It's 1990, your designing a game system, what would you do?

[malducci]

If slowdown is caused by the cpu running out of cycles in one frame, then why doesn't the framerate drop with the gameplay speed, instead of having sprites move by half distance per frame at 60fps?

 

You'd have to know your out of cycle time first, if you were going to make a setup dropped frame rate VS just slowing down. Not to mention the serious complexities of doing something like that on these old system.

[Multijointed Monster Maker]

Wouldn't the framerate automatically go down to 30fps, just because it's showing the same frame twice instead of once? I could swear I've seen games that looked 60fps during slowdowns, like they were intentionally put there, or could it just be that I just don't notice the difference between 30fps and 60fps.

[kool kitty89]

The screen has to always refresh at 60/50 Hz of course, and most slowdown I've seen slows the entire screen (to 30/20 fps, sometimes slower) I thik it might be a bit easier to drop framerate for 60 Hz than 50 Hz too, beyond 1/2 at least: 60 divides nicely bu 2,3,4,5,6,10, while 50 only by 2,5,10. (ie more integer factors)

 

Of course, in the case of bitmapped graphics (framebuffer), you get tearing if the proper frameare can't bemaintained without double buffering.

Edited April 4, 2010 by kool kitty89

[Herbarius]

Okay, I've had some thoughts about this.

 

Let's start with some basics: It's 1990 and we're designing a game system, as it is impossible to guess how you would have decided on these things in 1990, without knowledge of the future, let's say we somehow (e.g. through time-travel) are in 1990 with knowledge of (the un-altered timeline's) 2010.

Having established that we quickly identify our first major problem: In 1991, the SNES will hit the US market, and the European market in 1992. So, whatever we will do, even if superior, the SNES will surely beat our product into the ground.

So, let's think about our goals: Achieving the number one position will be impossible. So, our aim will be of getting in as a solid number two.

 

 

So, I'll leave it at this for now. I've already some ideas for my system, but I'll need another day or two to think it out as much as I'd like.

[kool kitty89]

somewhat disappointed that no one has mentioned the konix multi system...that definitely had possibilities (think of it as a 1990's wii wii)

Not really sure on the hardware side, though the use of the 8086 seems a bit weak, but there's not much information on the rest other than the general video modes and RAM configuration.

 

However, that's just from flare's side, the base console hardware, the ridiculous idea of all those expensive peripherals being standard was totally impractical... not sure how floppy discs would have gone over that well either, lading times, durability, piracy issues, etc.

 

It seems like it might have had promise, but the description is very vague, so a system with a "blitter" driving the graphics, and "DSP" driving the sound really doesn't say too much...

 

 

Okay, I've had some thoughts about this.

 

Let's start with some basics: It's 1990 and we're designing a game system, as it is impossible to guess how you would have decided on these things in 1990, without knowledge of the future, let's say we somehow (e.g. through time-travel) are in 1990 with knowledge of (the un-altered timeline's) 2010.

Having established that we quickly identify our first major problem: In 1991, the SNES will hit the US market, and the European market in 1992. So, whatever we will do, even if superior, the SNES will surely beat our product into the ground.

So, let's think about our goals: Achieving the number one position will be impossible. So, our aim will be of getting in as a solid number two.

 

Why so much emphasis on the SNES? The Genesis/MD made up a HUGE portion of the market, in fact there's some evidence that the MD/Genesis outsold the SNES worldwide excluding Japan's sales. (which makes up a huge chunk for the SNES and a rather modest chunk of the MD, hence putting the SNES well ahead worldwide regardless)

Undoubtedly there were several EU countries where the MD dominated the generation (and others where the SNES was more popular -Germany being a key example).

 

Plus, the Genesis would have been established on the market by that point as would the MD in Europe. (and in Europe there wasn't the Nintendo dominance to get past either -compared tot he US where Sega was just starting to crack Nintendo's stronghold)

Edited April 5, 2010 by kool kitty89

[WatchVenusSpa]

I wouldn't concern myself with how powerful the machine is as much as I would concentrate on creating epic games for the system. The system itself would be 16 bit and cartridge based with the next generation being 64 bit, (1995,) and CD based.

 

With the technical stuff out of the way time for the games.

