Stella Gets Some New Clothes

[Thomas Jentzsch]

Another approach: Maybe we can learn from the NTSC to PAL conversions.

 

In Sorcerer's Apprentice 1x/Fx have been replaced by 2x/5x, which might indicate that Fx was greenish.

 

However in Sprintmaster the Fx on the title has been replaced by 4x! Same for the pyramids in Desert Falcon or the bottom of the first screen in Jungle Hunt.

In Dark Chambers Fx is used for the hero's head skin (and replaced by grey in PAL). Same for Mario's shoes in Mario Bros. This would indicate some brownish/reddish color.

 

In Oscar's Trash race the trash bins' Fx has been replaced by 6x(!) in PAL. Same for the canyon in Save Mary.

 

Overall there is a tendency to are more brown/red Fx. But this has to be verified...

Edited September 30, 2013 by Thomas Jentzsch

[SeaGtGruff]

Well, the PAL palette on the 8-bit Atari computers definitely starts with a brownish-yellow hue 1 rather than the NTSC's greenish-yellow hue 1, and hue 15 is decidedly brownish. But the PAL 8-bit Atari computers have all 16 hues, arranged in their "normal" order, so I don't know how helpful it is to compare their palette with the PAL Atari 2600's.

 

On the other hand, the NTSC Atari 8-bit computers have a palette that's either the same as, or darn close to, the NTSC Atari 2600's palette (the color sections of the TIA and GTIA schematics should give an indication of how similar they are), and we have several screenshots from actual NTSC and PAL computers that people submitted a while back in one of the 8-bit forums-- either the main forum or the programming sub-forum. Those screenshots do show some variation in hue 15.

 

Of course, the existence of the color pot adjustment pretty much makes it all moot or academic. It would be great to know what hue 15 was "supposed" to look like, but AFAIK the only document we have from Atari that describes how it should actually be "calibrated" (as opposed to just describing its color as "gold" or whatever) is the 2600 field service manual, although that description sounds like a rule of thumb to me. That is, it's my understanding that even a miniscule turn of the color pot adjustment can produce dramatic changes in hue 15 (I've never tried it myself), so telling the field service technician to turn the adjustment until hue 15 is within a shade of hue 1 sounds like a way to say "get it somewhere in the ballpark of between hue 14 and hue 1, or between hue 1 and hue 2," since it's probably nearly impossible to make it match hue 1 exactly.

 

I wonder if any of the surviving original Atari engineers or programmers could shed any light on this question? Calling Allan Alcorn! Calling David Crane!

[Thomas Jentzsch]

That is, it's my understanding that even a miniscule turn of the color pot adjustment can produce dramatic changes in hue 15 (I've never tried it myself),

The funny thing on a PAL console is, that the pot seems to affect 2x most. I can get from greenish 2x to brownish 2x within a very few degrees. Maybe the pot works different on PAL?

[Thomas Jentzsch]

It would be great to know what hue 15 was "supposed" to look like, but AFAIK the only document we have from Atari that describes how it should actually be "calibrated" (as opposed to just describing its color as "gold" or whatever) is the 2600 field service manual

Not all documents are equal!

 

In this document I found:

 

 

(The bars should be goldenrod in color.)

 

Funny enough this document is from November 1982. The other one I found (without that sentence) is from January 1983.

 

BTW: The Diagnostic Cartridge for PAL shows a completely different color arrangement. Do we have a Field Service Manual for PAL?

[Trebor]

Hi SeaGTGruff,

Thanks for the input and being a part of the discussion. No matter what there is going to be much variety and disagreement with color palettes. My goal here is to create the 2600 color palette within reason from both facts and observation as it shows on a CRT.

Hue 1 is very finicky for both the 2600 and 7800. From two different 4-Switch Woody - one composite mod, one RF orignal, on the following CRT displays: Sony (Digital Controls) 27", RCA 19" (Analog), Panasonic 27" and Panasonic 36", along with a Commodore 1702 (Composite mod system only), shows Hue 1 as a 'gold'. More brown on the dark side, yellow on the light.

