Improper colors

[dshadoff]

Check out the one in the archive attached in the previous post from an hour or so ago.

[turboxray]

Sept 26, 2020 15:50:18 GMT elmer said:

Sept 26, 2020 15:00:19 GMT dshadoff said:

You have a MiSTer now ?  Or are you just watching the repositories ?

First, let me give a brief overview of what's happened in the past 2-3 months:

...

As for the question of "final"... I think we are. There are still measurements which would be nice to have, but they are merely expected to confirm what we believe.

Nope, I don't have a MiSTer, I just saw turboxray posting about the new palette on AtariAge and so took a look at the MiSTer repository.

Thanks for the update!

I look forward to the new palette appearing in some future Mednafen update, but even before that I'd like to try it so that I can see if my small tweeks to LoX UI colors need more tweeks, or if they should just be entirely removed.

You can use the palette directly with mednafen (if it's in the right format), but it would be nice for mednafen to have an option (or default to it) rather than needing to copy the palette file into the palette folder.

[nicole]

Question to people working on this: I don't use Mednafen as my preferred emulator, but it looks to me on my SuperGrafx that color 0b000000001 is perfectly equivalent to black, whereas in RGB it's a dark blue. Does that match what people are seeing on consoles, or is it possible that my SuperGrafx/SuperGraficeses in general are off in some way?

[turboxray]

Oct 10, 2020 18:18:27 GMT nicole said:

Question to people working on this: I don't use Mednafen as my preferred emulator, but it looks to me on my SuperGrafx that color 0b000000001 is perfectly equivalent to black, whereas in RGB it's a dark blue. Does that match what people are seeing on consoles, or is it possible that my SuperGrafx/SuperGraficeses in general are off in some way?

It depends. If you're talking about composite, then yeah blue of 001 is going to be darker on the composite signal vs RGB outputs of the VCE simply because the YUV map gives less brightness for it. It shouldn't be identical to black though. If it is, your set or whatever is adjusted too low (increase the brightness some. Use the 240p suite color bars to help adjust it).


Also, here's the final palette file for mednafen: pce.pal (1.5 KB)
Just place it in the 'palettes' subfolder. Luma and saturation have been scaled to be optimal, so it's not going to be as desaturated as your composite signal and doesn't have the NTSC interpreting NTJC-J signal low-end gamma crush either. 

[dshadoff]

Good timing, I was just preparing to updates a few threads today as well.

For those who want to review the old RGB values versus Y/B-Y/R-Y lookup values in the 6260, versus the new calculated RGB derivatives of those, I assembled it all into a spreadsheet table, here:

RGB-YUV-RGB' Table

[tatsuya79]

Thanks for all your work on this people, that's amazing what we can have now.

Here is the difference between raw rgb and composite color + ntsc svideo shader with crt (Ankoku Densetsu level 2):

   ___

Even on the logo the ntsc signal can change the result:

rgb colors    composite colors    composite+ntsc

[Black_Tiger]

Nov 28, 2020 11:50:49 GMT tatsuya79 said:

Thanks for all your work on this people, that's amazing what we can have now.

Here is the difference between raw rgb and composite color + ntsc svideo shader with crt (Ankoku Densetsu level 2):

View Attachment   View Attachment___

Even on the logo the ntsc signal can change the result:

rgb colors    composite colors    composite+ntsc

It would be interesting to see that color banded column that Chris Covell talked about and might have made a page for way back when.

It's at 21:38 in this video:

He was talking about techniques to get more shades and another example he used was attempting to port the title screen of Magician Lord.

[ndiddy]

Out of curiosity, does the official NEC stuff that outputs RGB (the Turbografx arcade board, the monitor with a built in PC Engine, etc) use the RGB outputs on the VCE without further modification?

[tatsuya79]

Nov 30, 2020 4:13:00 GMT Black_Tiger said:

It would be interesting to see that color banded column that Chris Covell talked about and might have made a page for way back when.

I don't have a save there but it mixes the color together and the banding is a lot less apparent.

A bit like here on my last save state, columns behind the final boss:

  

[tatsuya79]

Stage select code to the rescue.

  

The banding is still a bit visible.

[SignOfZeta]

Nov 30, 2020 5:22:49 GMT ndiddy said:

Out of curiosity, does the official NEC stuff that outputs RGB (the Turbografx arcade board, the monitor with a built in PC Engine, etc) use the RGB outputs on the VCE without further modification?

I can’t say for sure, but in the case of the monitor we can assume they didn’t Google “RGB mod PCE” and in fact actually designed a monitor with a PCE in it in which case it would be designed to pass a test pattern or something. I’d assume it’s a very quality job from 80s NEC. I’m assuming that, had the RGB booster been released, that the same thing would have been done there. (No promises!) The “tweaking” would have been done with the support components, not digitally. The same is probably true of the JAMMA version. In both cases consider that it’s totally normal to have level pots for R, G, and B so it would calibrate fine in the end regardless.

It’s also possible that the PCE monitor is as multi-sync as most old NEC monitors were and that it just fed it composite internally like the Laseractive but I’ve never seen one.

EDIT: some further looking has lead me to believe that the PC-KD863G did indeed show you PCE in RGB and is in fact the only official way to do so. It would be interesting to see the PCB...

[tatsuya79]

Sorry to hijack this thread again, just wanted to tell I made a 100% shader solution as a preset to get a composite signal simulation that wouldn't be too blurry.
It's using themaister ntsc adaptative shader so it will work with pce, nes, snes and megadrive too (the dithering gets merged to make transparency).
It's using a sharpen pass and a luminance morphing pass to enlarge bright pixels slightly (to fight the black edges getting too thick with enlargement).

It's in RetroArch slang shaders folder, shaders_slang\presets\crt-geom-ntsc-composite-sharp.

 

[tatsuya79]

...and the famous Startling Odyssey II grass.
It's both composite frames merged together using Photoshop. They alternate/flicker in reality.

 

[nicole]

I hope it's alright to bump the thread a bit.

I've been looking at my Pioneer LaserActive, and it looks like the PAC N-1's RGB signal is directly used to create the composite output there. Here's the thread on Twitter which has some screenshots:

So that's at least one case where the RGB palettes were the ones seen by the end users; for the tiny number of people who bought a LaserActive, anyway. (And really, don't they deserve worse colors)

[elmer]

I have finally checked out the new Legend of Xanadu 2 maps from the English translation with the more-accurate palette, and the differences are small, but interesting to see.

As drawn by Mednafen, using the old RGB-inspired palette ...




As drawn by Mednafen, using the new more-accurate palette ...