Nes rgb with composite colors?

[Smashbro29]

This is great but how do I actually apply them?

[CkRtech]

Smashbro29 -

Search on eBay for "altera mini blaster". Anything that says rev C on it is what you want. You need the blaster, the mini USB to USB cable and the ribbon cable. Nothing else. You will need to then modify / solder in the ribbon cable onto the NESRGB. Only 6 wires from ribbon cable are used. You'll need to cross reference blaster pin out to NESRGB one. Download software and blast away. NES needs to be powered on to program.

If you installed the NESRGB yourself, this is easy.

Good luck.

[FBX]

Smashbro29 -

Search on eBay for "altera mini blaster". Anything that says rev C on it is what you want. You need the blaster, the mini USB to USB cable and the ribbon cable. Nothing else. You will need to then modify / solder in the ribbon cable onto the NESRGB. Only 6 wires from ribbon cable are used. You'll need to cross reference blaster pin out to NESRGB one. Download software and blast away. NES needs to be powered on to program.

If you installed the NESRGB yourself, this is easy.

Good luck.

Some more information on the process can also be found on Tim's background page:

http://etim.net.au/nesrgb/background_fault/

And I added some information and the Altera Blaster pinout image on my palette project page here:

http://www.firebrandx.com/nespalette.html

[Smashbro29]

That's a shame, I got mine pre-done and can't do this stuff to save my life.

[bobrocks95]

It's soldering what, 10 pin headers? Try a soldering learners kit, it really pays to learn simple stuff like this rather than paying someone for every little tweak.

[Smashbro29]

It's soldering what, 10 pin headers? Try a soldering learners kit, it really pays to learn simple stuff like this rather than paying someone for every little tweak.

I have the iron but I keep failing.

[FBX]

I actually could use some clarification myself before I start soldering:

The encircled 3.3v pinout on the NESRGB board if I were to hazard a guess corresponds to pin 4 of the Altera Blaster, correct? Or do I leave this pin unhooked since the NES must be powered on?

Much appreciated for clarification on this!

[leonk]

I actually could use some clarification myself before I start soldering:

The encircled 3.3v pinout on the NESRGB board if I were to hazard a guess corresponds to pin 4 of the Altera Blaster, correct? Or do I leave this pin unhooked since the NES must be powered on?

Much appreciated for clarification on this!

I have it hooked up. Doesn't hurt.

[mvsfan]

that 3.3v wire tells the blaster that you have powered the nes on. You NEED to hook it up.

[FBX]

Thanks for the clarification! Got everything hooked up and installed, and it worked perfectly first try!

I installed the Playchoice 10 -- FBX Unsat. -- FBX Sat. version, and the saturated palette running on the NESRGB looks simply identical to my side-by-side test with the unmodded deck on my LCD 55" display! I understand there will be differences on a CRT display (as A.P. mentioned), but the unsaturated palette is better for that setup because you can adjust saturation on the CRT directly.

OCD is 100% satisfied, and a HUGE THANK YOU to Tim for making the alternate firmwares available!

[atheistgod1999]

Thanks for the clarification! Got everything hooked up and installed, and it worked perfectly first try!

I installed the Playchoice 10 -- FBX Unsat. -- FBX Sat. version, and the saturated palette running on the NESRGB looks simply identical to my side-by-side test with the unmodded deck on my LCD 55" display! I understand there will be differences on a CRT display (as A.P. mentioned), but the unsaturated palette is better for that setup because you can adjust saturation on the CRT directly.

OCD is 100% satisfied, and a HUGE THANK YOU to Tim for making the alternate firmwares available!

You sure it's 100%? It looks pretty good, but I noticed the terrain in Metroid looked a little too purple and red at the starting point. It looks bright blue on my CRT. I fired it up and looked at your version and what was getting output via composite to my BVM and it's definitely brighter and bluer through composite.

Could you upload pictures of the unsaturated version so I can compare it to composite better?

[leonk]

I installed the Playchoice 10 -- FBX Unsat. -- FBX Sat. version, and the saturated palette running on the NESRGB looks simply identical to my side-by-side test with the unmodded deck on my LCD 55" display! I understand there will be differences on a CRT display (as A.P. mentioned), but the unsaturated palette is better for that setup because you can adjust saturation on the CRT directly.