The same console systems on a LCD Panasonic 32", Pansonic 50" Plasma, LED Insignia 24", LCD Insignia 19", LED LG 23" (I might have some of the LCD/LED labels mixed-up), Hue 1 is a greenish-yellowish color.

From my experience, hue 15 for the NTSC systems can start green, becomes a olive green, to finally reaching a brownish-gold color. For the 7800, hue 15 starts green, but becomes brown very rapidly. The phase shifts and associated explanations I posted earlier goes into this in more depth.

In case there's confusion, I didn't reference any 8-bit Atari computers. I don't own a 5200 either. I have a ColecoVision though (Don't hate)

It's times like these I really wished Atari had a 'hard' palette. Then again other systems are not immune to these color difference issues either. The most standout-ish in my mind is the NES and the color of Super Mario Bros. 'sky'. Some televisions will show it as a crisp bright blue (having more green), others come across more purplish (having more red) in the color.

I don't discount any as wrong, just different. There are already 2600 NTSC palettes which lean 1x more greenish-yellow for whatever the reasons, and I can appreciate that. I'm not stating these are better or correct and the others are wrong, but another option for 2600 emulator users.

If their findings are the same for what these palettes provide matching my 4-switch Woody's on CRTs (along with some additional variety), that's fantastic; I filled a gap missing for them. While they may not be everybody's cup of tea, hopefully what I am seeing and experience are found to be suitable by others.

I mentioned analog controls over digital for some of the CRT just as a reminder to myself and sort of a footnote for others that the analog hue/tint dial on CRT's seem to be much more sensistive (Like the 2600/7800 pots), than digital controls/adjustments.

Ideally, Hue 1 and 15 should match with a 25.7 degree phase shift (180 degree colorburst) in place. However, the 2600 and 7800 systems start phased lower than that, and at least from the 2, 4-switch Woody's, phase much slower than the 7800. The 7800's (six of them), on the other hand, phase quickly to at least 26.7 degrees over time with some systems phasing even higher than that, ~27.7 more or less.

What I have for reference, the PAL begin point for color palettes is ~57 degrees. Under NTSC, the CRT begin point is ~33 degrees, and the begin point for NTSC LCD/LED/Plasma is ~45 degrees.

My friend who helped provide many of the tools I use went through the math and reasons with me a while back, explained everything, but honestly, quite a bit went over my head.

The beginning hue 1$ look like this (When in 'normal' order (i.e. 7800)):
PAL:


NTSC via LCD/LED/PLASMA:


NTSC via CRT:


Specifically with the 7800, interestingly enough, the numbers 'worked out' as the difference between the NTSC and PAL colors is essentially one row, or approximately 24 degrees. 57-24 = 33. That does stick out in my mind like a sore thumb...lol.

There is more, but I really would need go sift through all the notes, charts, and chew my friend's ear for awhile (Or more like him chew mine) in trying to re-explain everything.

As I mentioned earlier, I am working on the 27.2 and 27.7 phase shifts. From Thomas' feedback 26.7 looks like it is very close, and I'm figuring 27.2 or 27.7 will nail it or at least a combination of 26.7 and the one of the others for any shifting that occurs which is nonlinear for him.

Thanks too for the information on the brightness. I'm going to go forward and post the alternate sets as well in case more fall under Thomas' umbrella of the way grayscale and colors interact and compare with each other.

[Trebor]

 

Another approach: Maybe we can learn from the NTSC to PAL conversions.

 

In Sorcerer's Apprentice 1x/Fx have been replaced by 2x/5x, which might indicate that Fx was greenish.

 

However in Sprintmaster the Fx on the title has been replaced by 4x! Same for the pyramids in Desert Falcon or the bottom of the first screen in Jungle Hunt.