 

Since they were still a small company looking to break out I would sign a deal with id Software to create ports of their popular PC games, (Commander Keen, Dangerous Dave and the soon to be released Wolfenstein 3D,) as well as new creations for the system. The system would come with Commander Keen. Now we would have our flagship for the first few years on the eve of the 'mascot era'.

 

From there I would spend the money to license and port some of Sierra Online's bigger hits like King's Quest and Leisure Suit Larry. (While I'm at it I'd make a deal with the Williams' where their team could make 1-3 new creations for us a year.) By the time the new 64 bit CD system is launched I would bring Al Lowe in full time after Activision fires him upon buying Sierra from Ken and Roberta.

 

Then I would start buying the rights to arcade hits that had never been snapped up by NES and Sega to make sequels out of. (Venutre, Gorf, High Impact Football and Sinistar come readily to mind.) For High Impact I would team up with the NFL to create a game similar to what would become the NFL Blitz franchise. This is before EA bought the exclusive rights to all things NFL and there was a lot of money to be made out of a license like this.

 

From there I would focus heavily on marketing through contests and launch promotions. The Big N and Sega were pulling back on promotions that included the average fan so why not embrace gamers and get them involved? I'd lean towards a Sword Quest type of a promotion or something similar to the tournament in The Wizard to fuel interest in our newer games.

 

That's what I've got so far.

[kool kitty89]

I wouldn't concern myself with how powerful the machine is as much as I would concentrate on creating epic games for the system. The system itself would be 16 bit and cartridge based with the next generation being 64 bit, (1995,) and CD based.

It's a combination of cost to performance ratio, ease of programming (which depends on both hardware and the SDK tools), good marketing, and getting sufficient interest from 3rd party developers and sporting some 1st party published exclusives.

 

"bitness" only really matters in a marketing sense, and usually an overemphasis on the CPU architecture is made... You could have a perfectly good system by 1990s standards based around an 8-bit CPU (65C02, or perhaps Z80 derivative), though the ignorant media would porbably be more positive toward a weak CPU like a 8086 than a fast 65C02 or Z80 because it's "16-bit" . Plus, only a couple consoles used 64-bit CPUs (the argument against the Jag not being 16-bit and the N64 being so); prior to current gen, only the N64 and PS2 had 64-bit CPUs (internally, 64-bit ALU) and in some cases those CPUs were effectively used only for the 32-bit operations. (like N64)

 

It's really about how the system is organized, you could have a very poorly optimized system where an 8-bit CPU would drag it down considerably, and others where some 16-bit CPUs would be detrimental tot he architecture. (of course, any software rendering is CPU dependent)

 

IMO, it would simply take some good marketing and possibly some more educated marketing staff (with an actual vague understanding of the real strengths of the hardware) to get past such silly stigma.

 

 

Since they were still a small company looking to break out I would sign a deal with id Software to create ports of their popular PC games, (Commander Keen, Dangerous Dave and the soon to be released Wolfenstein 3D,) as well as new creations for the system. The system would come with Commander Keen. Now we would have our flagship for the first few years on the eve of the 'mascot era'.

Well, those are 3rd party titles, so unless you get exclusive contracts for home console ports, they will likely end up on other platforms later too.

 

From there I would spend the money to license and port some of Sierra Online's bigger hits like King's Quest and Leisure Suit Larry. (While I'm at it I'd make a deal with the Williams' where their team could make 1-3 new creations for us a year.) By the time the new 64 bit CD system is launched I would bring Al Lowe in full time after Activision fires him upon buying Sierra from Ken and Roberta.

Those aren't the typical genres that go over well on home consoles, especially at the time... Secret of Monkey Island was rather unpopular on the Sega CD for example. (granted it was an older game at the time)

 

Then I would start buying the rights to arcade hits that had never been snapped up by NES and Sega to make sequels out of. (Venutre, Gorf, High Impact Football and Sinistar come readily to mind.) For High Impact I would team up with the NFL to create a game similar to what would become the NFL Blitz franchise. This is before EA bought the exclusive rights to all things NFL and there was a lot of money to be made out of a license like this.

Sinistar never got "snapped up" by anyone, it got released in William's arcade hits for pretty much every console on the market in the mid '90s as well as being re-released in the Midway hits collection for several later consoles. Not sure about Gorf either, or any such sequels of those 2...

 

From there I would focus heavily on marketing through contests and launch promotions. The Big N and Sega were pulling back on promotions that included the average fan so why not embrace gamers and get them involved?