Are you suggesting unsaturated for CRT users, and saturated for those that use XRGB-mini?

[FBX]

I installed the Playchoice 10 -- FBX Unsat. -- FBX Sat. version, and the saturated palette running on the NESRGB looks simply identical to my side-by-side test with the unmodded deck on my LCD 55" display! I understand there will be differences on a CRT display (as A.P. mentioned), but the unsaturated palette is better for that setup because you can adjust saturation on the CRT directly.

Are you suggesting unsaturated for CRT users, and saturated for those that use XRGB-mini?

Yes. I figure since the mini only allows RGB level adjustment, the saturated palette is best for those that make use of that device.

Anyway, I noticed another color that the YUV palette gets totally wrong. At first I thought maybe I made a mistake in my own color conversion, but take a look at this comparison pic I made:



There's very little doubt the YUV palette is completely off on the dark green color. It isn't even close.

You sure it's 100%? It looks pretty good, but I noticed the terrain in Metroid looked a little too purple and red at the starting point. It looks bright blue on my CRT. I fired it up and looked at your version and what was getting output via composite to my BVM and it's definitely brighter and bluer through composite.

Could you upload pictures of the unsaturated version so I can compare it to composite better?

Will do tomorrow. Though you should remember, my palette is based on direct capture as opposed to the lack of accurate hue reproduction in a CRT. Granted, someone could claim "oh well see, you don't have an accurate palette to this person's CRT output", but my goal is to show that when you hook an umodded NES to CURRENT displays, this it what it should look like if the signal were pure RGB quality. you can see from my screen cap above that's exactly what it accomplishes.

[mvsfan]

I noticed Tim released 2 more palette files.

Are they still replacing Garish?

I still havent found my usb blaster yet.

[Josh128]

Is YUV supposed to be the same as "Improved"? Its definitely quite a bit off.

[FBX]

Is YUV supposed to be the same as "Improved"? Its definitely quite a bit off.

YUV is "natural" on the original NESRGB firmware selections.

I'm actually curious to find out what the truth is on those dark green/brown colors. Both the Framemeister and a direct video vectorscope test revealed the same green value used in my palette, while hooking the NES directly to my LCD display and having it translate the colors came out looking more brown like that YUV entry. That particular color swatch on the NES palette tree seems to be the hardest to get an accurate account of.

[Josh128]

Is YUV supposed to be the same as "Improved"? Its definitely quite a bit off.

YUV is "natural" on the original NESRGB firmware selections.

I'm actually curious to find out what the truth is on those dark green/brown colors. Both the Framemeister and a direct video vectorscope test revealed the same green value used in my palette, while hooking the NES directly to my LCD display and having it translate the colors came out looking more brown like that YUV entry. That particular color swatch on the NES palette tree seems to be the hardest to get an accurate account of.

I have a Sony Trinitron I can hook my NES to via composite to check. I left my palette switch with the ability to turn the NESRGB off and output the original composite of the NES out. I can check to see what it looks like if you are interested. Im curious now.

FWIW, I have "natural" and "improved" set on my NES, and the CV3 scene above also looks brown and not green when using "Improved". I guess improved may just be a more saturated version of natural?

[FBX]

Is YUV supposed to be the same as "Improved"? Its definitely quite a bit off.

YUV is "natural" on the original NESRGB firmware selections.

I'm actually curious to find out what the truth is on those dark green/brown colors. Both the Framemeister and a direct video vectorscope test revealed the same green value used in my palette, while hooking the NES directly to my LCD display and having it translate the colors came out looking more brown like that YUV entry. That particular color swatch on the NES palette tree seems to be the hardest to get an accurate account of.

I have a Sony Trinitron I can hook my NES to via composite to check. I left my palette switch with the ability to turn the NESRGB off and output the original composite of the NES out. I can check to see what it looks like if you are interested. Im curious now.

FWIW, I have "natural" and "improved" set on my NES, and the CV3 scene above also looks brown and not green when using "Improved". I guess improved may just be a more saturated version of natural?

I think "improved" might be the FCUX palette, but I'm not sure about that. Anyway, I got an idea to test the accuracy of the Framemeister's DVI mode color output, and this is to feed it a composite color source of the brownish-green value the YUV palette uses. If this converts to the more fuller green that I get from the console, then we know the Framemeister is at fault. The only problem is I need to figure out a way to get that brownish-green value onto a composite output signal so I can test it. Any suggestions on how to do this would be greatly appreciated.

