Atari 8-Bit Graphics Capabilities

[R0ger]

17 minutes ago, Gunstar said:

The flickering on these images is not nearly as bad as the old GTIA flicker you got with interlacing modes, not NEARLY, it's very minor in comparison, and would probably be even less with NTSC. Except you need PAL for the color blending here.

Main reason is that GTIA flicker is used to increase resolution. You can control it with blur, but generally you flicker 2 quite different brightness levels. My flickering is used to increase color depth. I can guarantee the brightness level to be closer, and I can use deflickering patterns. I was never friend of flickered modes. But on many CRT TVs this is just OK. And on LCDs, projectors, or emulators it's no issue whatsoever.

[_The Doctor__]

31 minutes ago, R0ger said:

Main reason is that GTIA flicker is used to increase resolution. You can control it with blur, but generally you flicker 2 quite different brightness levels. My flickering is used to increase color depth. I can guarantee the brightness level to be closer, and I can use deflickering patterns. I was never friend of flickered modes. But on many CRT TVs this is just OK. And on LCDs, projectors, or emulators it's no issue whatsoever.

this also explains why some people use the same modes for their flicker choices and one can look very noticeable and the other does not, perhaps they adjust the levels so they are more inline with each other to lessen the detracting affect..

[+Stephen]

1 hour ago, emkay said:

A bunch of posts read that caused some  de ja vu    Interlace, flicker ? In some cases it could impress ...

 

Sigh....

 

One of THE  features to have stable graphics manipulations on the Atari is  WSYNC .

 

How about some special super duper hires color mode that allows 100 sprites at 30x60 pixel.

Well, I know, it's not possible....

 

Back in the days I planned to do a graphics mode in ANTIC D with software colors. It would result in NO scanlines ( yeah ) and allowed to have a lot manipulation due to the low Cycle Stealing. But then the Atari had to wait more than 12 Years on the Attic and all graphics tools had been lost.     

Hmm - another post that caused some deja vu.  Emkay telling others he could do it better, and ... WAIT FOR IT ... in Mode D to boot.  You wonder why I seem to repeat this joke?  LMFAO.  I should invent a drinking game.  Every time you mention Mode D, I take a drink.  My liver would die in 2 months.

[PixelCrunch]

8 hours ago, Heaven/TQA said:

 

 

Jesus fucking Christ... ace...

 

re: loading screen

 

ok. Me biased.... but having Thalion intro screen or Lucasfilm intro gives you a first quality sign...

 

or bitmap brother logos...

 

 

Jesus christ, those are really well done art pieces.

 

My favorite one is the Coco art. It looks like something straight from the movie reel, but downsized.

[R0ger]

19 minutes ago, PixelCrunch said:

Jesus christ, those are really well done art pieces.

 

My favorite one is the Coco art. It looks like something straight from the movie reel, but downsized.

It's converted images. Not 'art'. If you want to see it in movement, I did test which only works in emulator, here: 

It's not Coco though. It's different kind of art ?

[Gunstar]

@R0ger Am I right in my understanding that RGBY 8x11F, and all other RGBY modes for that matter still have the 128 color Antic palette at their full disposal/limit, or does it create a larger virtual palette with what you are doing, like Colorview does with flipping three 16 shade R,G,B screens to create a virtual palette and on-screen colors of 4096?

 

Do on-screen colors, at least at this point in your RGBY modes, max out at 88 for 8x11 mode and go down from there? Or is this a misnomer on my part in my figuring that's exactly what 8x11 means, 8 shades/brightness levels by 11 colors(or vice-versa if I mixed it up), or have I forgotten, missed or mixed up that explanation? Or have I totally misunderstood?

 

perceptually some of those images seem like more than 88 colors or even 128. of course that is the case with some Rasta images too, to me. Even though I know it's limit is 128 color palette and maximum on screen colors, and most of the time average half of that.

