NESRGB New Firmware Palette Updates:

The place for all discussion on gaming hardware
Sefirosu789
Posts: 83
Joined: Sun Jul 30, 2017 5:05 pm
Location: UK

Re: NESRGB New Firmware Palette Updates:

Post by Sefirosu789 »

nakedarthur wrote:Yep, it's the same
Thanks, can't wait to try them out. & Thanks to FBX for making these.
User avatar
Link83
Posts: 342
Joined: Tue May 21, 2013 2:39 am

Re: NESRGB New Firmware Palette Updates:

Post by Link83 »

FBX, is there any chance you could create a quick flowchart/diagram showing the naming progression for all your palettes? I haven't followed this thread properly for a while and things seem to have moved on. For example (And this is just a guess):-

Unsaturated >>> Nostalgia >>> Smooth

Digital-to-PVM >>> Original Hardware >>> NTSC Hardware

PVM-to-Digital >>>PVM Style >>> PVM Style D93

>>>>>>>>>>>>>>>Composite Direct

Does that seem right?

I'm also slightly confused why the latest NESRGB palettes include Unsaturated v6, when IIRC there was an Unsaturated v7 and Unsaturated Final?
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: NESRGB New Firmware Palette Updates:

Post by FBX »

Link83 wrote:
I'm also slightly confused why the latest NESRGB palettes include Unsaturated v6, when IIRC there was an Unsaturated v7 and Unsaturated Final?
There shouldn't be any Unsaturated palettes I support unless somebody else happened to like them. I do recall someone wanted me to add V6 in the current download list on my web site. What I support is on my web page:


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

Any other palette I have done I consider outdated work (even the V6, though some people like it).
User avatar
Guspaz
Posts: 3136
Joined: Tue Oct 06, 2015 7:37 pm
Location: Montréal, Canada

Re: NESRGB New Firmware Palette Updates:

Post by Guspaz »

FYI, nowhere on that page does it say what palettes are actually included in the NESRGB firmware. You have to download the ZIP file to figure out which is which. Not all the palettes in the ZIP file are described on the website either (only the readme file). Unsaturated v6 is included in that ZIP file so there's no indication that it is any more or less supported than any of the other palettes.
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: NESRGB New Firmware Palette Updates:

Post by FBX »

Guspaz wrote:FYI, nowhere on that page does it say what palettes are actually included in the NESRGB firmware. You have to download the ZIP file to figure out which is which. Not all the palettes in the ZIP file are described on the website either (only the readme file). Unsaturated v6 is included in that ZIP file so there's no indication that it is any more or less supported than any of the other palettes.
Would you rather I just take the site down? Geez, sorry man.
User avatar
Guspaz
Posts: 3136
Joined: Tue Oct 06, 2015 7:37 pm
Location: Montréal, Canada

Re: NESRGB New Firmware Palette Updates:

Post by Guspaz »

Uh, I'm just pointing out that there is some information missing on the site, and it's thus easy to see where his confusion came from, since I also got confused trying to look it up, not knocking the work or anything.
User avatar
Link83
Posts: 342
Joined: Tue May 21, 2013 2:39 am

Re: NESRGB New Firmware Palette Updates:

Post by Link83 »

FBX wrote:
Link83 wrote:
I'm also slightly confused why the latest NESRGB palettes include Unsaturated v6, when IIRC there was an Unsaturated v7 and Unsaturated Final?
There shouldn't be any Unsaturated palettes I support unless somebody else happened to like them. I do recall someone wanted me to add V6 in the current download list on my web site. What I support is on my web page:


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

Any other palette I have done I consider outdated work (even the V6, though some people like it).
Thanks FBX, I understand the old palettes are now considered outdated, but is there any chance you could still detail the naming progression? I'm also not clear if the NTSC Hardware palette still exists in some form?