What do you mean by "pulling back" on the "average fan"??? DO you mean trying to appeal to a broader audience? That's what you'd be doing if you started pushing PC releases as major titles. Plus, I don't even see Sega doing that at the time, in fact, they seemed to be catering more to that than previously with the mediocre SMS marketing.

 

I'd lean towards a Sword Quest type of a promotion or something similar to the tournament in The Wizard to fuel interest in our newer games.

Perhaps a far more cost effective version... Sord Quest was extremely extravagant...

[WatchVenusSpa]

Well, those are 3rd party titles, so unless you get exclusive contracts for home console ports, they will likely end up on other platforms later too.

 

In 1990 the other big name console companies wouldn't touch the John's with a 10 foot pole so I think if one negotiated a percentage on par with their Apogee deal, one could have the best gaming team in the world making exclusives for your console.

 

Those aren't the typical genres that go over well on home consoles, especially at the time... Secret of Monkey Island was rather unpopular on the Sega CD for example. (granted it was an older game at the time)

 

Well to be fair it WAS the Sega CD, one of the bigger busts in gaming history. Growing the market to appeal to PC gamers is definitely something that X-Box and PlayStation did to overtake Nintendo in the late 90's.

 

Sinistar never got "snapped up" by anyone, it got released in William's arcade hits for pretty much every console on the market in the mid '90s as well as being re-released in the Midway hits collection for several later consoles. Not sure about Gorf either, or any such sequels of those 2...

 

That may be true but not in 1990.

 

What do you mean by "pulling back" on the "average fan"??? DO you mean trying to appeal to a broader audience? That's what you'd be doing if you started pushing PC releases as major titles. Plus, I don't even see Sega doing that at the time, in fact, they seemed to be catering more to that than previously with the mediocre SMS marketing.

 

I wouldn't just be pushing PC games as main titles, I'd be contracting their programmers to create some kickass games specifically for my console. As for pulling back, how many epic competitions existed for NES or Genesis gamers? I'd want to do things that were accessable for non "hardcore" gamers but also skill based games as well.

 

Perhaps a far more cost effective version... Sord Quest was extremely extravagant...

 

I would go big but not $150,000 big. What I would go for in the long run is an annual event to create a gamers equivilant of the Super Bowl while advertising my games and console all the way.

[kool kitty89]

Well, those are 3rd party titles, so unless you get exclusive contracts for home console ports, they will likely end up on other platforms later too.

 

In 1990 the other big name console companies wouldn't touch the John's with a 10 foot pole so I think if one negotiated a percentage on par with their Apogee deal, one could have the best gaming team in the world making exclusives for your console.

 

What I meant was, if they weren't contracted with exclusive licensing, any popularity of such titles on you console would spur ports to contemporary big name consoles.

 

Well to be fair it WAS the Sega CD, one of the bigger busts in gaming history. Growing the market to appeal to PC gamers is definitely something that X-Box and PlayStation did to overtake Nintendo in the late 90's.

I'd contest that about the Sega CD... I doubt it would be looked ont he way it is had Sega not followed up with tons of problems and major foibles...

Several other PC ports did well though, including adventure games, albeit those often had that streaming video gimmick of most CD based games at the time. (Dune, Rebel Assault, etc) So I guess I may have provided my own counterexample...

 

I will say I disagree with the latter though on Sony and MS. Nintendo was the first to push big with console FPSs (originally a PC genre -and prior to which all console releases had been PC ports), and several publishers (like Lucas Arts) were specifically releasing multiplatform PC/Nintendo titles. I think the entire industry, including Nintendo was becoming more and more blurred with PC and console gaming starting particularly with the 5th generation consoles. However, this was after a major genre popularity shift, and the rise of 3D.

 

Though this applies more prominently to the IBM PC/compatibles than some other popular computers at the time, especially Europe. (where consoles and contemporary computers were much more blurred in general, from the 80s with 8-bit computers and consoles, then the Amiga and ST followed by the 4th generation game consoles, with PCs finally popularizing on the verge of the 5th console generation)

 

Sinistar never got "snapped up" by anyone, it got released in William's arcade hits for pretty much every console on the market in the mid '90s as well as being re-released in the Midway hits collection for several later consoles. Not sure about Gorf either, or any such sequels of those 2...

