What made the NES so interesting?
nicole.express
nicole.express
This is something that a lot of even very plugged into retrogaming people don't really get about the NES and RGB mods. I think it's very counterintuitive for people now that there is no part of the system that operates in actual 'color' information. The NES' video output is composite from start to finish[1], with palette entries corresponding to approximate output frequencies for the color subcarrier.
So there is no canonical palette for the NES - the colors produced by games varied by region, a little bit by unit, and especially because NTSC had trouble with stable color even by TV.
So people imagine that RGB or HDMI mods of the NES are somehow tapping into a missing layer of raw color information, but what it's actually doing is intercepting communications with the PPU and modelling (on an fpga) an extra PPU that does operate in RGB, and then constructing a new video signal from that.
And then you have to pick a palette, because there isn't just one.
I have an RGB modded NES (a US toploader) but I doubt I will ever buy or make another one. At this point I think it's better to just use an upscaler, which will produce no less authentic a picture. The main benefit imo of the rgb mod is that there's less interference so the picture is clearer, even if you just output reconstructed composite from it.
[1] There is actually a model of NES PPU that outputs RGB, it was used in the PlayChoice arcade machines (which generally used RGB monitors, not composite) and it has a palette that most people agree is ugly as sin.
My thinking is that there was certainly an RGB palette that the designers — probably working on something like an Amiga — chose when creating the pixel art for the games; and that Nintendo likely has institutional memory, if not hard records, of what that design-time palette was; and so later palettes in RGB-based emulators were likely chosen to closely resemble those design-time palettes.
* As evidenced by the fact that the (RGB-based) Nintendo VS. System has a different palette than later emulation releases, even at the time Nintendo wasn't sure of the "correct" RGB palette (or has since come around and changed their mind about what that should be).
* There wasn't really any sort of standard devkit. Different developers, different games, different artists used different hardware and software tools to draw their graphics. Nintendo was notoriously crappy to third-party developers during the Famicom years, and Western developers in particular had little institutional support from them. Even if you had the correct palette that Nintendo was using between 1988 and 1991, that's not necessarily what Rare was using in 1990, or what Square was using in 1987, or...
* The Famicom was released in 1983, and development on it began in 1982. There was no Amiga, no Atari ST, no X68000, no VGA, no mainstream computers really capable of general purpose graphics editing. Developers often had to come up with bespoke systems for editing graphics at the time, including bespoke hardware - and it's entirely possible these bespoke systems used a similar non-RGB design as the Famicom, or even the same PPU, or were even just editing programs on a Famicom with some kind of serial-out hacked in.
* Philosophically speaking, the artists were aware that they were producing graphics for the Famicom/NES. That is, they weren't looking at whatever RGB output they had as the definitive version, but just as a rough guideline for how the image would appear on TVs. Given that, whatever RGB the devkit might've been spitting out isn't "definitive" in the sense of how the artist intended anyone to see it.
* I don't really trust Nintendo to make any extra effort to provide an "authentic" experience in their re-releases: witness, for instance, the low quality of the N64 emulation in the Switch Online Expansion Pack they just put out. The actual people who did the art may be gone or uninvolved in the re-releases; just because they own the copyright I don't think whatever Nintendo slaps together really represents the intent of the artists themselves, in the same way that I don't think, say, the release of Go Set a Watchman says anything about the intent of Harper Lee.
We have seen photos featuring people using nice, pro or broadcast grade CRT displays. Given the lack of an RGB authority, and the limited number of colors, it is likely each group just used whatever references that made sense to them.
And regions. PAL, SECAM, other variants all added noise.
I'll bet a standard consumer grade TV with RF was a small part of many dev processes too. Just a check on effects and or particularly troublesome color combinations. The pro or broadcast grade displays would often render those better than one might expect, but your average TV would likely display a hopefully interesting blur...
RGB images suggest this check on a basic RF TV because some of the pixel art choices may have been different, given better display signals were an option.
For me personally, viewing an NES on a better quality composite display works best for me. Personal preference, and that is due to having composite capable displays early on. I used composite more than anything else. But all that said, the NES RF modulator was respectable. Worked better than many did.
Often, the design choice was a long cable to the TV.