Edited July 20, 2019 by Gunstar

[VladR]

8 hours ago, popmilo said:

 

@VladR Here is cpu dma info for you:

 

Antic_Timings.txtFetching info...

 

Kernels can be made 100% stable and predictable. Yes it takes couple of scanlines to stabilize wsync, but after that with known PM positions and hscrol values you can be sure in each cycle execution.

 

Thanks ! That's the file I was looking for.

I'm confused why would it take more than 1 scanline to stabilize via WSYNC ? Meaning, you must do DLI with WSYNC on first several scanlines (not just one?), otherwise there's still jitter ? Or you mean something else ?

7 hours ago, Gunstar said:

The flickering on these images is not nearly as bad as the old GTIA flicker you got with interlacing modes, not NEARLY, it's very minor in comparison, and would probably be even less with NTSC. Except you need PAL for the color blending here.

So, you tried it on CRT TV ? You mean it doesn't flicker as bad as some demos did 30 years ago ? Because that's the only reference point I got (for flickering).

25 minutes ago, Gunstar said:

@R0ger Am I right in my understanding that RGBY 8x11F, and all other RGBY modes for that matter still have the 128 color Antic palette at their full disposal/limit, or does it create a larger virtual palette with what you are doing, like Colorview does with flipping three 16 shade R,G,B screens to create a virtual palette and on-screen colors of 4096? And on-screen colors at least at this point in your RGBY modes max out at 88 for 8x11 mode and go down from there? Or is this a misnomer on my part in my figuring that's exactly what 8x11 means, 8 shades/brightness levels by 11 colors(or vice-versa if I mixed it up), or have I forgotten, missed or mixed up that explanation? Or have I totally misunderstood?

perceptually some of those images seem like more than 88 colors or even 128. of course that is the case with some Rasta images too, to me. Even though I know it's limit is 128 color palette and maximum on screen colors, and most of the time average half of that.

Here's my understanding:

 

No Flicker modes:

RGBY.4x4 : 4 Hues, 4 Brightness: Bitmap mode

RGBY.4x6 : 4 Hues, 6 Brightness: Character mode

 

Flicker modes: (Hues*2), (Brightness*2 - 1)

RGBY.8x7F   : 8 Hues, 7 Brightness: Bitmap mode

RGBY.8x11F : 8 Hues, 11 Brightness: Character mode

 

You can only select from base 128 colors (the values you POKE into color registers). But, once blending happens, I believe it creates new color space due to the nature of the blending process (chroma:hue is subsampled also vertically on PAL) - hence you should get much more than 128 colors, as at that point you are not limited by ANTIC's capabilities, but rather by the PAL color format (the vertical subsampling of chroma).

 

That would make a great experiment - to compare results by color bleeding vs the standard 128 color palette! I'd love to see that !

[Gunstar]

5 hours ago, VladR said:

 

So, you tried it on CRT TV ? You mean it doesn't flicker as bad as some demos did 30 years ago ? Because that's the only reference point I got (for flickering).

Here's my understanding:

 

Basically, yes. Though as I said before, I'm feeding it through a converter to VGA and not direct composite/S-video, so it's possible that I'm getting some slightly different results when my converter processes the incoming signals and spits them out VGA on the other side, as I said, I have to turn off my converters 3D Y/C comb filtering/motion adaptive de-interlacing or some really strange effects occur. This happens with Space Harrier too. But I have not looked at the images in direct composite/S-video input to my TV/monitors as I'm using a PAL computer on 50/60HZ  screens that will not show color from a PAL video signal, hence my need to convert the signal to VGA; my converter, which can successfull process both NTSC and PAL video to VGA also spits out a PAL-60 VGA signal that does show color on my screens from the PAL 50Hz input video signal, regardless if I have it's comb filtering and de-interlacing turned on or off. So what I see is NOT the norm that most will see on these PAL images.

 

Though I'm referring to old static images made of interlaced GTIA modes of around 30 years ago rather than demos, but yeah, I'm sure similar techniques where used in those old demos too, as I recall. Or some of the HIP/TIP/RIP modes which I can never remember/keep track of what's what in those software/mixed modes, again judging by static images I recall compared to the flickering I see in these images. 