 

That may be true but not in 1990.

How so? Where were Sinistar or Gorf in 1990?

 

As for pulling back, how many epic competitions existed for NES or Genesis gamers? I'd want to do things that were accessable for non "hardcore" gamers but also skill based games as well.

Not sure about Sega, but I know Nintendo sponsored a lot of public gaming events. The NWC being the most obvious, Star Fox: Super Weekend Competition also comes to mind off the top of my head.

[WatchVenusSpa]

In 1990 Romero and Carmack were borderline blacklisted and it wasn't until 1993-1994 that the mainstream gaming companies wanted anything to do with them. (Post Wolfenstein.) They could have easily been signed to an exclusive deal to make 3rd party games for a console if someone actually made them an offer. (Remember, this was when they were living in a housing project complex in Madison, Wisconsin living from Apogee paycheck to paycheck.)

 

As for PC gaming, the landscape was about to change drastically and I would want to capitalize on it right away.

 

As for Gorf, Venture and Sinistar, NOBODY was making sequels to these games in 1990 if ever. Why not buy up the rights to these games and create new sequels and spinoffs for them? It worked for the Johns with Wolfenstein. (Essentially remaking a dead PC game and reinventing it completely with new technology.)

[kool kitty89]

Getting interest from Origin might be a good idea too. They had a couple titles that made it to consoles early on (some of the ultima series), but a lot of stuff that wasn't, or went several years without ports. If the blitter in the Flare 1 design was suitable, that could have facilitated a good port of wing Commander.

[Herbarius]

Getting interest from Origin might be a good idea too. They had a couple titles that made it to consoles early on (some of the ultima series), but a lot of stuff that wasn't, or went several years without ports. If the blitter in the Flare 1 design was suitable, that could have facilitated a good port of wing Commander.

I thought about the same thing. Also, those games which had been ported, should have GOOD ports. Look at the SNES port of Wing Commander, thats crappy. Gameplay mechanics altered because the console couldn't handle the battles in their original form (or maybe the programmers were just lazy). No, we have to have a QUALITY PORT of Wing Commander, even with (optional!) support for analog joysticks (sold seperately as accessory - or get the adapter for PC joysticks). And if only because Wing Commander is my favourite game (of that year).

 

Generally, I'd say look out for quality titles on the PC market that others are reluctant to port or made crappy ports of it and ensure quality port. Try to maintain good relations with PC game companies, maybe get some exclusives that way.

 

 

Did I already tell you that my console will come with an internal 3.5" disk drive? (However it uses cartridges for the games, the disks are for save games and user generated content (like maps/levels)).

Edited April 7, 2010 by Herbarius

[kool kitty89]

As for Gorf, Venture and Sinistar, NOBODY was making sequels to these games in 1990 if ever. Why not buy up the rights to these games and create new sequels and spinoffs for them? It worked for the Johns with Wolfenstein. (Essentially remaking a dead PC game and reinventing it completely with new technology.)

Not only that, but, unlike Gorf and Venture, Sinistar never got a home port in its heyday even. (some prototypes I believe)

 

 

I thought about the same thing. Also, those games which had been ported, should have GOOD ports. Look at the SNES port of Wing Commander, thats crappy. Gameplay mechanics altered because the console couldn't handle the battles in their original form (or maybe the programmers were just lazy). No, we have to have a QUALITY PORT of Wing Commander, even with (optional!) support for analog joysticks (sold seperately as accessory - or get the adapter for PC joysticks). And if only because Wing Commander is my favourite game (of that year).

I have the SNES port, and while it's definitely cut down, it's still reasonably playable. The Amiga port seems to be oddly poor, like it was a very basic port, taking little to no advantage of the blitter.

The Sega CD port seems to have been expertly handled, making use of the graphics coprocessor (basically a blitter with emphesis on scaling/rotation capabilities), but for some weird reason they replaced the John Williams music with some weaker alternate stuff. It's red book audio, but of arrangements not on par with the WC originals. The later win9x re-release has an excellent digital audio orchestral recording though. (hell, the Sega CD would have simply used recordings from the MT-32, which seems like what Lucas Arts may have done for the CD releases of Secret of Monkey Island) Though on can rip the ISO and replace those tracks for the Sega CD now too.

 

Do note that the later 3DO, Saturn, and PSX Wing Commander game releases did support the analog flight sticks on those consoles.

