Could 'Maria' have been an expansion?

[+darryl1970]

My understanding is that the 5200 started to take over the market right before it was killed. This could indicate that 2600 compatibility was no longer much of an issue, and people were starting to look for more advanced games. I believe that 2600 compatibility weakened the 7800's overall abilities (hence the choppy game play on many 7800 games). I don't want this to be a bashing session of the 5200 or 7800 (including the 5200 joysticks), but a marketing idea and THEN whether it would have been possible.

 

Let's say Atari completed the 5200 self-centering prototypes (similar to PS2 type mechanisms??), I think the joysticks would have been less to complain about as long as they didn't break. They already had the 2600 add-on. If Atari had used Maria as a "Super Game Module" of sorts, would that have saved money and created a better-performing machine. The 7800 could have been an add-on module that used the 5200's Pokey chip (no need for putting it on the cart), and more video memory could have been added instead of requiring all of the parts for a stand-alone machine (and 2600 compatibility). This might allow the higher resolution video modes to be used more often!

 

Is there away this could have been accomplished? I am not familiar with the hardware enough, but I understand the 2600 module is a 2600 that bypasses the 5200 hardware. Would the 7800 module have been able to only bypass the video chip from the cartridge port? Could they have used the bus on the back of the 5200 machines? I just think this would have saved Atari some. People who adopted the 5200 wouldn't have felt abandoned, and the 5200 would still play 2600 with a module. It would make the 7800 almost like an XM expansion, minus the Yamaha sound chip. Improved analog joysticks, along with the keyboard, would have allowed for more controller options. I just wonder why this was never thought of as an option. Additionally, it would have given Atari the image of always being expandable (one of Coleco's initial selling points!)

[ATARI7800fan]

I think it could of been possible, but we will never know for sure.

[GroovyBee]

Is there away this could have been accomplished?

 

No.

[+darryl1970]

Is there away this could have been accomplished?

 

No.

No way possible, or no because it wouldn't have been the same 7800 (and different to develop for?).

 

I realize you have done some pretty impressive things with the 7800, which indicate that the weak game play in some early games could have been programming issues. However, with stronger hardware backing Maria up, one wouldn't have as much to overcome. Someone with your talent could have come up with tempest 2000 on it (since somebody's already done a decent job of doing that on the regular 8-bit). Would more memory allow better use of the 320 modes?

[DanBoris]

Is there away this could have been accomplished?

 

No.

No way possible, or no because it wouldn't have been the same 7800 (and different to develop for?).

 

I realize you have done some pretty impressive things with the 7800, which indicate that the weak game play in some early games could have been programming issues. However, with stronger hardware backing Maria up, one wouldn't have as much to overcome. Someone with your talent could have come up with tempest 2000 on it (since somebody's already done a decent job of doing that on the regular 8-bit). Would more memory allow better use of the 320 modes?

 

The 5200 cartridge port doesn't have the necessary signals coming out to it to allow this sort of addon.

 

Dan

[Lynxpro]

Has anyone ever tried adding the Maria chip to an Atari 8-bit computer?

[The_Laird]

Has anyone ever tried adding the Maria chip to an Atari 8-bit computer?

 

What would be the point? It wouldn't work.

[+darryl1970]

Has anyone ever tried adding the Maria chip to an Atari 8-bit computer?

 

What would be the point? It wouldn't work.

 

Well, this forum is sometimes for "what ifs." That's the "point!" Could Maria have been designed to replace the 8-bit video hardware instead of the 2600. I understand that Dan says the 5200 cartridge port wouldn't have worked. What about the bus on the back of the 5200? Could Maria have been designed for an expansion to plug into the main data bus and replace the 8-bit hardware?

 

Is it the underlying (2600) hardware considerations that cause 7800 animation to be a lot less smooth than the 8-bits, Coleco, or NES? Galaga is a GREAT example. The animation is so slow... and on higher (faster) levels, it seems they are just jumping more pixels instead of going faster. (Much like 2600 game animation) Mario seems to glide across the screen in Donkey Kong. It is obviously a port of the NES, but the jumps seem so floaty and wrong. It's as if I watch everything in slow motion. (In contrast, look at how smoothly the 8-bits animate Mario, all of the barrels, and the fireballs.) Mario Bros for the 7800 is atrocious! The main characters seem to float around.. There is just something that is not as fluent in the 7800 games as they are in the 8-bit counterparts... It's kind of like I can SEE the screen refreshing, and I almost notice each sprite moving one at a time (like 2600), where the 8-bit games seem to move all at once -- like the arcade, 5200/8-bit, NES or even Colecovision.

 

Maybe it's a programmer flaw. Maybe Galaga was made when people weren't familiar with the system... Donkey Kong and Mario Bros could just be lack of good programmers by the time the games were poorly ported from the NES.

 

I just wonder if an 8-bit sprite upgrade would have been a better choice instead. Pokey would have been standard, and the embarrassment of TIA sound would not add to the already poor animation capabilities of the 7800 in its current carnation.

 

And finally, is it me, or does this fantasy question seem to upset people? I think the 7800 could have been much more. It wasn't the arcade experience that was touted back in 83 or 84 (whenever the magazines started talking about it). I was disappointed that the Monsters in Ms Pac looked just as lame as the 5200 one, without flicker. I expected more. The sound wasn't as good as my 5200 sounds. I expected more. I expected resolution as good as the Coleco at least. Like PacMan Collection quality for the CV, but no flicker///

 

In hindsight, it was stupid to worry about backward compatibility at that point, because I don't think people cared anymore. I think they would have just wanted a more arcade-like experience. (Kind of like PS3 and XBox360 don't have to worry so much anymore). It was big at first (like with ps1, Coleco, and 5200), but they were trying to make up for a mistake that no longer mattered -- the 5200 was starting to win. So, I that's why I wondered if ANY expansion of the 8-bit/5200 could have been an option...

[AtariNerd]

For starters, the version of the 6500 series of cpu the 5200 uses doesn't have the neccesary halt lines that the Maria needs to take control, so you'd have to include that in an add-on, probably along with an external memory module and logic chips... see where this is going?

Edited February 6, 2011 by AtariNerd

[Metal Ghost]

Well, there was a 7800 Expansion Module planned for the 5200 ( http://www.atariage.com/forums/topic/167048-atari-7800-adaptor-for-atari-5200-at-pax-east/page__p__2064628__hl__7800__fromsearch__1?do=findComment&comment=2064628 ), so the original plan was to cover both bases (i.e. 7800 as well as 7800 hardware for the 5200).