I had a few systems that parked the modulator right next to the signal input on the TV. TI Home computer had a huge one, but it really was great. I used it for many different systems. There was a modulator for PS1 that you could use with a very short connection. And the NES.
RF done that way worked well enough.
While I still feel the original quality was shameful, I recently upgraded my subscription and I have been enjoying it with no issues.
To add to that, the emulation of old games is handled by the NERD group at Nintendo, based in France (https://en.wikipedia.org/wiki/Nintendo_European_Research_%26...). Nintendo is not a monolithic block, and internal groups will have vastly different opinions on how things should look.
I don't think it's necessarily even very likely that they were looking at rgb from a devkit. These days people associate PVMs and other pro monitors with RGB but they aren't synonymous. Take a look through ebay listing for PVMs and you'll find plenty of composite or s-video-ish pro monitors from the 80s kicking around.
And even some microcomputer monitors were composite. If the dev tooling they used was based on a C64 or Apple IIe (which also used 6502 cpus), that'd be composite out for eg., even with a designed-for-c64/appleIIe monitor. And I'm pretty sure I've seen photos of people using Apple IIe or similar for NES dev.
Sure, but that's not what I meant.
Back when images were being bit-banged into computers, you didn't really care what the results looked like on the screen of your workstation, if your workstation wasn't the target system. You designed on paper — usually grid paper — and then iterated from there by flashing an EEPROM and checking the results on the machine.
When I say "the design colors", I mean the marker-pen colors used to color the grid paper. You might also call these the "pre-color-grading" palette.
Certainly, the artist would get a different idea about what they wanted a thing to look like, once the image had been passed through the "color grading" of the NES palette; and so would then iterate toward a design that most-closely lined up the vision in their head with what the NES could actually produce.
But the artist's original intent wasn't to achieve that iteratatively-color-graded final result image; it was to achieve the original colors they had put down on the grid paper. That original intent was just stymied by the system; and so they had to settle for the colors the system could do.
(I realize that modern designers working with retro art styles often design from scratch under palette constraints; but video games were much newer then, and so most of the professional art-and-design people working in the industry had never done digital art before, and had been trained / previously worked only in traditional art. The closest their designers would have got to the kind of constraints imposed by the NES would be in designing neon signs.)
* The power of that sticker was amazing! ;P
Before colour TVs, Brazil used the American 60Hz System M (B&W predecessor of NTSC), so when it came time to adopt colour they couldn't choose the European PAL because it would not be compatible with current B&W TV sets, and they didn't want to adopt NTSC* because by then it was clear it was an inferior colour encoding, so they slapped a PAL colour encoding on top of a SystemM B&W signal, resulting in a 525 lines 60Hz without NTSC colour distortions (kinda best of both worlds).
* Also they didn't want to adopt NTSC because they wanted to prop the national electronics industry, so a home made standard meant local TV makers didn't have to compete with imported TV sets.
In fact, if you had a composite device (with the yellow, red and black plugs), you could just get one of those cheap passive adapters to hook it up to SCART: https://media.s-bol.com/mwVL1yR45z6E/1144x1200.jpg
IIRC our childhood (Philips) TV had two SCART inputs: one supported both RGB and CVBS, the other only CVBS.
This is why we joked NTSC meant Never The Same Color.
The last thing you wanted to see in NTSC was bright red because you'd continue to see it smear for at least 1/4 of the screen to the right. NTSC "Safe" red wasn't really red. Run a broadcast safe filter on 255,0,0 and see how far off it goes. Should be somewhere around 180,0,0 IIRC. The days of looking at vectorscope to see the red point shoot off while everything else was much closer. When it became easier to get digital images to tape, we'd have to adjust the chroma down on the TBC to bring it back into legal limits which would mean the rest of the other colors would get squashed when the client didn't work in NTSC safe color palletes.
Weren't NTSC somewhat also grainier with more pronounced horizontal gaps between each scanline? Certainly seemed so.
Refresh rate is linked to mains frequency.
It doesn't have to be. Japan has a split power grid with 50Hz in the east and 60Hz in the west, but television frequency is 60Hz on both sides.
https://en.wikipedia.org/wiki/Electricity_sector_in_Japan#/m...