[leonk]

I think "improved" might be the FCUX palette, but I'm not sure about that. Anyway, I got an idea to test the accuracy of the Framemeister's DVI mode color output, and this is to feed it a composite color source of the brownish-green value the YUV palette uses. If this converts to the more fuller green that I get from the console, then we know the Framemeister is at fault. The only problem is I need to figure out a way to get that brownish-green value onto a composite output signal so I can test it. Any suggestions on how to do this would be greatly appreciated.

I recall years ago desktop PC video cards used to have a composite video out port (for hooking up to TV's). Maybe you can do a full screen fill on a PC and use one of these video cards to output to a TV. You'll need to confirm on a CRT that it is the correct color, and then feed that into the XRGB-Mini.

The alternative is some sort of VGA/DVI to composite converter box. Should be a few bucks on eBay. You need to test with CRT first to make sure the converter itself isn't doing anything strange with the colors (A/B compare to NES connected to same video output device)

[yxkalle]

Remember that converting RGB to composite is a lossy process, including color hue. RGB->Composite->RGB will produce false colors.

[darcagn]

I'm actually curious to find out what the truth is on those dark green/brown colors. Both the Framemeister and a direct video vectorscope test revealed the same green value used in my palette, while hooking the NES directly to my LCD display and having it translate the colors came out looking more brown like that YUV entry. That particular color swatch on the NES palette tree seems to be the hardest to get an accurate account of.

Looks like FBX may be finally understanding what other people (RGB32E, bobrocks, etc.) were saying about NTSC since page 1!

[FBX]

I'm actually curious to find out what the truth is on those dark green/brown colors. Both the Framemeister and a direct video vectorscope test revealed the same green value used in my palette, while hooking the NES directly to my LCD display and having it translate the colors came out looking more brown like that YUV entry. That particular color swatch on the NES palette tree seems to be the hardest to get an accurate account of.

Looks like FBX may be finally understanding what other people (RGB32E, bobrocks, etc.) were saying about NTSC since page 1!

I just want to know which is device (TV or Framemeister) is misinterpreting that particular swatch of green/olive colors. All the other colors appear to be consistent on both devices.

[Josh128]

Looks like FBX may be finally understanding what other people (RGB32E, bobrocks, etc.) were saying about NTSC since page 1!

Nah... I think hes right. If you want the most authentic colors of a US NTSC NES, his palette is as true as it gets. His captures are damn close (perhaps perfect, Im just eyeballing) to what I got when I connected my NES via PPU composite to both my plasma and a 19" flat tube CRT. There was a slight variation between the two, with the plasmas green being a tad bit darker, but it was still a shade of green. I didnt have a capture card (actually I do but its not set up) so I used a Canon SL1 DSLR camera to snap some shots. I know its not the most accurate thing to do but I compared the shots to the actual set running in front of me and they were/are essentially identical.

It was hard to get a perfect exposure, I think they are slightly overexposed-- this explains the white/washed out look of the blocks compared to his composite capture. In any case, I believe this rudimentary comparison with a completely different machine running on different displays pretty much confirms his findings.

First shot is a reference using NESRGB "Improved" on my F4500.


Next, original NES composite on the 19" Symphonic SD Flat CRT


Same as above, but with flash.


Last, original NES composite on the F4500

[FBX]

In your first shot, it's better to used the "natural" palette since that's same one used in Nestopia's "YUV" palette.

In any case, the only real clue that the Framemeister is showing the correct green color was the video vectorscope measurement from the NES itself. For that particular color entry, here's what the measurements were:


C Phase (degrees): 177.7
U: -0.512365282
V: 0.020578715
R Analog: 0.077455209
G Analog: 0.244250286
B Analog: -0.987160138

This translates to RGB 8-bit as 20, 62, 0, which is very similar results to the 35, 70, 0 value I arrived at from the Framemeister capture. The YUV palette uses 51, 55, 0.

[darcagn]

Looks like FBX may be finally understanding what other people (RGB32E, bobrocks, etc.) were saying about NTSC since page 1!