 

As to the hue/brightness of RGBY's, thanks for, presumably, helping me to understand number of hues/brightness levels and that some new color-space or virtual colors as I referred to them is happening to go beyond the hard 128 color palette.

Edited July 20, 2019 by Gunstar

[R0ger]

5 hours ago, Gunstar said:

@R0ger Am I right in my understanding that RGBY 8x11F, and all other RGBY modes for that matter still have the 128 color Antic palette at their full disposal/limit, or does it create a larger virtual palette with what you are doing, like Colorview does with flipping three 16 shade R,G,B screens to create a virtual palette and on-screen colors of 4096?

 

Do on-screen colors, at least at this point in your RGBY modes, max out at 88 for 8x11 mode and go down from there? Or is this a misnomer on my part in my figuring that's exactly what 8x11 means, 8 shades/brightness levels by 11 colors(or vice-versa if I mixed it up), or have I forgotten, missed or mixed up that explanation? Or have I totally misunderstood?

The palette resulting from PAL mixing itself is close to existing Atari colors. Color $20 and $0f will give something rather close to 2F.

But with flickering, different palette is created. I wouldn't say larger, but different. First the hues. Atari hues lie on a circle in color space. I'm picking 4 colors on the circle. Not even equally spaced. It's more like equilateral triangle for RGB, and yellow stuck between R and G. Flickering then creates hues between 2 base colors. So these new hues are somewhat at different angles, and they all have slightly worse saturation.

 

At the moment I use these colors for hues: 70,B0,E0,20 .. basically RGB are picked to be as strong and as single-component as possible. And then yellow is added. If you imagine the color space, the base colors create polygon. The flickering and dithering then can create virtual colors inside that polygon. So if some picture would be heavily skewed toward some color, the base colors could be picked differently. Though I like the system RGB + one extra color to extend the polygon toward area of interest.

 

Brightness levels are simpler. I use these 6: 0,4,8,10,12,14 .. as you can see, there is more detail in the bright values, as both flickering and dithering are more visible in bright areas. Flickering gives me extra levels very similar to original 2,6,9,11,13.

 

This also brings problem. When I'm dithering the images on PC, I'm using this 11x7=77 color palette, which I then recode into hue/bright values, and into 2 fields for flickering modes. But I have to compute the palette first. I have to simulate the way base colors combine into the resulting colors. I have very simple formula for that, which only picks other Atari palette color, but most of the time I use different method. I create XEX with all colors possible, I make screenshot in Altirra, with PAL artifacts and frame blending, and then I use this screenshot as reference.

 

This is old reference image from 11x7F mode:

 

 

Today I don't mix hues only in neighbor pairs only, I do all combinations. So instead of adding R-Y,Y-G,G-B,B-R I also add R-G and B-Y. So I actually have 10 hues, and the hues can't be ordered nicely like this, so today I use different reference pattern. The old one also had to be processed to remove the scanlines, the new one can be used straight from Altirra.

 

 

Quote

perceptually some of those images seem like more than 88 colors or even 128. of course that is the case with some Rasta images too, to me. Even though I know it's limit is 128 color palette and maximum on screen colors, and most of the time average half of that.

There is another level of color interpolation with dithering. And that's complex topic by itself. Rasta can't do it that well, at least it can't adapt dithering to current palette. It has to do dithering first, and then adapt palette to it. Also my approach has basically every color (minus character limits) available everywhere, which helps.

 

Edited July 20, 2019 by R0ger

[Gunstar]

 

Thank you @R0ger that improves my non-programmer understanding greatly.

[popmilo]

6 hours ago, VladR said:

I'm confused why would it take more than 1 scanline to stabilize via WSYNC ? Meaning, you must do DLI with WSYNC on first several scanlines (not just one?), otherwise there's still jitter ? Or you mean something else ?