I've personally never seen the ability to switch from mains frequency. That's a switch on the back of the PSU that typical goes from 240v or 120v.
The turbo button and mains frequency are 2 totally different things.
The AC power line frequency was used for the vertical refresh rate for two reasons. The first reason was that the television's vacuum tube was susceptible to interference from the unit's power supply, including residual ripple. This could cause drifting horizontal bars (hum bars). Using the same frequency reduced this, and made interference static on the screen and therefore less obtrusive. The second reason was that television studios would use AC lamps, filming at a different frequency would cause strobing.
It is similar to the effect that could be seen when using native lighting in a 50Hz mains country while recording video in NTSC. Not everyone would notice, but when you sit in an edit bay working with that footage all day, you notice everything.
Then again, some people don't notice the stuttering that is the result from 24->29.97 either. The modern 4th frame repeat drives me up the wall. </rant>
Though UK developers new what they were doing an optimised properly for pal. Goldeneye looks better on pal for example.
Australia is PAL too. I remembered waiting for this game and being so disappointed. It was $135 also which was a lot. Modern games are like $70.
SECAM = System Essentially Contrary to the American Method
PAL = Perfection At Last
I want a time machine to go back and prevent the US decision makers to not fret about keeping a color system backwards compatible with B&W. I don't care to use the time machine to prevent war, making stock picks or sports bets. I just want to make my future career a little less insane by stopping this idiotic (in hindsight) decision. If there's enough time, I'd try to convince them to keep working on progressive scan and avoid more complications to my future career by not needing to deal with interlacing issues.
And yeah, I also got a modded nes to use with an ossc but now there are good retro-console upscalers that'll take composite (notably the retrotink5x, which incidentally does have a deblur filter for the n64 iirc) and it just doesn't seem worth it anymore. I'll still mod my consoles (or get them modded by someone better at it than me) if they actually have a better video signal to steal, but I just don't see the point in faking one in the console instead of outside it.
I've been poking at trying to mod famicoms/NESs just for the purposes of getting a cleaner composite signal out of them, but it's tricky and I'm not that good at this stuff.
https://community.amd.com/t5/gaming/amd-fidelityfx-super-res...
What made the NES so interesting? It was the games, pure and simple! I've got probably 20 versions of classic arcade Q*Bert. But the iteration with the "purest" design? My goto to actually play when I want to go for a personal high score? Hands down, it's the NES.
Europe had RGB so I wonder if that was part of the reason the Mega Drive (Genesis) was so popular. The SNES doesn't look (much?) better than the Genesis when both in RGB, but SNES composite is much much better than Genesis.
I think they pretty often need to be recapped in modern times. More so than contemporary nintendo systems tend to need to, I think.
In the end most of this wasn't really super obvious on an 80s/early 90s consumer CRT though, and I don't think people were really comparison shopping on it.
Take the analog path from NES to TV in that common situation, then the second analog path inside the TV generating the color, then NTSC color/lum separation limitations, and "crisp" or "accurate" is something that couldn't describe the results.
On nesdev.com there's at least one thread where quite a few people are discussing about the best way to capture the color palette from video sources, and argument that there really isn't a "golden palette" because each TV was different and NTSC quirks: a pixel sent to a CRT TV could have different colors depending on its position and the colors/luminance before and after.
So CRT/NTSC fuzziness is part of the memory and nostalgia. I like the NTSC emulation options in FCEUX which do a really good job of the color fringing and smearing that was on CRT TVs and really make it "look" like how it was when I was young.
For instance the color palette was designed with aesthetics in mind: it had 56 colors that you'd want to use, compared to the 216 colors of the web safe palette that must have been chosen by a physicist because 60% of them make me want to puke.
The Atari 400 and 800 had an elaborate system for game-friendly video that couldn't touch what the NES did with tiles. They did amazing things with the TI-99/4A (like simulate framebuffer graphics with tiles so long as you didn't need to cover the whole screen... a trick you couldn't do with the character ROM talked about in this article) but it didn't have the brilliant end-to-end engineering of the NES...
... And of course the ability to extend the hardware with the cartridge took advantage of the falling cost of electronics AND fit with the business model that the cheap console is subsidized by expensive games.
Something like the C64 which used a "resistor bank" to pick the colors actually had thought into the individual choices of each color.