Edited February 6, 2011 by Metal Ghost

[Rybags]

It could work, but the 5200 would in effect just be a giant peripheral base to the unit, doing no more than supplying input from the joysticks.

 

In a marketing sense, it's fair enough to build such a unit for an "inferior" system, e.g. 2600 running on Coleco or 5200, because you're dealing with something that might cost the price of a couple of games.

 

In the case of the 7800 unit on a 5200, about the only saving you'd have is the lack of switches and joystick ports, you could have a small ROM in the 7800 that runs on the 6502 of the 5200, reads the ports and switches and relays the info through where it can be latched by the 7800 unit.

 

But still, you'd have the possibility of the 7800 unit needing it's own power supply, and you'd probably have had the thing costing 75% the price of a standalone 7800.

 

 

Re - interfacing Maria - not really in both cases (A8/5200). Maria needs exclusive access to the RAM, you can "turn off" Antic, but it'll still be doing 9 Refresh cycles per scanline.

Alternatively, I guess you could have some system where Maria has it's own RAM which is accessable also by the A8/5200, but you'd probably only have the access during VBlank.

You also have the problem in that both systems are generating their own video signal, and you can't really merge them - for starters they'd be at different stages of a frame, plus doesn't Maria actually run at 7 or 14 MHz and have slightly different line duration?

Edited February 6, 2011 by Rybags

[The Usotsuki]

Except isn't the 5200's CPU the same 6502C as the 7800? Or at least that's what Wiki claims.

[kool kitty89]

My understanding is that the 5200 started to take over the market right before it was killed. This could indicate that 2600 compatibility was no longer much of an issue, and people were starting to look for more advanced games. I believe that 2600 compatibility weakened the 7800's overall abilities (hence the choppy game play on many 7800 games). I don't want this to be a bashing session of the 5200 or 7800 (including the 5200 joysticks), but a marketing idea and THEN whether it would have been possible.

If anything, that's an argument to leave the 5200 as-is rather than have add-ons.

 

Compatibility was an issue, but in the end it probably wasn't make or break for the system in the long haul. (ie if they'd pushed on with it) However it was still an issue, and all the more so since competition had a module for VCS compatibility from the start. (and while plenty of platforms have done well without compatibiltiy and it's rarely ever been a make or break feature of a game console, it is always a good addition as long as it can be done reasonably efficiently without crippling the system -in the 7800 it was done OK, though it was a fair bit tacked on and the system could have been cheaper and more capable if 2600 compatibility had never been a feature -the 3200 was probably a more efficient route, at least for its time, actually evolving most of the old hardware and building on it -it seems like they were planning to morph TIA into a GTIA like chip with backwards compatible modes and possibly even add some enhancements over the normal GTIA)

 

That said, you could certainly argue that the 7800 was easier to use to correct the issues with the 5200 (hardware wise) than the necessary marketing to make the 5200 really transition smoothly into the role of Atari's new mainstay console (with the 2600 pushing into the budget range). The 5200 had other issues besides compatibility, and of the hardware ones there's the general odd cost cutting in some areas and considerable lack of it in other areas. (it very feasibly could and should have been implemented to be a low cost, further consolidated derivative of the 1200XL or original 600 prototype -no PIA, 16k DRAM, smaller motherboard facilitated by such -as compact as possible without resorting to inefficiently tight placement of components, compact but sleek form factor, 2 controller ports -given the fact they were pretty much unused on the A8 and being removed on the 1200, simple conventional RF cable and separate power supply, relatively simple/cost effective joysticks out of the box with more advanced ones as accessories, etc -they could go the analog route with resistors and switches to allow a simple 8-way analog joystick -effectively digital from the user's side of things- or actually have a digital mode where POKEY was configured to have the POT lines used directly as I/O lines -you have 8 lines and 2 controllers, so you could configure them in leu of PIA while GTIA read the triggers and any additional signals)

Once CGIA was completed, things could be even more consolidated too, plus POKEY could be stripped of the keyboard/SIO logic for just sound+POT hardware and a 28-pin package. (perhaps initially the same -or slightly modified- die with the main change being the small package, followed by a formal redesign of the chip that removed the unnecessary features and also shrunk the die for a newer manufacturing process and dropping to a narrow 28-pin DIP instead of a wide DIP)

 

After the fact, many things could still have been corrected, but they'd gotten weaker first impressions because of that and weaker sales and support in general. The 5100 was still only a modest fix well short of an ideal redesign of the 5200. (the new controllers were nice too, though the buttons hadn't been fixed)

 

 

But then there's all the non hardware related issues that, generally speaking, were the real problems with Atari at the time. Many of them are not directly tied to the 5200 specifically, though the 5200's problems were symptoms of many underlying issues at Atari Inc. There were various management problems, conflicts of interest, and other issues that needed to be addressed ASAP. They forced Atari into a downward spiral in late 1982 that was only finally getting corrected by Morgan in early '84 (with efforts starting in late '83). Had someone line Morgan come in by mid 1982 (let alone instead of Kassar from the start), disaster may have been averted entirely, but that wasn't the case. The biggest single issue was almost certainly the problematic distribution network that was showing demand/sales figures that were heavily inflated over what was actually occurring on the market. (it was growing to be sure, but not nearly as fast as Atari management was being led to believe -not only pushing for oversaturation on the hardware side, but also inducing the so-called glut of games on the market)

Other issues included dual management conflicts (and conflicts of interest) with Warner and Atari Inc: one blatant example is what happened with ET. (Kassar turned the license down after Universal asked too much, but Warner then made their own deal with Spielberg directly -tying into attempts to woo him to Warner Bros studios iirc- and on top of a massive license fee of $20-25 million he made the demand that the game be released by Christmas; Kassar made the call to produce the game in such large quantities only because it was felt that that was the minimum number that could make a profit after the heavy license fee -the alternative would be risking much less investment in massive production, but definitely taking a net loss -it would have been a safer bet though, and definitely the right move in hindsight, plus there was potential in additional production if the game did sell well or possibly revised versions of the game for the VCS or different games tying into the license like the later ET Phone home on the A8)

 

 

All those problems affected every area of Atari from the computers to the arcade to the home consoles and the related software. As such, Atari had much higher priority issues on their plate than retaining the near monopoly of the home game market. (as it was, without the 5200, the VCS was still holding strong as the market leader -if declining as well) Hell, even after rectifying the management issues, the issues in the computer side of things would probably be a higher priority to deal with than getting a new console out ASAP. (or you could kill 2 birds with 1 stone with something like a 1982 counterpart to the XEGS . . . or just release the 600 as originally intended and position that as the low-end consumer/game oriented computer -and high end game system- in 1982 -sort of like the 400 had originally been intended to be positioned as in 1979, but generally became impractical as such due to the high price -vs the 600 in 1982 which feasibly could have been in the same range as the 5200 at the time, or possibly lower -certainly in the $200 range by the end of the year -or maybe have the game version of the 600 even lower cost with the 400's cheap keyboard, preferably with provisions for an upgrade to full keyboard with minimal hassle and optimal cost efficiency -maybe even modular design such that the same case could be used for full and membrane keyboard units)