If you want the most authentic colors of a US NTSC NES, his palette is as true as it gets.

I have no problem if FBX makes new palettes to try to get closer to the look of the output of the NTSC decoders he owns. Better for the community if people like you and others find that it's closer to your NTSC decoders as well.

My problem is just that he (along with others in this thread) is pretending like his palette is an "accurate" palette. There is no such thing due to the nature of NTSC decoding. He's finally running into a discrepancy in this specific color, let's see how long it takes for him to realize there is no one true color.

[FBX]

Looks like FBX may be finally understanding what other people (RGB32E, bobrocks, etc.) were saying about NTSC since page 1!

If you want the most authentic colors of a US NTSC NES, his palette is as true as it gets.

I have no problem if FBX makes new palettes to try to get closer to the look of the output of the NTSC decoders he owns. Better for the community if people like you and others find that it's closer to your NTSC decoders as well.

My problem is just that he (along with others in this thread) is pretending like his palette is an "accurate" palette. There is no such thing due to the nature of NTSC decoding. He's finally running into a discrepancy in this specific color, let's see how long it takes for him to realize there is no one true color.

If I honestly felt that way about my palette, I wouldn't be bothered by that particular color entry looking different on some displays. I'd just 'pretend' there's no issue at all. Since I don't want to assume anything, I look into these things to see if there's some way to figure out which color is the closer approximation. You say there's no true color, but by that logic one could put hot pink in that entry and EVERYONE would know that is way off. So very clearly there is some sort of threshold we can go by. My concept is to try to get as close to that threshold as possible, and I certainly don't want to assume the brown-versus-green variance is as close as it gets.

Edit: Someone requested a Metroid comparison pic earlier. Turns out in this case there's very little variation from the YUV palette to what I got on my end:



Framemeister grab on the left, saturated palette in the middle, YUV palette on the right.

[atheistgod1999]

FBX: Actually, the terrain in Metroid wasn't too purple, my monitor's color was warm due to having f.lux. Looking at it in normal color warmth (like i was supposed to), it looks exactly like composite on my CRT. Sorry.

[RGB32E]

First shot is a reference using NESRGB "Improved" on my F4500.

Hue set too purple.

Next, original NES composite on the 19" Symphonic SD Flat CRT

Hue set too green.

Last, original NES composite on the F4500

Ick! Overexposed and slightly too purple.

Your pictures suggest that different NTSC decoders do different things with different signals due to the variance in hues with the different combinations.

My problem is just that he (along with others in this thread) is pretending like his palette is an "accurate" palette. There is no such thing due to the nature of NTSC decoding. He's finally running into a discrepancy in this specific color, let's see how long it takes for him to realize there is no one true color.

Accept that he's a composite video "truther" of sorts.

[darcagn]

If I honestly felt that way about my palette, I wouldn't be bothered by that particular color entry looking different on some displays. I'd just 'pretend' there's no issue at all. Since I don't want to assume anything, I look into these things to see if there's some way to figure out which color is the closer approximation. You say there's no true color, but by that logic one could put hot pink in that entry and EVERYONE would know that is way off. So very clearly there is some sort of threshold we can go by. My concept is to try to get as close to that threshold as possible, and I certainly don't want to assume the brown-versus-green variance is as close as it gets.

Yes, if you put hot pink in the place of brown, it would obviously be incorrect, but if you got hot pink as a result from an actual NES, you would say that display is horribly miscalibrated. No properly calibrated display would show hot pink for that color.

However, different displays, all of which are properly calibrated, can still show variances in color (like the brown/green issue) because NTSC encoding and decoding are not accurate at color reproduction.

The logic behind your argument here is flawed. It's like asking "what is the result of the roll of a six-sided die?" Well, obviously it can be anywhere from 1-6. But no, you want to know the exact integer here, not a range, when there is none. When I tell you there is no true correct answer, you say "well, obviously if you said the answer was 13, that would be wrong, so there must be a correct answer." There isn't. Different rolls of a die will end in different results, just like different NTSC decoders will interpret variations in color.

[yxkalle]

The colors can even differ a bit from one NES/Famicom to another, due to transistor gain variations, inaccuracies in the chip die and the crystal oscillator to mention a few... There will never be a palette that looks like what everybody remember their particular combination of console and television looked like, but this is close enough for me.