What NES got very right IMHO was:
- More than 16 colors
- Adequate horizontal resolution but not caring about the maximum since the point wasn't to display text (256 pixels, not 160 or 320)
- 2 bits per pixel with just enough palettes to be useful - no 1bpp graphics.
- 64 sprites - so many sprites even if they were 8x8
- Hardware scrolling and multiple nametable support
A lot of the magic in the NES was all the sprites - 64 is a lot for the hardware at the time, the fact they had the same resolution as the nametable, and easy support for scrolling. Also the CPU was kinda fast - 1.79MHz is speedy for a 6502 CPU.
I think it's got something to do with the eye being much more sensitive to green but RGB cubes wind up terribly unbalanced and "meaningless" in my way of thinking whereas the Nintendo colors are "meaningful" in that they are aligned with human perception. (Never seen anybody else do it but I run my Hue, Sengled and LIFX lights green on hot summer days to add less heat to the house... I figure the strange color rendition is not so bad because I'm either supplementing outside light from the windows or if is really hot the windows are covered with space blankets and I don't care how strange of a scene it is.)
According to Bob Yannes, there wasn't much thought that went into the C64 color selections. From https://www.pepto.de/projects/colorvic/2001/ :
Subject: "Re: VIC-II colors"
From: Robert 'Bob' Yannes
To: Philip 'Pepto' Timmermann
Date: 27.09.1999
I was involved with the development of the VIC-II, however the actual implementation of the design, including the Color Palette, was done by someone else. I have forwarded your message to him, but it is up to him if he wants to respond.
I can tell you that the design was based on the principle that adding a sine wave of a particular frequency and amplitude to an inverted version of the same sine wave at a different amplitude produces a phase-shifted sine wave of the same frequency. The amount of phase shift is directly proportional to the amplitudes of the two sine waves.
The VIC-II used the 14.31818 MHz master clock input (4 times the NTSC color burst frequency of 3.579545 MHz) to produce quadrature square-wave clocks. These clock signals were then integrated into triangle waves using analog integrators. The triangle waves were then integrated again into sine waves (actually rounded triangle waves, but good enough for this application). This produced a 3.579545 MHz sine wave, inverse sine wave, cosine wave and inverse cosine wave.
An analog summer was used to create the phase-shifts in the Chroma signal by adding together the appropiate two waveforms at the appropiate amplitudes. The Color Palette data went to a look-up table that specified the amplitude of the waves by selecting different resistors in the gain path of the summer. The end result was that we could create any hue we wanted by looking at the NTSC color wheel to determine the phase-shift and then picking the appropiate resistor values to produce that phase-shift.
Color Saturation was controlled by scaling the gain of the summer. When we picked the resistor values to determine the output phase-shift, we also scaled them to produce the desired output amplitude. Luminance was controlled using a simple voltage divider which switched different pull-down resistors into the open-drain output. We could create any Luminance we wanted by choosing the desired resistor value.
I'm afraid that not nearly as much effort went into the color selection as you think. Since we had total control over hue, saturation and luminance, we picked colors that we liked. In order to save space on the chip, though, many of the colors were simply the opposite side of the color wheel from ones that we picked. This allowed us to reuse the existing resistor values, rather than having a completely unique set for each color.
I believe that Commodore actually got a patent on this technique. It was certainly superior to the Apple or Atari approach at the time, as they ended up with whatever colors that came out — ours allowed the designer to freely select Hue, Saturation and Luminance.
Since all of this was based on selecting different resistor values and resistance varied from chip lot to chip lot, there was variation from one Commodore 64 to another. It wasn't as bad as it could have been though, since all of the Chrominance selection was based on resistor ratios, which could be kept constant even if the actual resistor values varied. Luminance was more of a problem. A trimmer resistor should really have been used to pull up the output. This would have allowed the Luminance to be adjusted for consistency from unit to unit, however Commodore didn't care enough about consistency to bother with adjusting each unit.
—
Robert 'Bob' YannesThis chip
https://en.wikipedia.org/wiki/ANTIC
is controlled by a "display list" that sets the strategy used to draw lines on a line-by-line basis. If you have 20 blank lines you can just skip past them and not waste any RAM representing them. If you want to put a line of text at the bottom of a video game it is completely easy.