In Dark Chambers Fx is used for the hero's head skin (and replaced by grey in PAL). Same for Mario's shoes in Mario Bros. This would indicate some brownish/reddish color.

 

In Oscar's Trash race the trash bins' Fx has been replaced by 6x(!) in PAL. Same for the canyon in Save Mary.

 

Overall there is a tendency to are more brown/red Fx. But this has to be verified...

If 2600 NTSC to PAL conversions is anything like 7800 NTSC to PAL conversions it will be a very mixed bag. Some games, you can hardly notice any difference (Dig Dug), then others they get some colors close to matching between regions, but some are really off from their NTSC counterparts at least for a few or more choices (Midnight Mutants, Donkey Kong).

 

Some 7800 PAL games actually had much better colors picked for them then their NTSC counterparts (Ms. Pac-Man stands out the most in my mind at this time).

[SeaGtGruff]

Hue 1 in the NTSC palette shouldn't change no matter how much you turn the color pot adjustment or how cold/warm/hot the machine is, because it's used for the color burst and hence should always have the same hue as color burst (that sounds like double-talk but it actually isn't). The *only* reason that hue 1 should ever look different is because of the "tint" setting on a TV, or similar sorts of settings on a monitor (such as the amount of red, green, and blue).

 

Interestingly enough, the color strips you show include hue 1 for PAL, which I presume is for the PAL 7800, since hue 1 on the PAL 2600 is black-and-white. The strips for PAL and NTSC look backwards to me-- the "PAL hue 1" looks close to the NTSC color burst, whereas the "NTSC hue 1" looks more like what I've seen for hue 1 in screenshots taken from PAL Atari 8-bit computers.

[SeaGtGruff]

The funny thing on a PAL console is, that the pot seems to affect 2x most. I can get from greenish 2x to brownish 2x within a very few degrees. Maybe the pot works different on PAL?

 

Which hue changes the least?

[Trebor]

Hue 1 in the NTSC palette shouldn't change no matter how much you turn the color pot adjustment or how cold/warm/hot the machine is, because it's used for the color burst and hence should always have the same hue as color burst (that sounds like double-talk but it actually isn't). The *only* reason that hue 1 should ever look different is because of the "tint" setting on a TV, or similar sorts of settings on a monitor (such as the amount of red, green, and blue).

 

Interestingly enough, the color strips you show include hue 1 for PAL, which I presume is for the PAL 7800, since hue 1 on the PAL 2600 is black-and-white. The strips for PAL and NTSC look backwards to me-- the "PAL hue 1" looks close to the NTSC color burst, whereas the "NTSC hue 1" looks more like what I've seen for hue 1 in screenshots taken from PAL Atari 8-bit computers.

 

Agree Hue 1 never changes (Without manually tint/hue intervention). It does though look different on a HDTV than a traditional CRT, more around -45 (much greener) degrees for an LCD/LED/Plasma, as opposed to ~-33 degrees (more browner/gold) for a CRT.

 

Correct, the strips were for the "i.e. 7800"

 

PAL 1$ hue in the strip is set to -57 degrees. That is the begin hue point of PAL from my understanding, and should not match or be close to NTSC (At least not on a CRT).

 

The NTSC CRT hue 1 (-33 degrees) is indicative of 7800 and 2600 NTSC consoles on the CRTs I listed. I'm not using or trying to simulate anything from Atari 8-bit computers; regardless of similarities...Sorry for any confusion.

[Trebor]

Hey SeaAtGruff,

 

Don't want you to think I'm dismissing or denying your or anyone else's experience with 1x being more greenish with an NTSC console. You may very well have a 4-switch Woody 2600 displayed on CRTs that are showing 1x as more greenish-yellow than gold. I can only tell you what I experience and witness and hopefully it also may seem reasonable to you some of the logic utilized.

 

Going back to the Hue 1 and 15 match, as mentioned earlier, I agree with you the Hue 15 (F) should be more brownish.