 

Sticking with the 3200 even if it meant delays also might have been the preferable alternative. (a low-cost computer would have been good in either case)

The 3200 was very much the same general concept as the 7800, though implementation was rather different. (a "quick fix" hack might have been to drop the new STIA chip in favor of plopping in both TIA and GTIA instead and later planning on merging them -or at least cutting cost with the release of GTIA, and then you'd really have something like the 7800 with CGIA instead of MARIA, and probably not using TIA's I/O at all in 3200 mode since you have the GTIA I/O as well as RIOT -if you enabled RIOT interrupts, that could facilitate CPU modulation tricks for driving GTIA's beeper/click sound channel to supplement TIA sound -it would be fixed volume though, but modulation tricks played on TIA channels could use hardware volume, or modulate volume for sample playback )

 

 

Let's say Atari completed the 5200 self-centering prototypes (similar to PS2 type mechanisms??), I think the joysticks would have been less to complain about as long as they didn't break. They already had the 2600 add-on.

The revised joysticks were more like the original (supposed) concept/design inspired by RC car/plane transmitters/controllers (2-axis spring loaded potentiometer assembly) and was packed into a tight module somewhat like the vectrex used iirc and very much like modern analog pot modules on gamepads from the current and past 2 generations. (except the N64 and Saturn 1st party controllers used mechano-optical and magnetic mechanisms rather than true analog oddly enough -Gravis's Xterminator used magnetic too, like the Saturn)

 

The PS2's pressure sensitive buttons are different than that, unless you just mean the dual analog sticks. (which should be pot modules like they were on the PS1 dual analog/dual shock controllers)

 

 

 

 

 

Is there away this could have been accomplished? I am not familiar with the hardware enough, but I understand the 2600 module is a 2600 that bypasses the 5200 hardware. Would the 7800 module have been able to only bypass the video chip from the cartridge port? Could they have used the bus on the back of the 5200 machines? I just think this would have saved Atari some. People who adopted the 5200 wouldn't have felt abandoned, and the 5200 would still play 2600 with a module. It would make the 7800 almost like an XM expansion, minus the Yamaha sound chip. Improved analog joysticks, along with the keyboard, would have allowed for more controller options. I just wonder why this was never thought of as an option. Additionally, it would have given Atari the image of always being expandable (one of Coleco's initial selling points!)

The add-on concept would have been a waste as the 5200 didn't have provisions for such unfortunately. (and again, it's rather overkill for the time, especially since the system needed to push for lower cost more than more capabilities)

Added RAM might have been the most useful thing for the 5200, or maybe provisions for added sound (as the Famicom used very well in Japan, as did the MSX to some extent among others), but it lacked the read/write and Phi2 lines needed for use of RAM or analog audio input. (except on late/modified 4 port and all 2 port models with VCS module support -with phi2 and analog audio input; read/write could be hacked fairly easily with phi2 there)

 

With the VCS compatible module models, any comprehensive add-on would really be a plug-in replacement system rather than piggybacking on the old hardware: the VCS module used nothing but the power and RF out of the 5200. (like the CV add-on, it was a total VCS in a box)

The planned 7800 module for the 5200 would have been the same: an entire 7800 in a box that plugged into the 5200, not very cost effective (probably moderately cheaper to produce than a full 7800), but a good idea to help bridge 5200 owners over to promote good PR and more satisfied customers. Actually, I think the VCS (or at least 7800) modules would have needed separate power input, so that's more cost and even less convenient. (if they could piggyback on the 5200's voltage regulator and PSU, that would have simplified cost and ease of installation, but I'm not sure that was the case)

 

 

 

As for using MARIA on later models of the computers, that might have been possible, but not not particularly efficient. (vs evolutionary enhancements of ANTIC+GTIA) And again, you could argue consolidation of the existing chipset was more important than enhancements anyway. (or break from the architecture entirely with a new design -which is what the advanced technology division was pushing for with several advanced 16-bit chipsets -and those were well beyond MARIA) More RAM a la XE would have been the most straightforward expansion, perhaps dual POKEYs. (a faster CPU might have been possible, but I'm not sure -it would depend on cost/availability of DRAM -3 or perhaps even 4 MHz rated CPUs shouldn't have been unreasonable, but the main issue was having fast enough memory to facilitate that, and anything short of 3.58 MHz would have needed a different system clock -unless they were already using 7.16 MHz)

And if they WERE t upgrade GTIA/ANTIC, higher dot clock (probably double resolution) versions of all the existing graphics modes might have been one of the more straightforward enhancements. (an added 80 pixel wide direct 256 color mode would have been nice too, let alone more conventional indexed color character modes -be it internal or external CRAM/tables- or at least bumping it to 16 color registers to allow 16 freely selected colors for 16 color character/bitmap modes)

Edited February 6, 2011 by kool kitty89

[+darryl1970]

AtariNerd, Metal Ghost, Rybags, The Usotsuki, kool kitty89, and any others I am missing... Thanks for an entertaining discussion. I figure we all have various levels of technical knowledge on the different Atari system, and our experiences "color" our memories. I have read posts from people in other countries that think the 2600 Pac-Man is better than the arcade, because that's what they grew up with. I remember seeing the 400/800 version of Pac-Man at a store kiosk, and I thought it was the most incredible thing ever! (LMAO) When I saw the 5200 was the same WITH intermissions, I was excited, and I had to sell my 2600 to get one. The 5200 holds my greatest memories. I liked the joysticks, but was blessed with the knowledge to keep the working when they broke.. lol. I thought the non-centering thing was not an issue..

 

The reason I share is that it shows where my allegiance was back in the day. I thought Coleco had nicer sprites, but Atari had better color, speed, sound, and programmers (Atari could squeeze a LOT out of 16k!) When I heard the 5200 had taken the lead, I was excited and proud. When I heard about the 7800, I was really excited, but I was bummed that they gave up on the 5200 just when it was starting to win. My thought was how excited I would have been with some kind of expansion (such as the super game module for Colecovision -- back then and the current Ops version). I would have been really pumped. However, having seen the "future" now, I know it didn't work so hot for Sega and the Sega CD.. lol. All I know is that I was disappointed by the 7800's low-res sprites, because it was the thing I hate most when I'd go between a C-64 and AtariXL. I really hated the 160 resolution -- especially since they were single color "Players". However, being a musician, I REALLY dispised the 7800 sound limits. I felt that it was a one step forward and two back.

