PAL Users - Need Your Input For 7800 Palette Implementation

[DEANJIMMY]

1 hour ago, Trebor said:

Makes sense - a very tiny correction counter-clockwise can still shave off a full degree of hue spacing (The color pot is typically very sensitive), so instead of warm (~26.7 degrees), it is set to cool (~25.7 degrees) now.

For the 7800 PAL Scart I expect most of them were wrong right from production. The video boards were added late in production by hand and the soldering does not look very profesional. So, they added the "Mod" and did not take care of fine tuning.

[Trebor]

6 hours ago, DEANJIMMY said:

For the 7800 PAL Scart I expect most of them were wrong right from production. The video boards were added late in production by hand and the soldering does not look very profesional. So, they added the "Mod" and did not take care of fine tuning.

I'm also thinking PAL systems may not have been necessarily wrong, per se, but different in another way.

 

NTSC systems specifically are furnished a Diagnostic Test Cartridge which checks for the encryption (which is absent from PAL) BIOS and then directions on how to set the color pot.  The effect has Hue 1 match Hue 15 (AKA Hue $Fx).  This happens when adjacent hues are set 25.714 (~25.7) degrees apart from each other, or what is often referred to as a "COOL" palette.  System warm up typically increases that by ~1-2 degrees.  That's how we derive at WARM palette (~26.7 degrees), and a HOT palette (~27.7 degrees).  Many NTSC consoles will ultimately end up as HOT, with a few being WARM.

 

PAL systems have no such cartridge.  If we are to evenly space adjacent hues from one another, the hues would be 24 degrees a part.  The reason for this is the color wheel is 360 degrees in total and there are 15 color hues (Plus a grayscale - the Hue $0x range).  360/15 = 24.  So if I set 24 degrees or about it to start (I.E. ~23.7 degrees), and the system 'warm-up' occurs, following the same pattern of ~2 degrees increase, the end result would be a adjacent hue separation of ~26 degrees (~25.7 degrees).  This would indeed provision what is interpreted as a "COOL" palette for PAL systems as the ultimate result.

 

Considering the color potentiometer is extremely sensitive, variances of ~0.5 - 1.0 degrees among systems can be expected as well, whether PAL or NTSC.

[pseudografx]

I wonder why the warming-up issue never occured on the Atari 8-bit line. I am not aware of any noticeable colour shifting on the 8-bit.

[Trebor]

10 hours ago, pseudografx said:

I wonder why the warming-up issue never occured on the Atari 8-bit line. I am not aware of any noticeable colour shifting on the 8-bit.

There is no warm-up factor for PAL systems in the 8-bit computer line. 

 

According to the Altirra Hardware Reference manual, NTSC 8-bit computers adjacent hues can vary from 23-26 degrees apart from each other; however, PAL 8-bit computers adjacent hues are hardcoded to 22.5 degrees of separation.

[pseudografx]

Oh, I see, thanks for the explanation.