 

If both consoles fell and stayed at phase shift 25.7 degrees - the ideal exact 180 degree color burst - then they both would shine golden not green under NTSC with a CRT. If hue 1 was more greenish-yellow, at an exact color burst 180, it would mean hue 15 would ultimately be and stay greenish-yellow too.

 

While the 7800 phase shifts much faster than the 2600 (At least with the consoles I own), they both look to push Hue 15 brown and tech docs for both systems show trying to get 1 and 15 matching as close a possible.

 

It doesn't take much for variation; I do see a greenish-yellow Hue 1, but only under LCD/LED/Plasma displays. All the CRTs I've seen have it as a 'gold' color. The worse I can say is these palettes won't be good for you, and something you don't experience. As mentioned before, I'm just putting out there 2600 palettes that match (As close as possible) what I see on the real hardware on CRTs. More greenish-yellow hue 1x palettes are already out there to be utilized for those whose system and displays match that.

 

Thomas has been gracious enough to give his input on PAL, as I am trying my best to ultimately provide to him what he experiences, while also covering the corresponding PAL palettes for the NTSC ones created.

 

Sorry if some of my points were/are confusing or I am not explaining myself clearly. However, a greenish-yellow hue 1, especially one that looks close to the PAL strip I posted at 57 degrees, is nowhere near what I see with my 4-Switch Woody's on CRTs. I do see a medium between the two for Hue 1 (~45 degrees) on LCD/Plasma/LED; and I may create those as well down the road. CRT and the 2600 is my focus now. You see and experience differently, and I respect that.

 

Thanks again for replying and contributing. It is appreciated.

[SeaGtGruff]

I agree with your comments/observations about hue 15 and the way it changes as the machine warms up, shifting from more greenish to more brownish.

 

As for hue 1, my main concern was that I wanted to make sure you hadn't somehow gotten your NTSC (YIQ) and PAL (YUV) formulas crossed, since your NTSC 7800 hue 1 looks like what I'd expect for PAL, and your PAL 7800 hue 1 looks like what I'd expect for NTSC. But aside from the issue of the tint adjustment on a TV, or the RGB adjustments on a monitor, there can be differences between monitors. I have two monitors side-by-side at work, with the Windows desktop split between them (left half, right half). They're two different brands of montors, and I've got them calibrated exactly the same, yet the colors on one do *not* match the colors on the other-- most easily seen by positioning a program window so it's partly on the left monitor and partly on the right-- and no matter how much I try to adjust one to match the other, I can't get them to match, therefore I just calibrate them the same and ignore the color differences. So yes, the colors will look different on different types of display devices, as well as on different display devices of the same type.

 

To change the subject, what I'm curious about right now is how you're calculating the PAL 2600 colors. I've never tried to calculate the PAL colors because I haven't seen any posts or documents that explain how they "should" be calculated. But based on what Thomas said about hue 2 changing the most when he fiddles with the color pot adjustment, I've been trying to figure out how it works, and the best idea I can come up with so far is as follows:

 

Hue 13 = 0 x phase shift delay

Hue 12 = 1 x phase shift delay

Hue 11 = 2 x phase shift delay

Hue 10 = 3 x phase shift delay

Hue 9 = 4 x phase shift delay

Hue 8 = 5 x phase shift delay

Hue 7 = 6 x phase shift delay

Hue 6 = 7 x phase shift delay

Hue 5 = 8 x phase shift delay

Hue 4 = 9 x phase shift delay

Hue 3 = 10 x phase shift delay

Hue 2 = 11 x phase shift delay

 

Or it might start with hue 15 = 1 x phase shift delay and go up from there to hue 2 = 12 x phase shift delay.