 

But I still love it, because it's part of my history, and things happened the way they did for bigger reasons than all of us, but it was fun fantasizing.. And thanks!

 

Kool Kitty89,You made a lot of interesting points. I guess expanding on the current 8-bit chipset would have been interesting. I guess they STILL could have named that Maria.. lol. At first I didn't get your comment that I was making an argument to continue with the 5200... But, the more I thought about it, you're right! I think it had a lot of untapped potential (look at Tempest X-Treme for the 8-bits). I guess, in my dreams, a 5200 with a Maria chip would be better than a 7800. Higher-Res, multi-colored sprite would have made it a Jaguar.. lol

 

One final question.... Is that that a parallel bus on the back of my 5200? I dread digging it out; but, if I remember correctly, it looks a lot like the one on the back of my 800xl...

[DracIsBack]

Is it the underlying (2600) hardware considerations that cause 7800 animation to be a lot less smooth than the 8-bits, Coleco, or NES?

 

Just curious ... how many 7800 games have you played? In some cases, it could be hardware. In others, it could be bad programming.

 

Maybe it's a programmer flaw. Maybe Galaga was made when people weren't familiar with the system..

 

It was one of the first games made for the 7800.

 

 

Donkey Kong and Mario Bros could just be lack of good programmers by the time the games were poorly ported from the NES.

 

They were indeed done in an era of penny pinching.

 

would not add to the already poor animation capabilities of the 7800 in its current carnation.

 

 

I'd disagree about it having poor animation capabilities.

 

It's sort of like me saying,

 

"Well, the A8 version of Desert Falcon has the chunkiest scrolling, therefore, it's far worse than the 7800 at animation because the 7800 version is pretty smooth."

 

There are better examples of scrolling on the A8 and one game is not reflective of the library.

 

or

 

"The A8 version of Dark Chambers has all kinds of flickering when the sprites move, whereas the 7800 version does not. The A8 must be terrible at moving objects."

 

 

I expected resolution as good as the Coleco at least. Like PacMan Collection quality for the CV, but no flicker///

 

1) Pac Man Collection was made in 200X, not in 1983 so it's irrelevant to this discussion ... no one in 1983 was playing that game either. But I agree, Ed did an awesome job.

 

As for resolution, this isn't cut and dry either.

 

On one hand, many games do indeed use the lower resolution 160 mode. However, other games use the higher 320 mode. You also have fewer sprites, fewer colors, fewer display tricks, a lack of hardware scrolling etc.

 

but they were trying to make up for a mistake that no longer mattered -- the 5200 was starting to win.

 

I think that's an urban myth. I have never seen any stats for either the Coleco or the Atari 5200 other than the usual "I heard from someone, who heard from someone, who heard from someone" variety. I don't believe that it was "starting to win" and have never seen any NPD data to back it up. Same for the Coleco proving it "far outsold the 5200". Just a lot of questionable claims by fanboys.

[Phredreeke]

Is 160x200 resolution really that bad? Does 96 additional pixels of horizontal resolution really make a difference on an 80s TV with an RF cable?

[kool kitty89]

AtariNerd, Metal Ghost, Rybags, The Usotsuki, kool kitty89, and any others I am missing... Thanks for an entertaining discussion. I figure we all have various levels of technical knowledge on the different Atari system, and our experiences "color" our memories. I have read posts from people in other countries that think the 2600 Pac-Man is better than the arcade, because that's what they grew up with. I remember seeing the 400/800 version of Pac-Man at a store kiosk, and I thought it was the most incredible thing ever! (LMAO) When I saw the 5200 was the same WITH intermissions, I was excited, and I had to sell my 2600 to get one. The 5200 holds my greatest memories. I liked the joysticks, but was blessed with the knowledge to keep the working when they broke.. lol. I thought the non-centering thing was not an issue..

The 5200 controllers aren't as bad as some make out, but they were far from ideal. (analog was unnecessary much of the time, and plain digital -or pull up resistors wiring switches to analog lines- would have been simpler, cheaper, and more reliable) For the few games analog was needed for, they could have offered accessories. (dedicated paddles would have been nice too)

The keypads were also generally unnecessary and could have been accessories to simplify the design and reduce cost.

The flex circuitry ended up being unreliable, not one of the better cost trade-offs over using a PCB. (the side buttons should have been VCS type tact/dome switches, or at least rubber dome switches/chicklet buttons over sturdy PCB contacts as later gamepads and such used -something like the 7800 buttons)

 

The Intellivision and Coleco controllers were problematic too, though the CV ones were OK for a lot of stuff at least. (best for games not using the buttons and allowing the stick to be used with your thumb)

The 7800 controllers fixed some things but broke others: they were a bit too big and the joystick was too stiff and bulky. (a slightly scaled down and shortened VCS stick would have been nice -prior to the joypads at least)

Some like the idea of built-in paddles a la 2700, but I really don't like joystick with bulky knobs and especially ones that turn. (full sized arcade sticks are one thing, but compact home controllers are another entirely -though I prefer arcade sticks that are locked from spinning as well) Having separate paddles and sticks is much better, and the VCS paddles were fine. (maybe have 2 buttons, but otherwise they're fine ergonomically and functionally)

 

 

However, having seen the "future" now, I know it didn't work so hot for Sega and the Sega CD.. lol. All I know is that I was disappointed by the 7800's low-res sprites, because it was the thing I hate most when I'd go between a C-64 and AtariXL. I really hated the 160 resolution -- especially since they were single color "Players". However, being a musician, I REALLY dispised the 7800 sound limits. I felt that it was a one step forward and two back.

The 7800 beats the C64's VIC-II (and ANTIC+GTIA) in pretty much every respect capabilities wise: more and larger sprites, more color for playfield and sprites, etc, etc.