 

Also, looking at Wolf3D on the SNES from a technical point of view, it's quite an accomplishment. Reasonably large screen and good framerate using only software rendering, not to mention working with only 128 kB of RAM; had it not been for the forced censorship, I'd imagine that port would be looked upon in a much better light. Also note that the game is actually running in mode 7, using the plane scaled to 2x, a 112x96 screen scaled to 224x196, with sprites overlaid for the status display and player's hand/weapon. (it's a 256 indexed color mode using 8-bit packed pixels -all other modes being planar) A neat trick they used there and kind of makes me wonder why it wasn't used for anything else.

 

 

The Archimedes derived console that Crazyace suggested would be rather well suited for a lot of the PC ports, all software rendering, like PCs, also supporting 256 color modes and similar resolutions as VGA (albeit some suggested games were EGA), then there's the differing CPU architecture, of course. (and sound hardware)

 

On the topic of potential PC developers going untapped would have to include Sierra as well. While a lot of the Adventure games may not have meshed particularly well on consoles, namely any requiring typed text inputs. (point an dc lick stuff would be OK, though a tad tedious without a mouse) They had a variety of other genres though in spite of the heavy adventure game emphasis, as well as some licensed conversions like Thexder and Silpheed. (those would both have been nice on home consoles, the former on the Famicom in Japan only, and the latter getting a remake half a decade later on the Sega CD, and another on the PS2 -the latter not holding up as well IMO)

 

I think Argonaut Software might have been an interesting one to consider as well, though perhaps some of those games were too involving compared to a lot of console games. (still, ports of Starglider or Starglider 2 would have been neat -more likely the latter unless the former was redone with filled polygons) Though given that a lot of their games wer published by different 3rd parties might have complicated things. (nintendo wouldn't have been an issue until a little later)

 

 

Did I already tell you that my console will come with an internal 3.5" disk drive? (However it uses cartridges for the games, the disks are for save games and user generated content (like maps/levels)).

Not sure if that's a really good idea, disk drives would add to cost significantly, plus make a liability for failure and recalls. Another choice to avoid the high cost of SRAM saves might be EEPROM, either on cart or in memory cards (possibly with a built-in reserve as well).

I'm not sure what kind of interfacing requirements there were at the time, ie if writes could be controlled by the CPU or if a dedicated controller would be required.

I believe several Genesis/MD titles used EEPROM as well as the 32x, and I'm pretty sure that was the only type of save memory used for the Jaguar, so it seems liekly that EEPROM would have been fairly practical for the time. (onboard, on-card, and/or as memory cards)

 

The obvious disadvantage over SRAM is long save times, though that would be a problem with disks as well. (and disks are less durable as well)

[gdement]

As far as going after licensees, I agree with the above points about PC games. In fact I think that should be the core of any 1990 strategy.

Nintendo had the major Japanese developers locked up, but most western developers weren't much involved with consoles ever since the 83 crash. It was Nintendo's key vulnerability.

In 1990, the best strategy would be to attract western computer game developers in general. The hardware should be tailored around that strategy. It was still early days for the Genesis and you could conceivably beat SOA at western licensing.

Even if you don't get exclusive deals (and you probably wouldn't), you can still design a system that western devs are more comfortable with than the Genesis or SNES.

 

CPU would be a 68K.

Sound: What was Commodore doing at the time? I wonder if you could contract development of a "double SID". In the meantime, maybe the first-gen consoles could ship with dual SIDs from Commodore overstock (if there was any).

Otherwise, some Yamaha thing would suffice, hopefully similar to the AdLib/Sound Blaster cards.

Graphics: I dunno.

Controllers: besides gamepads, also consider a keyboard attachment. Maybe include a cheap one with the console.

If not a full keyboard, then some substitute device that has plenty of buttons, and can take templates for different games. Hopefully it's something rugged enough to lay it on the floor in front of your TV, where it gets stepped on.

Also make a mouse. Gamepads are preferred for games that can use them, but many computer games need more than that.

 

 

Support from the west is something that I believe almost unwittingly made the Genesis so successful in our hemisphere. Trip Hawkins mentioned that the 68k processor is what got him interested in the Genesis, because his company was filled with people who were already experienced with that chip. Same would be true for pretty much any computer game company.