This chip
https://en.wikipedia.org/wiki/CTIA_and_GTIA
is controlled by the ANTIC and composites ANTIC's data stream it with sprites, tiles and characters.
It's a sweet system and crazy flexible, made it easy to make games like
https://www.youtube.com/watch?v=3_VDM8nC9sM https://www.youtube.com/watch?v=MOV5C_wvP4o
but it just didn't have it together like the NES system, or the even better tiles & sprites system in the original Game Boy.
Most likely a computer programmer. It's just all the combinations of six shades of Red, Green and Blue that are equally spaced out. 00, 33, 66, 99, cc and ff.
The NES colors are actually equally simplistic internally:
12 equally spaced color hues with 4 levels of luma (plus 4 grays, 2 whites and 10 blacks wasting space in the pallet). But it works so much better because it's done in the YUV color space, rather than the RGB color space. Then the analog hardware doesn't manage to replicate it's target YUV colors accurately (or consistently across models, regions and even individual machines/tvs), which adds character.
Once I tried to come up with a good set of 256 colours for a fantasy console and found almost all of the existing options sucked.
I Documented an analysis and what I came up with here http://blag.fingswotidun.com/2018/04/giving-kwak8-256-colour... ( You can mouse-over the spinning colour cubes to view them in Lab space)
I think I could come up with a better set if I were to do it again today, but I was pretty pleased with this attempt.
At the time I had a C64 and an Atari 2600, but would spend every quarter I could dig up at the arcade at the bowling alley next door playing pretty much anything, but especially Super Mario Bros.
The Nintendo was the first console to bring the arcade home and it was absolutely revolutionary.
I played it on the NES first (and it was released there before arcades, though I don't think I played until 1989 personally). But I didn't own an NES at the time, so it wasn't that I could play it for free at home.
SMB is a much better game than Rampage. I would also play Excitebike at the arcade, also not nearly as good as SMB.
I guess in an arcade I was looking for simple high score games, I wanted to dream of putting my initials on, but I also wanted to be able to walk away when I died (where in SMB if I got to 8-2 then died I'd wanna go again/continue).
I could never get into Excitebike. Controls always felt so awkward.
What it was, was the most successful console in America in that era but that was largely due to the crash in the console market over there (something that didn’t affect the rest of the world).
Also the ability to do simple stored audio wave patterns and not just sine waves and white noise probably mattered a lot.
Another little detail is the controller layout. Nintendo was the first to figure out that "shitty rubber joystick" wasn't a selling point, and a simple d-pad was infinitely preferable for real games.
Platformers on a single or two button joystick, for example, just do not work as well.
If you liked this, you'll likely also enjoy https://youtu.be/ar9WRwCiSr0
I'm not giving a summary since that video is most fun to watch if you don't know the punchline in advance.
(No relationship to the video other than that I liked it.)
Brilliant move making the pins so powerful.
You'd be amazed at how many times Nintendo almost died in the cradle!
It’s also interesting seeing what influence the American offices of Sega and Nintendo had in the 90s.
C64 and Amiga are the precursors of the Raspberry, that while not 100% good; they atleast leave you in charge of your machine to a larger extent:
http://move.rupy.se/file/commodore.png
Another interesting tidbit is that Nintendo wanted Commodore to design the Famicom but Tramiel bailed in the last second when the Commodore guy was allready in Japan on his way to that first meeting.
Last but not least, the C64 had S-video before it even had that name, separating the chroma and luma. Not even the Amigas had that 10 years later.
Besides, Nintendo did have the Famicom Disk System in Japan, which used rewritable floppies.
And not even 0.0000001% of NES/Famicom owners had a keyboard and a disk drive and out of those probably 0% did any coding since Nintendo to this day does not allow you to run what you want on your hardware unless you hack it.
That was in August 1985, about three years after the C-64. The Atari ST came out a few months earlier, and had excellent monitors, the problem was you really wanted two of them, one for high resolution (640x400), non-interlaced black and white, and the other for RGB low resolution (320x200) color. They were not compatible with each other.