 

This actually makes sense to me, because if we were starting with hue 2 (color burst) and hue 3 (alternate color burst) and going up from there, fiddling with the color pot adjustment should leave hues 2 and 3 more or less untouched, with hues 12 and 13 changing wildly, and with the possibility that the even hues could be overlapped with the odd hues and vice versa, as follows:

 

Hue 2 = Yellow

Hue 4 = Orange

Hue 6 = Red

Hue 8 = Purple

Hue 10 = Blue

Hue 12 = Cyan

 

Hue 3 = Green

Hue 5 = Cyan

Hue 7 = Blue

Hue 9 = Purple

Hue 11 = Red

Hue 13 = Orange

 

Also, we'd never be able to adjust the color pot to get any colors in the 90-degree area between the two color bursts (135 to 180 to 225 degrees).

 

But if hue 13 is phase-shifted the least, and hue 2 is phase-shifted the most, as suggested further above, then we could potentially adjust the color pot to get colors in that 135-to-225-degree area.

 

Unfortunately, I can't wrap my head around how that would work with the color bursts.

 

Furthermore, it's possible that the reason hue 2 "changes the most" is because it's the hue where small changes to either side are most easily noticeable-- i.e., go a little bit "left" or "right" of the other colors and you're still in the reds, or blues, or greens, etc.; but go a little bit "left" or "right" of yellow and you get orangish-yellow or greenish-yellow, which look quite different from each other.

 

I really wish we had some PAL TIA schematics to look at!

[Thomas Jentzsch]

Which hue changes the least?

I will test again tonight. But IIRC the blue to green colors change least.

 

BTW: I am still waiting for an answer to the question why for NTSC the phase step multiplied with the colors result into ~360°. But for PAL the same phase step is used, and with 3 colors less, we end below 290°. Aren't the PAL colors not meant to go full circle? On my console they easily can do.

Edited October 1, 2013 by Thomas Jentzsch

[SeaGtGruff]

I will test again tonight. But IIRC the blue to green colors change least.

 

BTW: I am still waiting for an answer to the question why for NTSC the phase step multiplied with the colors result into ~360°. But for PAL the same phase step is used, and with 3 colors less, we end below 290°. Aren't the PAL colors not meant to go full circle? On my console they easily can do.

 

Hmm, that sounds like maybe green is the starting point:

 

Hue 3 = 0 x phase shift delay

Hue 5 = 1 x phase shift delay

Hue 7 = 2 x phase shift delay

Hue 9 = 3 x phase shift delay

Hue 11 = 4 x phase shift delay

Hue 13 = 5 x phase shift delay

Hue 12 = 6 x phase shift delay

Hue 10 = 7 x phase shift delay

Hue 8 = 8 x phase shift delay

Hue 6 = 9 x phase shift delay

Hue 4 = 10 x phase shift delay

Hue 2 = 11 x phase shift delay

 

Or maybe hue 3 is 1 x phase shift delay, up to hue 2 being 12 x phase shift delay.

 

As far as your question, I don't know how Trebor is calculating the PAL colors, but if he's assuming that hue 2 is the 135-degree color burst and hue 3 is the 225-degree color burst, that's a difference of 270 degrees. If there are 12 hues, that's 11 steps, and 270 / 11 is about 24.5 degrees per phase shift.

 

However, if the pattern follows the above layout, adjusting the color pot should allow you to move hue 2 closer to hue 3-- 360 / 11 or about 32.7 degrees per phase shift.

 

What I don't understand is why Atari didn't go for 15 colors:

 

Hue 1

Hue 3

Hue 5

Hue 7

Hue 9

Hue 11

Hue 13

Hue 15

Hue 14

Hue 12

Hue 10

Hue 8

Hue 6

Hue 4

Hue 2

 

Maybe that wouldn't work because it's 8 down and 7 up, and they needed the sides to be equal-- but if so, they could have left out hue 1 (as they did) and still gotten 14 colors. And they apparently didn't have trouble getting 15 colors on the PAL computers or PAL 7800.

 

The PAL TIA schematics would hopefully give us some answers, or at least let us see what they were doing.

