Questions about color encoders in Nestopia and Sega Li

Discuss emulation of the Nintendo Entertainment System and Famicom.

Moderator: Moderators

strangenesfreak
Posts: 155
Joined: Thu May 03, 2007 3:07 pm
Contact:

Questions about color encoders in Nestopia and Sega Li

Post by strangenesfreak » Mon Jul 21, 2008 8:25 pm

  • How uncommon is Nestopia's "15° Canonical" encoder in consumer TVs? Is it uncommon enough to be generally less accurate to the TV experience than "Consumer" encoders?
  • Nestopia's "Alternative" encoder boosts yellow, which I have read to be an effect from NTSC-J, so is it supposed to represent a general NTSC-J encoder? If so, why is the yellow boost much stronger than Sega Li's "Sony CXA2025AS/J" and "Sony CXA2095S/J" encoders? Perhaps they were made for different monitor color temperatures?
  • I am assuming that Sega Li's "industrial" encoder is supposed to be canonical to NTSC standards (correct me if I'm wrong), as is Nestopia's "canonical" encoder, but why are they different? If it could be found out, which is more accurate?

User avatar
hap
Posts: 355
Joined: Thu Mar 24, 2005 3:17 pm
Contact:

Post by hap » Tue Jul 22, 2008 3:10 am

In Sega Li, the only decoder properties I use are R-Y/G-Y/B-Y angles and gains, obtained from sources online, no adjustable black level or "yellow boost" (which is probably artificial in Nestopia). I convert these properties to a matrix and feed them to blargg's nes_ntsc.

Point 3: Yes, the industrial decoder is standard NTSC, the gain levels were guessed though. Anyone knows the correct values? :)

Code: Select all

R-Y angle: 90.0, gain: 0.6
G-Y angle: 246.0, gain: 0.3
B-Y angle: 0.0, gain: 1.0

*edit*, I did some fiddling, G-Y angle is probably 236 even..

NewRisingSun
Posts: 1303
Joined: Thu May 19, 2005 11:30 am

Post by NewRisingSun » Mon Aug 18, 2008 8:44 am

How uncommon is Nestopia's "15° Canonical" encoder in consumer TVs? Is it uncommon enough to be generally less accurate to the TV experience than "Consumer" encoders?
In the U.S. until the late 90s, certainly. Japan is different.
Nestopia's "Alternative" encoder boosts yellow, which I have read to be an effect from NTSC-J, so is it supposed to represent a general NTSC-J encoder?
The only by-the-book difference between American NTSC and Japanese NTSC is the setup (7.5% U.S., 0% Japan) and the white point (D65 U.S. since the 1970s, D93 Japan). Any boosting business will be at a TV manufacturer's discretion. Best forget about it.
Anyone knows the correct values?

Code: Select all

R-Y: 90°, 0.570
G-Y: 236°, 0.351
B-Y: 0°, 1.016
Actually, the gains should be twice as high, but for some reason Nestopia requires them at half value.

User avatar
jargon
B&: This is not your blog
Posts: 208
Joined: Fri Dec 07, 2007 11:40 pm
Location: 480/85260
Contact:

Post by jargon » Mon Aug 18, 2008 7:39 pm

This thread is a little confusing.

Could someone please explain what "*-Y", "angle", and "gain" exactly means? :)
Last edited by jargon on Thu Aug 21, 2008 8:10 pm, edited 1 time in total.
Cheers,
Timothy Robert Keal alias jargon

Image
Miser's House Anthology Project

strangenesfreak
Posts: 155
Joined: Thu May 03, 2007 3:07 pm
Contact:

Post by strangenesfreak » Mon Aug 18, 2008 8:32 pm

NewRisingSun wrote:
Anyone knows the correct values?