In the PC world you started to see digital RGB EGA with 64 colors about this time, but until VGA monitors came out there wasn't really anything to compare to the Amiga / Atari ST in the color department. VGA was released in 1987, and of course took a few years to become common. An amazing number of PC video games were still CGA (four color low res digital RGB) at the time, but that of course all changed a few years later when VGA became standard.
Again everything can do anything.
Of course these days HDMI conversion is better, but that is also the case for nearly any video signal you want to display from more than a couple of decades ago.
Has anyone attempted to calculate utility per byte? Certainly, the relative evolution of society would play a role (a cave-dweller might have found a few hieroglyphs entirely fascinating), but even with a certain adjustment factor in terms of the general complexity of dominant sources of information, Nintendo games would assuredly excel.
As evidence, folks are still speed-running the original Super Mario Bros from 1985. Lindy effect or efficiency of "utility" per byte?
I'm an expert on the Atari 2600 (wrote a homebrew game.) This channel's video on it isn't quite accurate. He overstates the case of how primitive the machine is, claiming that it takes only 39 bits to display the 2600's screen. That's not really true. The playfield (20), sprite (8x2), missiles (2x1), and ball (1) do total to 39 bits of pixel data... but a lot more than that is involved, many registers modify those pixel bits. There are two X-position latches for all five objects, four color registers, two sprite width/repeat registers, two sprite reflection registers, one register to repeat/mirror/priority the playfield, plus a few more bits in the "vertical-delay" (really each another latch) and timing (horizontal and vertical blank and sync) registers.
Here's the list, with a total of 45 byte-sized registers, although not all bits are used and some are for nongraphical purposes (audio, collision detection). https://github.com/munsie/dasm/blob/master/machines/atari260... (starts at line 88)
If he didn't like the game, the game would not get published no matter what.
This process ensured that the NES had only quality games.
Yamauchi was 7th dan in Go, and at one point was the richest man in Japan.
Probably some rubberstamp thing then. NES had _plenty_ of bad games though, actually the majority of games published were bad.
See LJN.
The required you to put a licensed catridge on top of the game catridge, and the unlicensed catridge would piggyback on the licensed catridge for loading.
The SNES was the most "special" of the Nintendo machines in that it hit the sweet spot where the games were sophisticated enough to be really interesting yet still simple enough to be made by small teams and thus not end up as run by committee corporate turds.
I cannot complain enough, this is a simple web page. there is absolutely no reason i cannot capture a static image of it.
beyond that, i sort of enjoyed a deeper dive, but the subject matter was relatively uninteresting to me. Technology used to have to be bent to the will of the group programming for it, and i'm sure with enough time and the lack of a release schedule from the three big CPU manufacturers, we'd have amazing results with x86 and Arm now, too. But it was a different time, and i do appreciate history.
Sure do wish i could save the site as a .png, though.
With the European GDPR(?) stuff and general cookie alerts and floating ad boxes, i find that a good 20-30% of all web pages cannot be correctly "captured" as an image. Realistically, since i can literally scroll the actual webpage and see everything, there is no reason that a screenshot wouldn't capture the exact same thing, but it's 2022 and here we are.
Sorry, my pastebin (projectftm.com) is one of the few that allows both arbitrary resolution screenshots as well as pinch/ctrl-scrollwheel zoom to arbitrary sizes. My screenshot bot uploads to a regular nginx webserver, however the screenshot id it generates is a 404. I haven't figured out how to adjust the timeout interval, as i think that's on the bot's home host side, and not the firefox side. the API has a timeout for receiving data from a remote host and if firefox takes longer than that to produce a screenshot it just gives up, even though the script has created a link.
I'm nearly positive it doesn't have anything to do with your site, even though the page weighs in at at least 19000KB as a .png. I'm weird, i think .png (or even .gif/.jpg) full page renders are functionally equivalent to a "print to pdf" except more useful.
Sorry i sort of derided your efforts. My parents bought me a Sega master system rather than a nintendo, and the past year has me worn out on the "6502-era" systems.
[0]http://projectftm.com/#-b3CluYULJcy0JiPUjjUCg this is my personal pastebin, the screenshot linked here is 10% of my total storage on the pastebin, so this may not be available by the time you see this reply. you can @genewitch me on twitter or gmail if you want to see what i mean, and this link is "dead". Many spring tidings!