So, you tried it on CRT TV ? You mean it doesn't flicker as bad as some demos did 30 years ago ? Because that's the only reference point I got (for flickering).

Here is information from Altirra manual:

Quote

A write to WSYNC [D40A] halts CPU execution until the end of a scan line, allowing the CPU to synchronize to
the display. One more cycle elapses before the CPU is halted until cycle 105, when execution resumes around
the start of horizontal blank. Because the CPU usually gets to execute the first cycle of the next instruction, this
appears as if the instruction started on cycle 104. There are, however, three circumstances that can change this
behavior:
• If the cycle immediately after writing WSYNC is blocked.
In this case, the CPU doesn't get to execute the first cycle of the next instruction, and that instruction
starts from the beginning as usual on cycle 105.
• If playfield DMA extends to cycle 105.
Wide playfields, normal playfields with horizontal scrolling, and narrow playfields with high horizontal
scroll values can encroach on cycle 105. This causes a one-cycle delay in the CPU restart.
• If refresh DMA extends to cycle 105 or 106.
The first scan line of a character mode line can incur solid playfield DMA during the active region such
that refresh DMA is pushed all the way to the end of the scan line. This can cause refresh DMA to
occupy cycle 105, resulting in a one-cycle delay for the CPU. If playfield DMA is already occupying cycle
105, however, then it will push refresh DMA to cycle 106, resulting in a two-cycle delay.
These factors mean that there can be up to a three cycle variance in when an instruction following a write to
WSYNC finishes execution, not counting interrupts. Therefore, if you are attempting to use a write to WSYNC to
establish an event at an exact time on a scan line, your best bet is to write to WSYNC twice during vertical blank
or during blank mode lines, ensuring that no DMA interference occurs.

So it depends on where on screen that write to wsync happens. Depending on display mode of that scanline and hscroll, different things can happen.

 

If you're in upper border and there's nothing wonky going on, then sure you can have stable code with single wsync write

 

Edited July 20, 2019 by popmilo

[Gunstar]

@VladR

Here are some out of focus, and camera altered color, pictures of my 50/60Hz LCD monitor, with a PAL 50Hz S-video signal, converted to PAL 60Hz VGA through my Ambery video-to-VGA converter/upscaler. One of each is doing a terrible job of conveying the difference between having the converter's 3D Y/C comb filtering & de-interlacing and up-scaling turned off (it's one button all or none) and one of each with it on, again, they are terribly out of focus and also color butchered by my cheap camera phone.

 

I didn't bother to take screen shots on my CRT monitor or of the flicker 8x11F mode pictures, since you can't see flickering and what effects it has going through my VGA converter with a photo. And also because it's a high-res CRT VGA monitor, not a standard definition CRT TV/monitor.

 

MyIDE 2 main menu screen shot, and a Rastaconverter image screen shot. The top picture of each with it off the bottom with it on. You will have to click on each picture and then zoom in x2, due to the poor focus of the pictures, to see any significant difference, which isn't truely representative of either, in real life, but gives you an idea. Essentially, with the 3D processing effects and up-scaling turned on, it makes a 160x240 Rasta image look 320x240 with ant-aliasing.

Edited July 20, 2019 by Gunstar

[+Stephen]

-- DELETED --

[+Stephen]

Sorry for replying to myself - I guess I can't edit posts after a certain time now.  Vlad, as I suspected, I cannot get my camera to sync to a 25/50Hz screen, only 30/60Hz.  But what I can say, is I am VERY susceptible to flicker even on 60Hz stuff.  These images are as close to non-flickering as I have experienced on the machine.  My wife also agreed.  If FlickerTerm is a 10, these images are a 1 or 2 at worst.  I did make a 30 second video of the static text screen, then loading and displaying the image.  Even though the entire video is flickery, it shows no extra when the image displays.  Perhaps I will upload that.

