extra chip resources. Might be there wasn't enough.

 

Remember, whe can always use missiles (in 5th player mode) to mix with gr. 9. There are 32 colors possible on 1 line (16 shades of 2 colors)

 

Again, this might be nice for enhancing the picture, but it won't be very helpful for the purpose of displaying 256 colors at once. No matter what you do. With any gr. mode besides 9, or with player-missiles, you only get 128 colors. The extra 128 colors must be displayed with gr.9.

 

This means that short of using cycle counting (for chaging the HUE register in the middle of a scan line), you still must use 16 bands of gr.9

[emkay]

(I've had a better look)

Yup, color 1 and 15 are practically the same. Furthermore I've noticed that shades 0 and  1 are so dark they look the same, but when you turn up the monitor brightness a little you actually see difference.

 

So we have a total of 15 colors with 16 shades in gr.9 or 8 shades in other modes = 120 or 240 colors.

877847[/snapback]

 

 

I'm not shure, what computers you have, but my XL(s) has as colour "1" a rust/gold and as "15" a green/brown. To have a real "yellow", colour 15 is to use. With colour "1" you get a light gold colour.

On XE machines the last colour gets more into green...

[+CharlieChaplin]

hmmm,

maybe he is talking of SETcolor while he writes Color...?!?

 

Anyway, on my Atari 800XL (PAL) the color that comes after black is brown. Now if you step down all colors until you reach black again, the color before black is the same brown color again. Or shorter: the colors before and after black are brown and they seem to be the same brown colors... (at least for the human eye).

 

Only with an oscilloscope one can find out a difference (at least someone told me so), but for the human eye these brown colors are the same. Funny thing is, this seems to be typical for PAL systems. when I read color tables written on paper for NTSC systems, the first brown should be a more red-brown...

 

afaik, this (PAL) bug not only appears with colors, but also with luminances - at least for the human eye two luminances seem to be the same. thats why I often say that NTSC Ataris have 16 x 16 = 256 colors, while PAL Ataris have only 15 x 15 = 225 colors...

 

And thats some kind of explanation to me why 256 color pix originally painted or converted on a NTSC system don`t look very good on a PAL system...

 

Otherwise said, in the last 20 years I have converted dozens of GIF (and other) pictures into 256 color atari pix (with my PAL Atari 800XL). There were lots of PC pictures with red and brown tones which appeared only as brown tones on my Atari (allthough Atari has red tones!). And there were also lots of PC pictures with blue and violet/lila (leela?) tones that appeared only as violet/lila on my Atari...

 

-andreas.

[Kr0tki]

afaik, this (PAL) bug not only appears with colors, but also with luminances - at least for the human eye two luminances seem to be the same.

881475[/snapback]

Rather than human eyes, this has more to do with brightness/contrast setting of your monitor. Like analmux said, it is possible to tune the monitor to see 16 different luminances.

[deathtrappomegranate]

violet/lila (leela?) tones that appeared only as violet/lila on my Atari...

 

-andreas.

 

lila=lilac?