It would also be great if your NES palettes page could have a small paragraph with more details about the origins of the Smooth palette, like you have for the PVM Style D93, Composite Direct and NES Classic palettes. Currently its just described as a "culmination of all those efforts" in the top section, but it doesn't really explain what the colors are based on, or what the aim/goal was of the Smooth palette (Also some comparison shots of each palette side by side would be much appreciated)

Lastly I would be interested in your thoughts as to the best combination of palettes for the NESRGB if i'm not interested in the PlayChoice-10 or NES Classic palettes?
Smashbro29
Posts: 532
Joined: Thu Dec 15, 2011 2:46 am

Re: NESRGB New Firmware Palette Updates:

Post by Smashbro29 »

It's been a while for me, wasn't the goal to find the most accurate colors possible?
User avatar
Guspaz
Posts: 3136
Joined: Tue Oct 06, 2015 7:37 pm
Location: Montréal, Canada

Re: NESRGB New Firmware Palette Updates:

Post by Guspaz »

Accurate to what? The Japanese video standard? The American video standard? The European video standard? On what display? From what hardware? Using which connection? There is no one answer. FBX has to balance all of the many possibilities to find what is most agreeable to the most people, and he's done that.

NTSC and NTSC-J don't even use the same colour temperature, with Japan using a bluer 9300K whitepoint versus North America using 6500K. So these games didn't even look the same on Japanese versus American televisions back in the day.
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: NESRGB New Firmware Palette Updates:

Post by FBX »

Smashbro29 wrote:It's been a while for me, wasn't the goal to find the most accurate colors possible?
After years of trying, I finally stopped being an idiot and realized it's impossible. Composite NTSC color output gets 'interpreted' in a myriad of ways depending on the display hardware, and some color output resulting on the CRT you can never duplicate in the RGB input of a PVM for example. The best you can do is make a few palettes that get pretty close to a few brands of CRTs, and somebody just picks one they like. For example, the Smooth palette is about as close as you can get to NTSC composite on a PVM. I'd say better than 95% accuracy in that scenario, but you can never get to 100 due to how some of the sky blues are handled between composite video input and RGB input on the same monitor.
User avatar
austin532
Posts: 842
Joined: Wed Mar 12, 2014 12:44 am
Location: Arizona, US

Re: NESRGB New Firmware Palette Updates:

Post by austin532 »

Pretty sure it was RGBSource who was the idiot. Claiming he had an accurate palette due to having better capturing hardware. It's unfortunate we can't get it to 99% but 95% is still pretty damn good. It was a fun journey.
Framemeister 240p scanline settings: http://shmups.system11.org/viewtopic.ph ... start=9600
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: NESRGB New Firmware Palette Updates:

Post by FBX »

austin532 wrote:Pretty sure it was RGBSource who was the idiot. Claiming he had an accurate palette due to having better capturing hardware.
I remember that. At the time, I was basing my palette off Framemeister output since that was the HOT piece of hardware everyone was hooking NESRGB mods into. Later on, I got other capture devices (including highly respected ones) and found the RGB conversion clipping was still an interpretive dance. Even Nintendo's own 'official' palette they used on the NES mini looked pretty awful, even when you excuse the white nose effect.
User avatar
Link83
Posts: 342
Joined: Tue May 21, 2013 2:39 am

Re: NESRGB New Firmware Palette Updates:

Post by Link83 »

FBX i'm not sure if you missed my above post or not, but if possible I would really appreciate your thoughts.
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: NESRGB New Firmware Palette Updates:

Post by FBX »

Link83 wrote:FBX i'm not sure if you missed my above post or not, but if possible I would really appreciate your thoughts.
Oh sorry about that.

Well first, I went through so many revisions of the older palettes that I can't really remember much about them. I didn't really keep a diary or anything like that. Once I felt my latest effort was better than the prior, I deleted the older work and didn't spend any more thought on them.

Next, the "Smooth" palette made use of the Analogue Nt mini's RGB output hooked into a PVM. On that same PVM, I had an unmodded NES hooked in via composite video. I then used a full-screen color palette selector ROM in both consoles, and dialed in the RGB values using the SD card on the Nt Mini to load in changes on-the-fly. Once I got it dialed to where pressing the button between the two inputs showed as close to identical as possible, I then started loading games and seeing how it looked on the PVM. I remember Legend of Zelda looked freaking identical between the two sources, but a couple of other games had a couple colors that didn't quite look perfect, so I manually adjusted I think 4 different entries until it was more visually appealing.