 

1st image is my 32" Samsung 4K showing the images via FireFox web browser.  Now that I think about it, I may have Windows 10 colourblind filter on which is why the image looks so colour saturated and heavy in the reds.

 

Other images are displayed via my 1088XEL configured as PAL.  It has a UAV board for video, and is playing through a medical grade (600 line) 20" Sony PVM using luma/chroma.  The monitor has a very minor purity issue in top right corner, and unless it's my eyes, the focus seems a tad soft.  Last image (#7) is a Rasta Convertor image, showing no "scanlines" from the flicker mode (apologies Gunstar - IIRC you always use NTSC palette so this will not look as good as it should).  As you can see, the Atari just cannot be made to produce red without modifying the hardware!

 

 

[Albert]

1 minute ago, Stephen said:

Sorry for replying to myself - I guess I can't edit posts after a certain time now.

The default edit window is 60 minutes, which should be sufficient to make any corrections to a post.  I had to add this restriction after several previous members edited away their posts--in one instance, several thousand posts were removed (which I later restored from a backup with some custom scripts I wrote).  Subscribers can edit their posts for 30 days, as I figured if someone was willing to pay to become a subscriber, they're going to be less likely to vandalize their own posts (of course, I could be proven wrong here at some point, but 30 days is still better than several years worth of posts being maliciously altered).

 

If you need the ability to edit specific posts, send me a PM or report the post and state that you would like permission to edit it.  I or a moderator can then grant you the ability to edit the post indefinitely.  Most of the time this is used for the first post in a thread that needs frequent updates.  All posts in the marketplace, high score club, and development forums can be edited indefinitely by members, since it's useful to be able to do so in those threads.

 

 ..Al

[+Stephen]

32 minutes ago, Albert said:

The default edit window is 60 minutes, which should be sufficient to make any corrections to a post.  I had to add this restriction after several previous members edited away their posts--in one instance, several thousand posts were removed (which I later restored from a backup with some custom scripts I wrote).  Subscribers can edit their posts for 30 days, as I figured if someone was willing to pay to become a subscriber, they're going to be less likely to vandalize their own posts (of course, I could be proven wrong here at some point, but 30 days is still better than several years worth of posts being maliciously altered).

 

If you need the ability to edit specific posts, send me a PM or report the post and state that you would like permission to edit it.  I or a moderator can then grant you the ability to edit the post indefinitely.  Most of the time this is used for the first post in a thread that needs frequent updates.  All posts in the marketplace, high score club, and development forums can be edited indefinitely by members, since it's useful to be able to do so in those threads.

 

 ..Al

Hmm - my subscription must have ran out then.  I don't recall getting an email about that.  I recall the fiasco of people taking their ball and going home!  OK - fixed.

[Albert]

3 minutes ago, Stephen said:

Hmm - my subscription must have ran out then.  I don't recall getting an email about that.  I recall the fiasco of people taking their ball and going home!

Yeah, looks like it ran out a few days ago. 

 

 ..Al

[Gunstar]

Ok, I managed some much better pictures this evening with the lights off.

Here are RGBY 8x11f images, PAL S-video to VGA PAL-60 output at 1024x768 XGA on an LCD screen, brightness set at 45 and contrast at 90. NO upscaling or other effects, with my 1200XL's SV 2.1 video upgrade. Except the one image "Face" which has two pics and one is with my converter's upscaling/deinterlacing/comb filtering ON-but it does re-introduce some flickering that you can't see in the picture. As you can see, with my upgraded 1200XL video, on LCD VGA, the color saturation is much richer and works well for these images. The camera alters the colors seen in real-life a bit, but overall a good representation of what I see on my monitors, minus a little sharpness.