"EOA" and Accolade both made games for the system in it's early days without Sega's approval. Even when they signed an agreement, EA used their ability to go rogue as leverage to get better terms.

Despite their rough relationship, EA probably sold millions of Genesis consoles in North America, and hopefully SOA began to appreciate this fact.

Through the course of the 16-bit era, western developers became increasingly prominent on both Genesis and SNES, but I believe their presence overall served to level the playing field for Sega. They were a ripe resource that Nintendo didn't control.

Edited April 7, 2010 by gdement

[carmel_andrews]

KK, in regards to the Konix system, i was referring to the updated one (using either a 2 or 386) that was going to be put out by a company called MSU (or was it MXU) who bought the rights after konix went into administration

[kool kitty89]

In 1990, the best strategy would be to attract western computer game developers in general. The hardware should be tailored around that strategy. It was still early days for the Genesis and you could conceivably beat SOA at western licensing.

Even if you don't get exclusive deals (and you probably wouldn't), you can still design a system that western devs are more comfortable with than the Genesis or SNES.

Such developers would include Europe as well as North America, of course. Also, that Nintendo didn't have a hold on Europe like Japan and North America (divided with Sega and home computers). Though, also, that the architectures of many such home computers were more akin to contemporary consoles in some aspects than PCs are. (the atari ST would obviously correspond to a PC's software rendering though) The C64 in EU was probably the most architecturally similar to contemporary consoles; the Amiga did have some hardware sprites and an emphasis on 2D acceleration, but with a blitter rather than dedicated tile based display hardware and heavy emphasis on sprites. (more flexible)

 

CPU would be a 68K.

Sound: What was Commodore doing at the time? I wonder if you could contract development of a "double SID". In the meantime, maybe the first-gen consoles could ship with dual SIDs from Commodore overstock (if there was any).

Otherwise, some Yamaha thing would suffice, hopefully similar to the AdLib/Sound Blaster cards.

Graphics: I dunno.

Well, you could go with a powerful CPU and all software rendering and a simple bitmap display (as with the crazyace's suggestion), or a modest CPU with powerful coprocessors (blitter, etc). As for sound, an off the shelf yamaha chip might be fine, though I'd tend to favor ones other than the OPL2 of the Adlib; an OPL3 would have been great if available yet, otherwise a YM2612 or pair of such might be good (genesis used it as well as several arcade machines and the FM towns), with that you've also got an integrated DAC and ability to disable channel 6 and use 8-bit linear DAC directly. (for PCM playback) Or a custom sound chip, perhaps of a DMA driven type like the Amiga used. (again, crazyace's suggestion applies here too)

One thing about the archimedes derivative though, would be that Acorn was not openly licencing their hardware at the time (I don't think even the ARM chips had yet been licensed to 3rd parties), so that would hinge on managing a partnership with Acorn. (which might be possible, especially if any such licensed product was restricted to the video game market -not threatening direct competition)

 

Controllers: besides gamepads, also consider a keyboard attachment. Maybe include a cheap one with the console.

If not a full keyboard, then some substitute device that has plenty of buttons, and can take templates for different games. Hopefully it's something rugged enough to lay it on the floor in front of your TV, where it gets stepped on.

Also make a mouse. Gamepads are preferred for games that can use them, but many computer games need more than that.

I don't like the sound of that... Most attempts at such tended to fail or at least detract from the console. I'd stay away from text heavy games and such. Some more complex games with many commands (like X wing or wing commander) could probably work well on a many button gamepad, with one with a keypad (ie Jaguar), or more preferably a large number of action buttons, with button combos used to further expand such in-game commands. (like the Jaguar Pro controller with keypad minimized or removed or saturn -I'm partial to Gravis's xterminator, Microsoft's Sidewinder gamepad would be a somewhat similar earlier example though) A mouse peripheral would be nice for point and click games (and indeed, the MD, and SNES had mice), perhaps an analog joystick as well.

 

After playing newer incarnations of such games (like X-wing Alliance) with fully mappable buttons on modern gamepads or joyticks, I much prefer that to working with a keyboard and simple 2-button joystick (especially with throttle support). Again, the exterminator is kind of a jack of all trades for this, and IMO it's a real shame there's no modern counterpart since Gravis disappeared. (though there are some uncommon USB incarnations of the xterminator)

 

 

Support from the west is something that I believe almost unwittingly made the Genesis so successful in our hemisphere. Trip Hawkins mentioned that the 68k processor is what got him interested in the Genesis, because his company was filled with people who were already experienced with that chip. Same would be true for pretty much any computer game company.