Code: Select all

R-Y: 90°, 0.570
G-Y: 236°, 0.351
B-Y: 0°, 1.016
Actually, the gains should be twice as high, but for some reason Nestopia requires them at half value.
When I tried to enter in the doubled gain values into Nestopia's settings, B-Y's doubled gain of 2.032 was clipped to 2, which seems to be Nestopia's color decoder limit for angle gains. The luminance ranges across the different hues became oversaturated and disproportionate of each other after doubling the gains, which might also be why Nestopia halves the gains.

strangenesfreak
Posts: 155
Joined: Thu May 03, 2007 3:07 pm
Contact:

Post by strangenesfreak » Tue Sep 02, 2008 7:15 pm

  • How close to (or far from) Japanese TV yellow boost is Nestopia's yellow boost? Is the yellow boost in Nestopia generally typical or atypical to most of these TVs?
  • Why does Nestopia's yellow boost differentiate colors 20h and 30h in red, green, and yellow emphasis? Could this be a bug, or does this actually happen on Japanese TVs? I don't see why it should happen, since those two colors are supposed to be the same.
  • What is the most common color temperature for American CRT TVs - specifically those around the NES era? On modern websites, I've read 7100K or even 9300K, like Japanese TVs.
[/list]

NewRisingSun
Posts: 1303
Joined: Thu May 19, 2005 11:30 am

Post by NewRisingSun » Mon Sep 08, 2008 12:54 am

How close to (or far from) Japanese TV yellow boost is Nestopia's yellow boost? Is the yellow boost in Nestopia generally typical or atypical to most of these TVs?
Atypical. Just forget about it.
What is the most common color temperature for American CRT TVs - specifically those around the NES era? On modern websites, I've read 7100K or even 9300K, like Japanese TVs.
Most Japanese and American TVs of the 1980s use a white point of 9300K+27MPCD, meaning the same. The most common decoding characteristics seem to be something like 120°/246°/-10° with gains of 0.75/0.33/1.00 (roughly). (This will look too "red" on an sRGB monitor.)
With gains that high, most TVs will use a "relative colorimetric" rendering intent, whereas Nestopia's decoder seems to use a "saturation" rendering intent.

_Zaphod_
Posts: 4
Joined: Fri Oct 10, 2008 6:20 pm

Post by _Zaphod_ » Fri Oct 10, 2008 7:15 pm

Not atypical at the time.

Compare the RGB NES palette with the yellow boosted palette, and you will find they are VERY similar. this is evidence that the RGB NES PPUs were designed to replicate that amount of yellow boost.

MANY japanese developed games look right with that yellow boost, and look wrong with any other setting.

strangenesfreak
Posts: 155
Joined: Thu May 03, 2007 3:07 pm
Contact:

Post by strangenesfreak » Sat Oct 11, 2008 2:58 pm

_Zaphod_ wrote:Compare the RGB NES palette with the yellow boosted palette, and you will find they are VERY similar. this is evidence that the RGB NES PPUs were designed to replicate that amount of yellow boost
Maybe some of the yellow boosted hues (5-A) may look similar in hue, but the luminances look very warped in the RGB palette in comparison to a yellow boosted NTSC palette in Nestopia, especially hues 3 and 4. So I don't think the RGB palette is a good standard here.

Also, as mentioned in this thread, I found a bug in Nestopia's yellow boost that differentiate colors 0x20 and 0x30 in red, green, and yellow emphasis settings. Someone else reported this in Nestopia's bug tracker, concluding that Nestopia was calculating yellow boost based on the NES's palette values and not on the PPU's output values. Yellow boost is supposed to be directly a TV effect, not a Famicom PPU effect; if you play with the hue slider, you'll see that colors the TV should yellow boost are not boosted anymore. So Nestopia's yellow boost is probably not reliable regardless.

NewRisingSun
Posts: 1303
Joined: Thu May 19, 2005 11:30 am

Post by NewRisingSun » Sun Oct 12, 2008 5:36 am

The fact that Nintendo changed the color numbers used in Playchoice 10 games for the RGB PPU is proof that the RGB palette is not authoritative.