 

Edited July 21, 2019 by Gunstar

[Gunstar]

4 hours ago, Stephen said:

Sorry for replying to myself - I guess I can't edit posts after a certain time now.  Vlad, as I suspected, I cannot get my camera to sync to a 25/50Hz screen, only 30/60Hz.  But what I can say, is I am VERY susceptible to flicker even on 60Hz stuff.  These images are as close to non-flickering as I have experienced on the machine.  My wife also agreed.  If FlickerTerm is a 10, these images are a 1 or 2 at worst.  I did make a 30 second video of the static text screen, then loading and displaying the image.  Even though the entire video is flickery, it shows no extra when the image displays.  Perhaps I will upload that.

 

1st image is my 32" Samsung 4K showing the images via FireFox web browser.  Now that I think about it, I may have Windows 10 colourblind filter on which is why the image looks so colour saturated and heavy in the reds.

 

Other images are displayed via my 1088XEL configured as PAL.  It has a UAV board for video, and is playing through a medical grade (600 line) 20" Sony PVM using luma/chroma.  The monitor has a very minor purity issue in top right corner, and unless it's my eyes, the focus seems a tad soft.  Last image (#7) is a Rasta Convertor image, showing no "scanlines" from the flicker mode (apologies Gunstar - IIRC you always use NTSC palette so this will not look as good as it should).  As you can see, the Atari just cannot be made to produce red without modifying the hardware!

No, I always use PAL with Rastaconverter. Those colors seem awfully washed-out to me, especially if it's UAV video output (I know it's not the fault of your wonderful Sony Trinitron). Which I thought qualified as modifying the hardware, but apparently not chroma color saturation. And as you can see, with my modified hardware, I have rich colors and true red, from even an Atari PAL palette. And that doesn't change even on a CRT monitor either via S-video2VGA input on my VGA CRT or my (currently non working) CBM 1084-SP monitor via chroma/luma S-video. The color saturation improvement and true reds is all due to the SV 2.1 upgrade, 1200XL version with the 1200's fixed chroma boost circuit; the converter/up-scaler does not increase saturation from the video input when converting. However, my phone-camera does over-saturate the reds a little bit compared to in real life, as well as change the hue of the images a bit.

Edited July 21, 2019 by Gunstar

[+Adam+]

On 7/19/2019 at 5:14 PM, CharlieChaplin said:

Back to topic. Maybe someone can post some nice screens that a) show all available standard A8 gfx. modes (one pic/screen each for Gr. 0 to Gr. 15) and b) several screens with new/modern gfx. modes like HIP, RIP, TIP, CIN, MAX and many others...?!?

 

For the standard modes it would be nice to have some game or demo screens (e.g. "this example shows a Gr. 0 screen:", "this example shows a Gr. 1 screen:", etc.), for the modern gfx maybe one could use some picture or demo screens... ?!?

You mean something like Gury's page about standard graphics modes?: http://gury.atari8.info/ref/graphics_modes.php

 

Some time ago there was a Wikipedia entry about Atari 8-bit software-driven graphics modes (without pictures) - it seems it doesn't exist now, I just found its copies like this page: https://enacademic.com/dic.nsf/enwiki/5875720

 

On 7/19/2019 at 5:39 PM, VladR said:

Wow, just wow ! This is some incredible work. It absolutely deserves its separate thread!!!

I agree, outstanding work by R0ger. But I'm wondering, why still there is no separate thread?

[+Philsan]

4 hours ago, +Adam+ said:

You mean something like Gury's page about standard graphics modes?: http://gury.atari8.info/ref/graphics_modes.php

 

I mentioned that nice page in this thread too. If someone searches in Google "Atari graphic modes", he lands on that page.

Therefore, I think some additional infos would be useful, otherwise people would think Atari max resolution is 160x192 with 5 colors...

 

For example,

Atari has overscan, therefore modes are not limited to 320x192, they can be up to 336x240 pixels.

Thanks to display lists, graphics modes can be mixed together.

Thanks to display list interrupts, you can have tens of colors.

Mixing display modes, colors and sprites (overlapped sprites generate colors) you can have many combinations.