[Trebor]

Hi Thomas,

As promised here are PAL 27.2 and 27.7 phase shift palettes. Please provide feedback when able regarding what you see on your Vader and Sony compared to these palettes. It is greatly appreciated.

 

Both Composite and RGB are provided below...


PAL 27.2:

 

PAL 27.7:

 

[Thomas Jentzsch]

Can you post the palette files too, please? Or the ordered palette from my ColorTest program.

Makes it easier to compare. I can hardly see any difference between 27.2 and 27.7.

 

Preliminary results:

• The red to blue colors look about OK.
• The switch from blue to green starts too early. Bx is completely blue and 9x has just a minimal green tint.
• 7x is about OK, a bit too much green (need the palettes to compare better), 5x looks good, 3x has less yellow (looks almost like 5x).

Edited October 1, 2013 by Thomas Jentzsch

[Trebor]

Sorry, meant to attach them after the captures...

 

Here is the complete NTSC & PAL Stella set running from Phase Shift 24.7 to 27.7 with 0.5 increments:

Stella2XX_Palettes_20131001.zip

[Thomas Jentzsch]

 

Which hue changes the least?

 

Blue, Dx

 

Also I don't think the PAL pot works like the NTSC pot. To me it looks like it defines the green mix. I can make almost all colors pretty greenish, except 3-5 colors (the ones around Dx).

Edited October 1, 2013 by Thomas Jentzsch

[SeaGtGruff]

 

Blue, Dx

 

Also I don't think the PAL pot works like the NTSC pot. To me it looks like it defines the green mix. I can make almost all colors pretty greenish, except 3-5 colors (the ones around Dx).

 

Okay, if hue 13 (blue) changes the least, it sounds like the order may be as I suggested in post 36-- hues 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, and 2. I'm not prepared to calculate an actual PAL palette yet, but the following illustrations show how I think it may be set up and what different phase steps might look like. I'll give similar illustrations for the NTSC setup first, for comparison.

 

The first illustration is the color section from the NTSC TIA schematics:

 

 

Here is the YIQ/YUV color cube, with dots positioned at phase steps of 28 degrees, beginning at color burst. The colors wrap around the wheel a bit-- hue 14 is close to hue 1, and hue 15 is close to hue 2:

 

 

And here are the corresponding hues at the edges of the color cube:

 

 

Here is the color cube again, but with dots at steps of 20 degrees. The colors don't go all the way around the wheel:

 

 

And here are the corresponding hues:

 

 

The colors are more "pure" than the 2600 produces, but the above hue strips should give some idea of how turning the color pot adjustment will change the NTSC colors.

 

Now here's a simplified picture of how I think the PAL TIA schematics *might* look. It's patterned after the NTSC TIA schematics but with the hues rearranged to try to match what you said you're seeing-- that the high-numbered hues change the least, and the low-numbered hues change the most, which suggests that the hues are in a reverse order:

 

 

Here's the color cube with dots at steps of 20 degrees, but this time beginning at the far right-- directly opposite the NTSC color burst, or along the U axis. Going counter-clockwise from that point, the black dots represent the hue steps. But since the PAL color burst alternates between 135 and 225 degrees, I'm assuming that every other step position is reflected onto the opposite side of the wheel-- i.e., minus degrees instead of plus degrees-- represented by the red dots. As a result, the hues end up wrapping around the wheel at the -U end:

 

 

Here are the corresponding hues:

 

 

Finally, here's the color cube with steps of 12 degrees. This time the hues don't go all the way around the wheel:

 

 

And here are the corresponding hues:

 

 

I don't know if this model is correct, but hopefully you can verify whether the behavior you're seeing when you adjust the color pot is consistent with this hypothetical model. And I have no idea how the PALI and PALS pins of the PAL TIA chip fit into this hypothetical picture, and/or how the phase angles as I've drawn them (originating at 0 degrees) are correlated with the alternating PAL color burst signals.

[Thomas Jentzsch]

I have to admit I don't completely understand everything. But the bottom hues are about right.