Interestingly enough, the final result was very similar looking to Wavebeam, only designed without oversaturation. Wavebeam is more saturated for an appealing look on a digital display, so I always recommend it paired with smooth on a 3-selector palette NESRGB board. The 3rd is up to you, though I'd probably go with composite direct since that's a simple capture card conversion.
User avatar
tjstogy
Posts: 341
Joined: Tue Sep 01, 2015 1:27 am
Location: New York

Re: NESRGB New Firmware Palette Updates:

Post by tjstogy »

FBX did a killer job on this.
Patiently waiting for your attempts at an Atari 2600 RGB palette, FBX... :idea:
User avatar
Link83
Posts: 342
Joined: Tue May 21, 2013 2:39 am

Re: NESRGB New Firmware Palette Updates:

Post by Link83 »

FBX wrote:
Link83 wrote:FBX i'm not sure if you missed my above post or not, but if possible I would really appreciate your thoughts.
Oh sorry about that.

Well first, I went through so many revisions of the older palettes that I can't really remember much about them. I didn't really keep a diary or anything like that. Once I felt my latest effort was better than the prior, I deleted the older work and didn't spend any more thought on them.

Next, the "Smooth" palette made use of the Analogue Nt mini's RGB output hooked into a PVM. On that same PVM, I had an unmodded NES hooked in via composite video. I then used a full-screen color palette selector ROM in both consoles, and dialed in the RGB values using the SD card on the Nt Mini to load in changes on-the-fly. Once I got it dialed to where pressing the button between the two inputs showed as close to identical as possible, I then started loading games and seeing how it looked on the PVM. I remember Legend of Zelda looked freaking identical between the two sources, but a couple of other games had a couple colors that didn't quite look perfect, so I manually adjusted I think 4 different entries until it was more visually appealing.

Interestingly enough, the final result was very similar looking to Wavebeam, only designed without oversaturation. Wavebeam is more saturated for an appealing look on a digital display, so I always recommend it paired with smooth on a 3-selector palette NESRGB board. The 3rd is up to you, though I'd probably go with composite direct since that's a simple capture card conversion.
Thanks FBX, that really helps :) So the Smooth palette sounds like its quite closely related to the older "NTSC Hardware" palette (Perhaps with some tweaks carried over from the Unsaturated/Nostalgia palettes) as both seem to be a good match for CRT users.

I just had one last question if you dont mind - did you submit the palettes Tim used to create the NESRGB 2.2 firmwares? If so might you know if the included Wavebeam palette is the March 2017 or October 2017 version? Apparently nakedarthur made some tweaks to the Wavebeam palette in October 2017, but I dont know what colors were actually changed :?
User avatar
nakedarthur
Posts: 218
Joined: Tue Jul 21, 2015 8:20 pm

Re: NESRGB New Firmware Palette Updates:

Post by nakedarthur »

Funny coincidence to see these palettes brought up again after a couple years now. I actually just uploaded an update to the .pal version of Wavebeam that fixes a couple problems. There are a bunch of black entries in the NES palette, and while most palettes 0 them all out as pure black, it turns out some games actually use them for detail as a near black color (Bee 52 is a good example). So anyway, I fixed that and made a couple other adjustments I've noticed over the past couple years as I've been playing. I will have to look into getting new NESRGB firmware made eventually.
Last edited by nakedarthur on Fri Aug 16, 2019 3:46 pm, edited 1 time in total.
Drunk_Caterpillar
Posts: 39
Joined: Fri Dec 23, 2016 12:07 am

Re: NESRGB New Firmware Palette Updates:

Post by Drunk_Caterpillar »

nakedarthur wrote:Funny coincidence to see these palettes brought up again after a couple years now. I actually just uploaded an update to the .pal version of Wavebeam that fixes a couple problems. There are a bunch of black entries in the NES palette, and while most palettes 0 them all out as pure black, it turns out some games actually use them for detail as a near black color (Bee 52 is a good example). So anyway, I fixed that and made a couple other adjustments I've noticed over the past couple years as I've been playing. I will have to look into getting new NESRGB firmware made eventually.
Do you have a link to that? Not 100% sure where to download the latest Wavebeam from.
User avatar
nakedarthur
Posts: 218
Joined: Tue Jul 21, 2015 8:20 pm

Re: NESRGB New Firmware Palette Updates:

Post by nakedarthur »

Drunk_Caterpillar wrote: Do you have a link to that? Not 100% sure where to download the latest Wavebeam from.
Sorry, added a link above
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: NESRGB New Firmware Palette Updates:

Post by FBX »

nakedarthur wrote:Funny coincidence to see these palettes brought up again after a couple years now. I actually just uploaded an update to the .pal version of Wavebeam that fixes a couple problems. There are a bunch of black entries in the NES palette, and while most palettes 0 them all out as pure black, it turns out some games actually use them for detail as a near black color (Bee 52 is a good example). So anyway, I fixed that and made a couple other adjustments I've noticed over the past couple years as I've been playing. I will have to look into getting new NESRGB firmware made eventually.
When I tested those entries on an unmodded NES, all of those blacks were the same output on the PVM. I'd like to see an example of which entry actually shows up as above flat black because I have never encountered this. Use a full-screen color test ROM please. Thanks!

Edit: Researching the subject on NES DEV, the ONLY way Bee 52 would show a non flat-black is if it used emphasis bit settings. According to NES DEV Wiki:
Note that $1D black is affected by color emphasis, but $0F black is not.
The problem with this is any other game that does not use emphasis will end up having your dark grey value instead of black. Also Bee 52 wasn't listed as a game that uses color emphasis, though it may be that you just discovered it does when nobody else noticed. Even if that is the case, the resulting output is not a balanced dark grey. Only one of the three color channels is emphasized while the other two are slightly dimmed. So instead, you would have 0, 0, 8 or 0, 8, 0 for examples. However, color emphasis I believe is already accounted for in color emulation, so that's why the .pal format never includes entries for them.

Edit 2: Did some googling about Bee 52 and indeed it uses entry $1D for the off-black emphasis. A separate palette has to be made for it from what I read.

Edit: 3: More from NES DEV Wiki:
$D is dark gray, and $E-$F are mirrors of $1D (black). The canonical code for "black" is $0F or $1D.
Additionally, Bee 52 is one of a few games to also use $0D, which apparently is a no-no because it's a loss of voltage, causing different behaviors on different displays (and also depending on how much of the screen is assigned that color entry). I know on my PVM, there's no reaction to it.
User avatar
nakedarthur
Posts: 218
Joined: Tue Jul 21, 2015 8:20 pm

Re: NESRGB New Firmware Palette Updates:

Post by nakedarthur »

I don't think it has anything to do with the emphasis. They are using a different color for off-black spots in the map. Here is the scene using Index palette (look in grass and below the triangle thing):

Image

I also looked at the data Artemio captured directly from scope and it matches what I'm seeing. While we haven't gotten IQ decoding, brightness was figured out.

0D: 0,0,0 (true black)
0E-0F: 3,3,3
1D: 9,9,9
1E-1F: 10,10,10
2E, 2F, 3E, 3F: 17, 17, 17
Last edited by nakedarthur on Sat Aug 17, 2019 1:21 am, edited 1 time in total.
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: NESRGB New Firmware Palette Updates:

Post by FBX »

nakedarthur wrote:I don't think it has anything to do with the emphasis. They are using a different palette index for off-black spots in the map. Here is the scene using Index palette (look in grass and below the triangle thing):
Can you elaborate on "palette index"? I can't find anything on NES Dev referring to indexes. Also another OCD palette maker specifically stated he had to make a separate palette for Bee 52, because if you apply that off-black to some other games, it comes out looking wrong. You'll end up seeing blocky patches of very dark grey and patches of black on the screen. This was a great example someone posted on those forums, where using off-blacks caused the intended look to get WAY off base:

https://forums.nesdev.com/download/file ... &mode=view

Also please note that while scopes can measure voltage, if the CRT does not react to it (because it's not strong enough) then it's going to look just as black as the all the other black entries. This was in fact the behavior of my PVM.
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: NESRGB New Firmware Palette Updates:

Post by FBX »

I think what I'm going to have to do is load up Bee 52 and Maniac Mansion on my PVM and see if there's ANY visible differences in the blacks on my unmodded NES. If a PVM cannot pick up the difference, it's highly unlikely a consumer grade TV would.
User avatar
nakedarthur
Posts: 218
Joined: Tue Jul 21, 2015 8:20 pm

Re: NESRGB New Firmware Palette Updates:

Post by nakedarthur »

Sorry I updated the post, I was just referring to memory address in decimal. I will check out the nesdev post, sounds interesting.
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: NESRGB New Firmware Palette Updates:

Post by FBX »

nakedarthur wrote:
0D: 0,0,0 (true black)
0E-0F: 3,3,3
1D: 9,9,9
1E-1F: 10,10,10
2E, 2F, 3E, 3F: 17, 17, 17
According to NES DEV threads I've been reading, the 0D entry is not "true black" but rather a mistake in the hardware, where not enough voltage is provided. Very few games used that entry as a result, with most of them being Camerica games because they had to back-engineer the NES in order to code games for it. In other words, they simply had no access to technical notes warning not to use 0D. There are also several threads about these games causing display glitches on some CRTs as a direct result of using 0D for black. The entry has been jokingly referred to as "Blacker than Black" as a reminder to stay away from it.

Getting back to the other entries, as I mentioned, NES DEV Wiki claims that all the E and F entries are "mirrors" of 1D. If you don't want your black affected by emphasis, then you're supposed to use 0F instead of 1D. With all this said, it could be very possible the NES DEV Wiki has incorrect or outdated information, and that perhaps those values may actually be more accurate. I still need to check everything out on my PVM though. I will work on it overnight tonight.
User avatar
nakedarthur
Posts: 218
Joined: Tue Jul 21, 2015 8:20 pm

Re: NESRGB New Firmware Palette Updates:

Post by nakedarthur »

Cool makes sense, I see the Castlevania title screen uses 0F. Did all those show up as 0s when doing your Composite Direct palette? The 3,3,3 from Artemio's capture would definitely seem like true black on about any display. 1D-1F is where it gets interesting I guess. I would think it would be hard to eyeball even on PVM without a side by side comparison with 0F?
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: NESRGB New Firmware Palette Updates:

Post by FBX »

Just checked on my 20M2U. There is ZERO difference in any of the black entries on that monitor. Not even 0D causes issues on it (where it would with other CRTs). I even cranked the brightness all the way up to check it as a grey image for even the slightest difference. There was not even the slightest hint of change.

This was with the fullscreen color ROM test. Next, I'm going to run Bee 52 and Maniac Mansion. I'll report back in a few mins.
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: NESRGB New Firmware Palette Updates:

Post by FBX »

Update:

Okay so this is really weird. Maniac Mansion shows no change in any of the blacks on that intro screen with the meteor. Bee 52 does in fact show a brightness change on the tree trunks and grass, but I have to crank the brightness on the PVM way up to see it, otherwise it all looks black. Gonna have to play around with this on Nestopia now to experiment with those entries.
User avatar
FBX
Posts: 2347
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: NESRGB New Firmware Palette Updates:

Post by FBX »

Wow! Discovered something not covered on NES DEV Wiki by using a different color test ROM that lets you compare all the blacks:

1D AND 0D are BOTH affected by emphasis bits! I had to crank the monitor brightness way up to see it, but it's blatantly obvious once you do. Setting all the emphasis bits on for 0D makes it EVEN darker! Likewise, setting all the emphasis bits on for 1D turns it into the same shade as default 0D. Would love to know if Bee 52 is doing this??


Okay so here are the final conclusions as far as the PVM testing goes:

0E, 1E, 2E, 3E, 0F, 1F, 2F, 3F are all the same exact shade on the PVM even when brightness cranked. There's no difference. They also do not respond to emphasis bit settings.

1D with emphasis bits turned off is identical in shade to the E and F entries. Turning on all emphasis bits turns 1D into the same shade as 0D. Turning different emphasis bits on and off will change the hue of 1D.

0D by default is darker then 1D by default. 0D also responds to emphasis bits, where turning them all on makes it even darker. Turning different emphasis bits on and off will in also change the hue of 0D.

Edit: I finally found this same information I concluded on NES DEV Wiki, only instead of being in the palette section, it was in the NTSC video section:

https://wiki.nesdev.com/w/index.php/NTSC_video
User avatar
Link83
Posts: 342
Joined: Tue May 21, 2013 2:39 am

Re: NESRGB New Firmware Palette Updates:

Post by Link83 »

Sorry to bump this thread, but I was just curious if any of these recent discoveries about the emphasis bits have been incorporated into any NESRGB palettes? and has the NES Dev Wiki palette section been updated?
Post Reply