[+Stephen]

9 hours ago, Gunstar said:

No, I always use PAL with Rastaconverter. Those colors seem awfully washed-out to me, especially if it's UAV video output (I know it's not the fault of your wonderful Sony Trinitron). Which I thought qualified as modifying the hardware, but apparently not chroma color saturation. And as you can see, with my modified hardware, I have rich colors and true red, from even an Atari PAL palette. And that doesn't change even on a CRT monitor either via S-video2VGA input on my VGA CRT or my (currently non working) CBM 1084-SP monitor via chroma/luma S-video. The color saturation improvement and true reds is all due to the SV 2.1 upgrade, 1200XL version with the 1200's fixed chroma boost circuit; the converter/up-scaler does not increase saturation from the video input when converting. However, my phone-camera does over-saturate the reds a little bit compared to in real life, as well as change the hue of the images a bit.

Yeah - the colour saturation is rather low, especially when compared to your images!  The pictures I posted were quite representative of what I saw though.  That's definitely an advantage of using LCD vs CRT for images like this.  Sometime very soon I'll be digging my EclaireXL back out, and that uses HDMI.  I'll leave myself a note to revisit this thread when I get it running again.

[Gunstar]

Now, earlier in the thread I was saying the Atari has 15 graphic and text modes, 11 bitmap modes and 5 text modes, like it shows on that link at the beginning, and then someone corrected me saying it was really 8 bitmap modes and 6 text modes, and that the GTIA modes weren't really modes, like Atari claimed in early ads,(400/800 was 8 graphic and 5 text and 1200XL, and later XL ads with GTIA, was 11 graphic and 5 text, and Atari may have started advertising the 400/800 line with 11 and 5 once the GTIA was introduced, I'll have to look. But I have a lot of these classic ads that are changing desktop wallpapers and a couple I have printed on metal sign-plates. But I've seen Atari graphic mode charts like the in this one http://gury.atari8.info/ref/graphics_modes.php as long as I've owned an Atari 8-bit. Now whatever, I'll except that the GTIA modes do or do not qualify as real modes, but what about that 6th text mode? How come it never shows up on these charts? What is the Antic mode?

[flashjazzcat]

You mean the descender text mode? That's ANTIC mode 3.

 

[Gunstar]

5 minutes ago, Stephen said:

Yeah - the colour saturation is rather low, especially when compared to your images!  The pictures I posted were quite representative of what I saw though.  That's definitely an advantage of using LCD vs CRT for images like this.  Sometime very soon I'll be digging my EclaireXL back out, and that uses HDMI.  I'll leave myself a note to revisit this thread when I get it running again.

Well, if you get a 1200XL and do the SV 2.1 upgrade for it, you can have chroma-boost that looks pretty  rich, with no color-bleed, almost as rich as RGB. But of course do to scanlines and half being skipped in 240p images, it is half-bright, half color  saturation on a 480i CRT. But of course you can help it a bit with the brightness, contrast and color controls before you start bleeding or washing out. Just with straight S-video on a good CRT like the one you have. I had brilliant color with the same computer on my 1084S, and only started using LCD TV's and a video2VGA converter in 2016 when I first finished repairing my 1200XL after being broken and in storage since 2012. And I was still using the 1084S, for the first few weeks, and then I had a game running in "attract" mode and left to eat dinner. I came back and there was no power to the 1084S. I just put it aside until recently, I've started rebuilding the power board now. Once I finish, it will be the monitor for this 1200XL of mine again, since it handles both NTSC and PAL 50/60HZ perfectly, and even Atari's with different PAL/ANTIC chip hybrids. My 800 will get the video2vga converter on an LCD, since it is NTSC, and only gets color on my LCD's if is converter through my converter anyway. My new, "cheap" HDMI converter won't recognize the 800's chroma signal either! 

 

Anyway, I have been assuming that video upgrades like the UAV and Sophia also vastly improve the chroma saturation on all the XL/XE's that don't have the 1200XL's chroma boost circuit? Isn't there a chroma-pot or something on the UAV board? Is it only the RGB versions that improve chroma saturation or do they not do it either?