 

The 7800 sound was bad in hindsight (intended to add only via cart expansion), but in the context of the mid 1984 release it wasn't that huge of a deal. (plus they could have used an add-on module rather than on-cart chips for a 1 time cost to consumers, though none of that ever happened) I think they probably could have crammed in a POKEY in '84 with some trade-offs, or at very least a tiny SN76489 PSG. (if they went off the shelf) But that's not what GCC ended up doing. (the easiest way to add more board space would have been to make the RF modulator external)

 

The 7800 was designed to be extremely cost cut, so it made some sense in that context. (albeit, with the 3200 instead of the 5200 in the first place, TIA sound would have been much more acceptable, and that's if S-TIA didn't add more sound hardware -or the TIA+GTIA route, but that would add the GTIA click/beeper channel and RIOT's timer could have been useful for interrupts for CPU modulation of that 3rd channel -or the TIA channels)

 

 

The resolution was fine for the time . . . though a better argument is that the 320 wide mode should have been a 256 wide mode with more color flexibility. (the 160 wide mode can push a lot more color -among other things- than the NES can per scanline or on screen, let alone the old TMS9918 VDP)

 

The 7800 didn't get nearly enough support for it to demonstrate its full capabilities. (the 5200/A8 either -aside from some homebrew stuff, and the CV more or less through the MSX1 and SG-1000)

 

 

 

One final question.... Is that that a parallel bus on the back of my 5200? I dread digging it out; but, if I remember correctly, it looks a lot like the one on the back of my 800xl...

It's a very limited expansion port (see pinouts online) and most likely was intended for a disk and/or casette interface.

 

 

 

 

 

 

Is it the underlying (2600) hardware considerations that cause 7800 animation to be a lot less smooth than the 8-bits, Coleco, or NES?

 

Just curious ... how many 7800 games have you played? In some cases, it could be hardware. In others, it could be bad programming.

Animation could be a ROM limit. (trade-off between graphical detail and animation with a limited ROM size)

 

Maybe it's a programmer flaw. Maybe Galaga was made when people weren't familiar with the system..

 

It was one of the first games made for the 7800.

And in spite of that it looks better than the NES version in some ways. (and Joust in pretty much all ways -generally better sounding too)

 

would not add to the already poor animation capabilities of the 7800 in its current carnation.

 

 

I'd disagree about it having poor animation capabilities.

Animation isn't a programming limitation, it's an art decision when ROM space is limited. (animation = ROM space)

 

 

"The A8 version of Dark Chambers has all kinds of flickering when the sprites move, whereas the 7800 version does not. The A8 must be terrible at moving objects."

Well, on that note, the 7800 IS more powerful by a good margin. (best case on A8 with hardware+software sprites -be it bitmap or charmode stuff- will be more limited than the best case on the 7800 in every respect)

 

I expected resolution as good as the Coleco at least. Like PacMan Collection quality for the CV, but no flicker///

 

1) Pac Man Collection was made in 200X, not in 1983 so it's irrelevant to this discussion ... no one in 1983 was playing that game either. But I agree, Ed did an awesome job.

 

As for resolution, this isn't cut and dry either.

Resolution is significant, but not the end all be all. Many other things make a difference from art design to color depth. (Doom on the Jaguar runs at 160 pixels wide ) The SNES, most PCE/TG-16 games, SMS, CV/SG-1000/MSX1/etc all run at the same resolution and dot clock (ie same pixel shape), but with drastically different graphics.

 

On one hand, many games do indeed use the lower resolution 160 mode. However, other games use the higher 320 mode. You also have fewer sprites, fewer colors, fewer display tricks, a lack of hardware scrolling etc.

Were there any commercial 7800 games in high res? The limitations of those modes are generally not worth it compared to the flexibility of the 160 wide modes. (ie most common games of the period will look much better at 160 width, and again better than C64 counterparts -or A8 for that matter, and better than the NES with trade-offs -more color, more sprites without flicker, etc)

 

but they were trying to make up for a mistake that no longer mattered -- the 5200 was starting to win.

 

I think that's an urban myth. I have never seen any stats for either the Coleco or the Atari 5200 other than the usual "I heard from someone, who heard from someone, who heard from someone" variety. I don't believe that it was "starting to win" and have never seen any NPD data to back it up. Same for the Coleco proving it "far outsold the 5200". Just a lot of questionable claims by fanboys.

The 5200 might have pulled through in the long run, but the 7800 could have been better (neither would have been preferable with something other than the 5200 from the start followed by a true successor to that in the late 80s -or nothing but the VCS and computers until they got their other problems settled, or straight jump to the 7800 -but the 3200 would effectively be the same context but sooner).

There's no sold sales figures for either platform though. (and again, there's no telling if the 5200 could have been as successful -or more- than the 7800 with any number of variables, though I think most would agree that sticking with the 3200 in the first place would have cut out a lot of the confusion and been better in the long run than switching things after the fact))

 

In any case, Atari had much bigger issues than competition at the time. The #1 problem was the internal management issues and related problems with the industry. (especially the distribution network and related inflation issues -and dual mangement conflicts on top of fundamentally problematic management)

 

I already addressed this above though.

Edited February 7, 2011 by kool kitty89

[+darryl1970]

Is 160x200 resolution really that bad? Does 96 additional pixels of horizontal resolution really make a difference on an 80s TV with an RF cable?

 

Honestly, it's a HUGE difference. Try drawing a Mario that looks just like the arcade, colecovision, or NES. He'll look like a big-nosed freak. Look at the Monsters on Coleco Pac-Man collection vs the 7800 version. Both great games, but the 7800 can't produce the roundness nor the monster bottoms accurately, in an area that would fit between two corridors. It really does make a difference. I love my 5200/8-bits, but I have always been disappointed by the 160 resolution, as it looks obviously blocky. (And is the same as the 2600) I was especially disappointed when I saw that the original 7800 Ms Pac-Man monsters were the same as my 5200. These things made a big deal to me back then...

 

I would say it DOES... And the Coleco and NES are only 256 resolution!

[+darryl1970]

Is it the underlying (2600) hardware considerations that cause 7800 animation to be a lot less smooth than the 8-bits, Coleco, or NES?

 

Just curious ... how many 7800 games have you played? In some cases, it could be hardware. In others, it could be bad programming.

 

I have played MOST of the games on REAL hardware. I do own a 7800. Emulation is pretty spot on too.

 

Maybe it's a programmer flaw. Maybe Galaga was made when people weren't familiar with the system..

 

It was one of the first games made for the 7800.

As was Dig Dug, Joust, Ms. Pac-Man, etc. I was trying to excuse that Galaga nightmare. What an embarrassment, from the poor jerky alien formations to the giant, FUZZY sprites. The player ship looks like it melted.

 

Donkey Kong and Mario Bros could just be lack of good programmers by the time the games were poorly ported from the NES.

 

would not add to the already poor animation capabilities of the 7800 in its current carnation.

 

I'd disagree about it having poor animation capabilities.

 

It's sort of like me saying,

 

"Well, the A8 version of Desert Falcon has the chunkiest scrolling, therefore, it's far worse than the 7800 at animation because the 7800 version is pretty smooth."