 

And the hues above those describe what happens when I turn the pot too much towards green pretty well too.

 

BTW: The opposite pot direction mixes up the colors.

[SeaGtGruff]

I have to admit I don't completely understand everything. But the bottom hues are about right.

 

And the hues above those describe what happens when I turn the pot too much towards green pretty well too.

 

BTW: The opposite pot direction mixes up the colors.

 

So it sounds like my hypothetical model might be close to the way it actually works.

 

It would be great if you could post some actual photos showing the screen after turning the pot different directions/amounts, especially some of the more extreme settings. That should indicate, for example, whether hue 13 is always constant-- which would mean it's somehow being used as the "base"-- and how the other colors are being shifted from that base (be it hue 13 or something else).

[Thomas Jentzsch]

I will post some pictures with different pot settings tonight. Fingers crossed my camera doesn't get the colors completely wrong.

 

The pot is very sensitive. Between greenish and brownish 2x there are just ~5° degrees.

[Trebor]

That last PAL bar looks very similar to the Stella247 palette, just with stronger saturation.

 

As I mentioned previously, from my NTSC 2600 systems, the palette looks to fall around Stella247 (Cooler system on for a little time) and Stella252 (Warmer system on for a while). I thought PAL systems may behave similarly.

 

Here's the Stella247 palette with different saturation levels under PAL:

Stella Default (50) --> 75 --> 100

 

 

Your monitor, video card, and of course system and (CRT) television sets will have differences amoung them.

 

Adding stronger/weaker saturation, brightness, contrast all becomes very subjective and can be handled via Stella's video controls.

[SeaGtGruff]

I will post some pictures with different pot settings tonight. Fingers crossed my camera doesn't get the colors completely wrong.

 

The pot is very sensitive. Between greenish and brownish 2x there are just ~5° degrees.

 

If my theory about how the PAL colors are generated is correct, then the pot *should* be much more sensitive than on the NTSC 2600. If you study the vectorscope graticules where I put the black and red dots around the wheel for PAL, you'll notice that you can cover essentially twice the distance around the wheel with a particular phase shift size and number of phase shifts as with NTSC, due to the way half of the hues are being reflected to the other side of the wheel (i.e., you're covering twice the ground with the same number of steps).

 

On the other hand-- assuming my theoretical model is accurate (and until it's demonstrated to be accurate I'll keep calling it theoretical)-- the distance from hue 13 to hue 12 will be half the distance between the other hues. If you have trouble grasping that (studying those vectorscope graticules should help), think of it in terms of the hue numbers-- from hue 13 to hue 12 is 1 step (13 - 12), but for the other hues it's 2 steps (13 - 11, or 11 - 9, or 12 - 10, etc.).

 

So if you want to go all the way around the circle as evenly as possible, such that hue 2 and hue 3 are the same distance apart as the other hues are (with the exception of hues 12 and 13), you'd want the phase step to be about 15.65 degrees, or 360 / 23 (i.e., 11 hues spaced 2 steps apart, but from hue 12 to 13 is only 1 step, so 2 x 11 + 1 = 23).

[Trebor]

Here is 'Stella247A' under PAL having the same phase shift as Stella247 in the previous postings, but utilizing "x1, x3, x5, x7, x9, xB, xD, xF" as the 'brightness' columns for color, instead of "x0, x2, x4, x6, x8, xA, xC, xE".

 

Here are the various saturation models included below like the prior post:

(50) Default --> 75 --> 100

 

 

And here is one with a gamma and saturation boost (Again, all controlled under Stella's video options):

 

 

Here is the updated set: Stella2XX_Palettes_20131002.zip

 

Next set posted will have all the 'alternate' columns posted for each of the seven phases, but I probably won't get to it for several days.

 

Those are awesome models and work with your approach to the palettes, SeaGtGruff; very cool.

[Thomas Jentzsch]

This time I got you.

 

And you could be very well right with your half-step theory around the hues 12 and 13.