Given that no one has ever named a television model that actually works this way, I still think the yellow boost business is a dud. I once mentioned the idea based on a single untrustworthy source and had it as experimental code in my algorithm when blargg created his library, and that's why it's still there after all that time, unfortunately.

User avatar
blargg
Posts: 3715
Joined: Mon Sep 27, 2004 8:33 am
Location: Central Texas, USA
Contact:

Post by blargg » Sun Oct 12, 2008 6:30 am

NewRisingSun wrote:Given that no one has ever named a television model that actually works this way, I still think the yellow boost business is a dud. I once mentioned the idea based on a single untrustworthy source and had it as experimental code in my algorithm when blargg created his library, and that's why it's still there after all that time, unfortunately.
I removed "hue warping" from nes_ntsc in version 0.2.1 in 2006. I thought Nestopia always had its own extended color algorithms, which this yellow boost came from.

strangenesfreak
Posts: 155
Joined: Thu May 03, 2007 3:07 pm
Contact:

Post by strangenesfreak » Sun Oct 12, 2008 2:58 pm

NewRisingSun wrote:Given that no one has ever named a television model that actually works this way, I still think the yellow boost business is a dud.
So does that mean that yellow boost is just a myth after all? According to _Zaphod_, on certain Japanese TVs, yellow colors look like they are indeed boosted; he cited Fire Mario's bright-yellow clothes as an example:
_Zaphod wrote:Many games were developed for consumer Ntsc on the NEs/Famicom.

WHat this means is that many games orginally developed in japan were made for a yellow boosted palette, which was common on JP tvs at the time...

On a JP consumer TV, fire mario wears yellow clothes instead of white. Ths is replicated in the arcade version. ON US tvs, he wears white instead. Luigi wears white on either, though.
Maybe _Zaphod_, if he has one of these JP consumer TVs, could tell us the model of that TV to see for sure if JP TVs indeed boost yellow or not.

NewRisingSun
Posts: 1303
Joined: Thu May 19, 2005 11:30 am

Post by NewRisingSun » Sun Oct 12, 2008 10:42 pm

So does that mean that yellow boost is just a myth after all?
Yes, or "dud", as I phrased it.
According to _Zaphod_, on certain Japanese TVs, yellow colors look like they are indeed boosted
Manufacturer and model, please. Make sure you set the saturation knob to the center position when trying.
On a JP consumer TV, fire mario wears yellow clothes instead of white. Ths is replicated in the arcade version. ON US tvs, he wears white instead.
No, he doesn't; it's just a rather pale yellow. If it appears white on your monitor, your monitor is maladjusted.

strangenesfreak
Posts: 155
Joined: Thu May 03, 2007 3:07 pm
Contact:

Post by strangenesfreak » Mon Oct 13, 2008 8:21 am

I did PM _Zaphod_ if he had any of those TVs, and he said he didn't. Yeah, I could see how oversaturation could produce an illusion of this yellow boost...could it be that Japanese TVs are simply set to be more saturated than American TVs? I personally think that might be the case, because yellow does "boost" if you increase saturation in Nestopia. Naturally, since the yellowish hues are generally the most saturated hues in the NES, a saturation increase would be noticable on the brighter colors of yellowish hues, which are 7 and 8 in the NES under normal emphasis settings.

NewRisingSun
Posts: 1303
Joined: Thu May 19, 2005 11:30 am

Post by NewRisingSun » Mon Oct 13, 2008 9:31 am

could it be that Japanese TVs are simply set to be more saturated than American TVs?
Not by default, but many people I know just turn saturation all the way up on their TV, Japanese or otherwise. This over-saturating the picture only works however if your TV sets doesn't clip colors. Here's an example of what I mean:
Image
left: saturation at 100% (default). Accurate hues, but pale color 0x37.
middle: saturation at 200% with simple RGB clipping, as used in cheap TVs. Yellow "boosted" color 0x37, but the browns drift to red because of the RGB clipping.
right: saturation at 200% with hue preservation, as used in better TVs; boosted yellows while still accurate browns.

Post Reply