 

There are better examples of scrolling on the A8 and one game is not reflective of the library.

 

or

 

"The A8 version of Dark Chambers has all kinds of flickering when the sprites move, whereas the 7800 version does not. The A8 must be terrible at moving objects."

 

I worded this wrong. I should not have used the word, "ANIMATION", because I am not talking about missing frames. The 8-bits seemed to move everything at once -- each screen refresh seemed different. The alien in Galaga seem to have large gaps in their flight pattern... The only way I can explain this is by plotting a sign wave. If one were to move one degree at a time, you'd have a pretty solid wave; however, if you skip every 4 degrees, the points would be spread apart with gaps. It feels like there are a LOT of gaps in the alien's pattern compared to the NES.. On the higher levels, it feels like the aliens just skip more pixels vs opening up the throttle on them. These are animation traits I notice in quite a few 7800 games, but not 8-bit..

 

Another example would be something as simple as Mario's jump in DK. It's almost as if Mario follows an awkward pattern... And he just kind of floats in the air vs a solid jump. The 8-bits even included the velocity of the arcade. (I think the NES had it close, but the barrels moved too slowly)

 

Another example, which could be poor programming is Mario Bros on the 7800. It seems like Mario's jumps are jerky. I figure that is more of a programming thing, or does the 7800 really not have enough power to just move everything smoothly? I question if Galaga is a lack of processing. It get the impression that Maria is impressive for display and color... HOWEVER, ANTIC, GTIA, & POKEY free up the processor more to calculate things like flight patterns and jump. Does the programmer have to handle display sync more on the 7800 than the 8-bits (kind of like the 2600)?

 

Additional, like the 2600, the controls on many games seem less responsive. It's kind of like a delay. Is this because Pokey takes some of the reading duties off of the processor?

 

I expected resolution as good as the Coleco at least. Like PacMan Collection quality for the CV, but no flicker///

 

1) Pac Man Collection was made in 200X, not in 1983 so it's irrelevant to this discussion ... no one in 1983 was playing that game either. But I agree, Ed did an awesome job.

 

I disagree, because it is still the Coleco's native resolution and the same quality of graphics COULD have been achieved at ANY point of the CV's lifespan. It was simply an example that the 7800 couldn't match at 160 resolution.

 

As for resolution, this isn't cut and dry either.

 

On one hand, many games do indeed use the lower resolution 160 mode. However, other games use the higher 320 mode. You also have fewer sprites, fewer colors, fewer display tricks, a lack of hardware scrolling etc.

 

I never saw a 320 game (or I should say recognized) until the homebrew Frogger. Does that mode eat up more of the 7800's "thinking" time too?

Edited February 7, 2011 by darryl1970

[Phredreeke]

The Amstrad CPC version of Donkey Kong does a decent job of drawing Mario in 160 pixel resolution.

Edited February 7, 2011 by Phredreeke

[+darryl1970]

The Amstrad CPC version of Donkey Kong does a decent job of drawing Mario in 160 pixel resolution.

 

Very good point! I guess my mind was kind of in a rut, because it's obvious that the Donkey Kong references are more of an artist flaw than limit of the 160. The same argument could be posed for BOTH C-64 versions. (Like wise the birds in Joust were nicely designed) However, it's still not as "sharp" as Coleco's or Nintendo, but the 320 mode would have.

 

I think I am stuck in that disappointment that the Pac-Monsters never had the "right" bottoms to them. There is a post on here were somebody was creating a "new" Pac-Man for the 8-bits, which had nicely rounded eyes and the bottoms were correct. However, he is scrolling the screen so they can be wider. Mario was a bad example, because Mario CAN be wider, since he doesn't have to "fit" in the Vertical corridor. (I know...Apparently Namco/Tengen didn't think is was necessary for that on the NES version either.)

 

Nice catch on the Mario, because I was impressed with what Ocean did with the Amstrad/C64 versions. I JUST wish Ocean would have knocked Mario one pixel shorter on the "lean forward" frame, so he didn't walk like he had a stick somewhere... lol. I digress, although it would be a nice, simple hack...

[Bruce Tomlin]

I recall seeing that they already had a 7800 module design in the works for the 5200, though it was apparently early enough in the 7800 development process that it didn't have a boot ROM (so it pre-dated the digital signature).

 

They already had to re-design the 5200 just to make the 2600 module work, but just like it this would be a whole 7800 minus the RF modulator, so what's the point? The 5200 was large enough as it was, and adding a vertical module just made it worse.

 

The only thing it would have gotten you is that parents wouldn't see two video game systems and say "Why don't we get rid of your old one? Your cousin would love to have it." (Actually, that's exactly what happened to me, except there wasn't a 2600 module then, so all the 2600 games went out the door.)

[kool kitty89]

Is 160x200 resolution really that bad? Does 96 additional pixels of horizontal resolution really make a difference on an 80s TV with an RF cable?

 

Honestly, it's a HUGE difference. Try drawing a Mario that looks just like the arcade, colecovision, or NES. He'll look like a big-nosed freak. Look at the Monsters on Coleco Pac-Man collection vs the 7800 version. Both great games, but the 7800 can't produce the roundness nor the monster bottoms accurately, in an area that would fit between two corridors. It really does make a difference. I love my 5200/8-bits, but I have always been disappointed by the 160 resolution, as it looks obviously blocky. (And is the same as the 2600) I was especially disappointed when I saw that the original 7800 Ms Pac-Man monsters were the same as my 5200. These things made a big deal to me back then...

 

I would say it DOES... And the Coleco and NES are only 256 resolution!

There's absolutely MASSIVE trade-offs though, the CV's color and graphics capabilities are extremely constrained, even heavy trade-offs against the A8, let alone the 7800's much more flexible color/sprite capabilities. (the 7800 has some significant advantages over the NES for that matter)

 

Good art design is a big part of it too, and catering to pixel aspect ratio is important in all cases. (including the NES and CV where pixels are still a good bit wider than square -MUCH more so in PAL where even the 320 wide C64/A8/Amiga pixels -or others at 7.16 MHz- are very slightly wider than square -ST is slightly taller than Square and MUCH taller than square on NTSC TVs -custom calibrated monitors are a separate issue, though also tend to be tall due to removing overscan evenly as with the Amiga and CGA -or various other 320x200 modes including VGA 13h)

I'm not sure all of the above pictures are properly scaled to depict true NTSC resolution. (many emulators show A8/7800 pixels too wide as such, with 160 res being 2x1 pixels on a 320x240 screen -or scaled up from there- where it should actually be narrower than that; likewise many NES/CV/etc emulators will use pure square pixels instead, thus not making the pixels as wide as they are on real TVs -both 3.58 MHz/160 and 5.37 MHz/256 pixels are between 1:1 and 2:1 PAR with 3.58 being roughly 1.72:1 and 1.15:1)

 

Again, it's a shame that the 7800 didn't offer a 5.37 MHz mode with more flexibility than 7.16 MHz 320 pixels would allow, or short of that, maybe allow added flexibility with 7.16 MHz clipped to 256 pixels wide and more overscan (like the CoCo among others -apple II does that at 280 pixels wide), let alone allowing the full-featured 160 pixel res set to 7.16 MHz (and thus more than 1/2 the screen as horizontal overscan/boarder, but great for vertically oriented arcade games or other cases), though the latter could also have been great (if not better) in 5.37 MHz with 160 pixels and ~35% horizontal boarder and NES/SNES/CV/etc shaped pixels. (given how much 7.16 MHz tends to artifact in NTSC, especially with video encoders and TVs of the time, offering only 3.58 MHz and 5.37 MHz dot modes would have been preferable, and having 160 pixels and full color/graphics capabilities in 5.37 MHz mode would be great -having all 3 would be great too though- actually, allowing 256 pixels -or a bit less even- for 3.58 MHz could mean making a game with zero horizontal boarder and possibly easing H-scroll since you'd hide the edges off-screen)

 

 

In any case, with a bit of tweaking, you can make good trade-offs for lower res stuff, let alone games designed for the system in the first place. (or more custom tailored to it).

Like this:

http://www.youtube.com/watch?v=qD-ok15kdZA

 

And again, that's a conversion of graphics not tailored to the resolution vs something like Giana Sisters on the C64 which was designed for that resolution.

 

And, of course, the 7800, A8, and C64 versions of Donkey Kong look much better than the CV version, in spite of the lower resolution. (or the fact that the 7800 version is a rather bare bones/hacked conversion of the NES game rather than a direct arcade conversion focusing on the 7800 specifically -same for Mario Bros)

 

 

All systems also suffer from varying degrees of artifacts in composite video or RF, or the additional washed-out color of the 7800 due to the mechanism used to short TIA and MARIA video together. (avoiding the need for any sort of switching mechanism) The NES has some nasty dot crawl, but clean RF (and the dot crawl isn't very noticeable on poorer/blurry TVs).

 

 

 

 

I worded this wrong. I should not have used the word, "ANIMATION", because I am not talking about missing frames. The 8-bits seemed to move everything at once -- each screen refresh seemed different. The alien in Galaga seem to have large gaps in their flight pattern... The only way I can explain this is by plotting a sign wave. If one were to move one degree at a time, you'd have a pretty solid wave; however, if you skip every 4 degrees, the points would be spread apart with gaps. It feels like there are a LOT of gaps in the alien's pattern compared to the NES.. On the higher levels, it feels like the aliens just skip more pixels vs opening up the throttle on them. These are animation traits I notice in quite a few 7800 games, but not 8-bit..

The NES game was a later, higher-budget game though, and still it's pretty funky with flaws: namely the really choppy movement of the ships in formation and the occasional amount of flicker.

 

 

 

 

But back to the original topic: they WERE planning MARIA as an add-on for the 5200 in the form of the 7800 module. (a rather good marketing/PR move to bridge the gap for 5200 and allow it to be phased out more smoothly)

 

 

 

I recall seeing that they already had a 7800 module design in the works for the 5200, though it was apparently early enough in the 7800 development process that it didn't have a boot ROM (so it pre-dated the digital signature).

 

They already had to re-design the 5200 just to make the 2600 module work, but just like it this would be a whole 7800 minus the RF modulator, so what's the point? The 5200 was large enough as it was, and adding a vertical module just made it worse.

 

The only thing it would have gotten you is that parents wouldn't see two video game systems and say "Why don't we get rid of your old one? Your cousin would love to have it." (Actually, that's exactly what happened to me, except there wasn't a 2600 module then, so all the 2600 games went out the door.)

It was a smart move to improve PR and brand loyalty for 5200 owners (with the 5200 being discontinued), especially if offered at tight price (or maybe a rebate with proof of purchase of a 5200).

 

I agree on the vertical module form factor, maybe they could have made it into something more like Curt's XM or such without adding to cost.

Edited February 10, 2011 by kool kitty89

[+darryl1970]

Is 160x200 resolution really that bad? Does 96 additional pixels of horizontal resolution really make a difference on an 80s TV with an RF cable?

 

Honestly, it's a HUGE difference. Try drawing a Mario that looks just like the arcade, colecovision, or NES. He'll look like a big-nosed freak. Look at the Monsters on Coleco Pac-Man collection vs the 7800 version. Both great games, but the 7800 can't produce the roundness nor the monster bottoms accurately, in an area that would fit between two corridors. It really does make a difference. I love my 5200/8-bits, but I have always been disappointed by the 160 resolution, as it looks obviously blocky. (And is the same as the 2600) I was especially disappointed when I saw that the original 7800 Ms Pac-Man monsters were the same as my 5200. These things made a big deal to me back then...

 

I would say it DOES... And the Coleco and NES are only 256 resolution!

There's absolutely MASSIVE trade-offs though, the CV's color and graphics capabilities are extremely constrained, even heavy trade-offs against the A8, let alone the 7800's much more flexible color/sprite capabilities. (the 7800 has some significant advantages over the NES for that matter)

 

Good art design is a big part of it too, and catering to pixel aspect ratio is important in all cases. (including the NES and CV where pixels are still a good bit wider than square -MUCH more so in PAL where even the 320 wide C64/A8/Amiga pixels -or others at 7.16 MHz- are very slightly wider than square -ST is slightly taller than Square and MUCH taller than square on NTSC TVs -custom calibrated monitors are a separate issue, though also tend to be tall due to removing overscan evenly as with the Amiga and CGA -or various other 320x200 modes including VGA 13h)

I'm not sure all of the above pictures are properly scaled to depict true NTSC resolution. (many emulators show A8/7800 pixels too wide as such, with 160 res being 2x1 pixels on a 320x240 screen -or scaled up from there- where it should actually be narrower than that; likewise many NES/CV/etc emulators will use pure square pixels instead, thus not making the pixels as wide as they are on real TVs -both 3.58 MHz/160 and 5.37 MHz/256 pixels are between 1:1 and 2:1 PAR with 3.58 being roughly 1.72:1 and 1.15:1)

 

Again, it's a shame that the 7800 didn't offer a 5.37 MHz mode with more flexibility than 7.16 MHz 320 pixels would allow, or short of that, maybe allow added flexibility with 7.16 MHz clipped to 256 pixels wide and more overscan (like the CoCo among others -apple II does that at 280 pixels wide), let alone allowing the full-featured 160 pixel res set to 7.16 MHz (and thus more than 1/2 the screen as horizontal overscan/boarder, but great for vertically oriented arcade games or other cases), though the latter could also have been great (if not better) in 5.37 MHz with 160 pixels and ~35% horizontal boarder and NES/SNES/CV/etc shaped pixels. (given how much 7.16 MHz tends to artifact in NTSC, especially with video encoders and TVs of the time, offering only 3.58 MHz and 5.37 MHz dot modes would have been preferable, and having 160 pixels and full color/graphics capabilities in 5.37 MHz mode would be great -having all 3 would be great too though- actually, allowing 256 pixels -or a bit less even- for 3.58 MHz could mean making a game with zero horizontal boarder and possibly easing H-scroll since you'd hide the edges off-screen)

 

 

In any case, with a bit of tweaking, you can make good trade-offs for lower res stuff, let alone games designed for the system in the first place. (or more custom tailored to it).

Like this:

http://www.youtube.com/watch?v=qD-ok15kdZA

 

And again, that's a conversion of graphics not tailored to the resolution vs something like Giana Sisters on the C64 which was designed for that resolution.

 

And, of course, the 7800, A8, and C64 versions of Donkey Kong look much better than the CV version, in spite of the lower resolution. (or the fact that the 7800 version is a rather bare bones/hacked conversion of the NES game rather than a direct arcade conversion focusing on the 7800 specifically -same for Mario Bros)

 

 

All systems also suffer from varying degrees of artifacts in composite video or RF, or the additional washed-out color of the 7800 due to the mechanism used to short TIA and MARIA video together. (avoiding the need for any sort of switching mechanism) The NES has some nasty dot crawl, but clean RF (and the dot crawl isn't very noticeable on poorer/blurry TVs).

 

 

 

 

I worded this wrong. I should not have used the word, "ANIMATION", because I am not talking about missing frames. The 8-bits seemed to move everything at once -- each screen refresh seemed different. The alien in Galaga seem to have large gaps in their flight pattern... The only way I can explain this is by plotting a sign wave. If one were to move one degree at a time, you'd have a pretty solid wave; however, if you skip every 4 degrees, the points would be spread apart with gaps. It feels like there are a LOT of gaps in the alien's pattern compared to the NES.. On the higher levels, it feels like the aliens just skip more pixels vs opening up the throttle on them. These are animation traits I notice in quite a few 7800 games, but not 8-bit..

The NES game was a later, higher-budget game though, and still it's pretty funky with flaws: namely the really choppy movement of the ships in formation and the occasional amount of flicker.

 

 

 

 

But back to the original topic: they WERE planning MARIA as an add-on for the 5200 in the form of the 7800 module. (a rather good marketing/PR move to bridge the gap for 5200 and allow it to be phased out more smoothly)

 

 

 

I recall seeing that they already had a 7800 module design in the works for the 5200, though it was apparently early enough in the 7800 development process that it didn't have a boot ROM (so it pre-dated the digital signature).

 

They already had to re-design the 5200 just to make the 2600 module work, but just like it this would be a whole 7800 minus the RF modulator, so what's the point? The 5200 was large enough as it was, and adding a vertical module just made it worse.

 

The only thing it would have gotten you is that parents wouldn't see two video game systems and say "Why don't we get rid of your old one? Your cousin would love to have it." (Actually, that's exactly what happened to me, except there wasn't a 2600 module then, so all the 2600 games went out the door.)

It was a smart move to improve PR and brand loyalty for 5200 owners (with the 5200 being discontinued), especially if offered at tight price (or maybe a rebate with proof of purchase of a 5200).

 

I agree on the vertical module form factor, maybe they could have made it into something more like Curt's XM or such without adding to cost.

 

A lot of good stuff. I enjoyed that. Cool SMB 8-bit demo. My original theme for this post was to see if it could have been possible to make Maria an "XM" of sorts for the 5200... As in still using the 5200 hardware, but upgraded graphics. I figured the 7800 expansion would have been a repackaged 7800, like the 2600 module was.

 

I do recall those Atari pixels looking pretty wide on my TV.. As for the emulators not displaying the proper aspect, I did check the "Maintain Aspect ratio" for all of the shots. The Nintendo emulator was the only one that wouldn't cooperate, which is why I did not include it in the comparisons. Of course it COULD still be off, but that IS the way I remember them, and I spent many TOO LONG nights trying to create the BEST looking "Player" graphics on my XL. I thought Mode 15 was pretty sad too. But I loved it for the day.. Not quite as bad as the Vic 20.. LOL.

 

Again.. Thanks for sharing. That's a lot of good points.

[kool kitty89]

A lot of good stuff. I enjoyed that. Cool SMB 8-bit demo. My original theme for this post was to see if it could have been possible to make Maria an "XM" of sorts for the 5200... As in still using the 5200 hardware, but upgraded graphics. I figured the 7800 expansion would have been a repackaged 7800, like the 2600 module was.

Not so much like the XM, but yes Atari Inc had plans for exactly that: a 7800 in a box to plug into the 5200. (makes most sense from a marketing/PR perspective to bring over 5200 users and look good for the press -and even from a business perspective in as far as getting 5200 users to convert formats -of course, by the time Warner and Tramiel worked things out over the GCC contract, it didn't really matter anymore)

 

There were even some working prototypes, including some mock ups and preproduction models even:

http://www.atariage.com/forums/topic/167048-atari-7800-adaptor-for-atari-5200-at-pax-east/

 

I do recall those Atari pixels looking pretty wide on my TV.. As for the emulators not displaying the proper aspect, I did check the "Maintain Aspect ratio" for all of the shots. The Nintendo emulator was the only one that wouldn't cooperate, which is why I did not include it in the comparisons. Of course it COULD still be off, but that IS the way I remember them, and I spent many TOO LONG nights trying to create the BEST looking "Player" graphics on my XL. I thought Mode 15 was pretty sad too. But I loved it for the day.. Not quite as bad as the Vic 20.. LOL.

"maintain aspect ratio" doesn't necessarily mean it aspect corrects for what a normally calibrated TV will show. (for several emulators, it means square pixels -very few emulators offer realistic NTSC aspect ratio modes, but a few do -and all such modes require antialiasing since you're scaling the pixels)

 

And yes, 3.58 MHz pixels are pretty wide, but not 2:1 aspect ratio wide.

Edited February 10, 2011 by kool kitty89