"EOA" and Accolade both made games for the system in it's early days without Sega's approval. Even when they signed an agreement, EA used their ability to go rogue as leverage to get better terms.

Despite their rough relationship, EA probably sold millions of Genesis consoles in North America, and hopefully SOA began to appreciate this fact.

Through the course of the 16-bit era, western developers became increasingly prominent on both Genesis and SNES, but I believe their presence overall served to level the playing field for Sega. They were a ripe resource that Nintendo didn't control.

Undoubtedly that's a factor, but good marketing most definitely was a major factor as well, plus strong 1st party software support (both Japanese and SOA's STI), with a greater emphasis on some genres, like sports titles. (which EA did cater to as well)

That said, it's also pretty clear that the MD/Genesis relied a lot more on North American and European developers than Nintendo was, at least in the early 90s.

 

As for the unlicensed thing, I believe EA did get official licensing on the Genesis before releasing any games, though for some reason Populous got released without forward compatibility with Sega's later introduced TMSS lockout. Possibly due to that game being developeded prior to EA's licensing. (though not released until after Sega's deal was made) I think Michael Katz mentioned that SoJ was pretty pissed off at EA reverse engineering the hardware and making their own development tools, but Katz took the "if you can't bet 'em joun 'em" proactive approach with EA instead. (and EA ended up developing several Sega published titles early on, like Joe Montana Football)

 

KK, in regards to the Konix system, i was referring to the updated one (using either a 2 or 386) that was going to be put out by a company called MSU (or was it MXU) who bought the rights after konix went into administration

Hmm, you mean this:

http://www.konixmultisystem.co.uk/index.php?id=msu

 

That sounds like a lot later than 1990.

 

The original Flare 1 design sounds pretty impressive for the time (powerful blitter, math coprocessor, color depth, etc), but who's to say how that would have transitioned to the market. Regardless, the whole thing got muddled with Konix with the overemphesis on peripherals. (granted, that's what the company did, sort of like Gravis, Logitech, etc taking interest in a game console) It does sound like it was fairly oriented towards polygon rendering as well, most descriptions seem a bit vague though.

Interesting that the original design used a Z80 CPU supplemented by a number of custom chips/coprocessors. (apparently an additional ALU, blitter, VDC, and DSP for audio) It appears the DSP was intended both for driving sound and as a math coprocessor for 3D.

 

Sounds like it might have been a lot more realistic than the Panther though. (requiring 32-bit SRAM, and sprite oriented) The Jaguar seems like it went back to some of the Flare 1 design concept merged with the Panther with new hardware.

Edited April 7, 2010 by kool kitty89

[Multijointed Monster Maker]

With the exception of 32-bit RISC cpus, I'd say the Hitachi 6309 had the best performance/cost ratio at the time. I know a lot of people would argue the 68000, but I disagree.

[potatohead]

Deffo 6309, even the Moto 6809 had good characteristics. Defender was written on a 6809, BTW.

 

Funny, I was reading through this thread thinking of the 63 and 6809 CPUs as good targets for a game console.

 

One really great thing about these chips is that custom hardware doesn't have to interrupt the CPU. That's a huge performance gain. The other thing about them, compared to the 68000 series, is having short instructions. Overall instruction density is higher on the 6309, and can be faster for "read it", "decide on it" kinds of structures.

 

IMHO, the 64K address space is a bit of a limit, but the powerful addressing modes make relocatable and re-entrant code very doable, allowing for some creative use of memory banking.

 

If it were me, I would have put video and audio signals on the cart, along with enough address lines, and the clock, so that additional custom hardware could be used with that native feel. More expensive carts are a better sell than some expansion that not everybody has, and economies of scale over time improve the use of them, while improving the base console.

 

-custom sound chip, 4 channel stereo, 8 bit samples on DMA access (remember this can be done with no CPU interrupt on those CPUs), plus some waveform / frequency / distortion options.

 

-custom video chip, 16 color tile graphics, scrolling, base memory pointers, 16 color sprites, 10 bit color specification, 64 color palette registers(16 for screen, 2 banks of 16 for sprites), screen interrupts, display list, sprite list (point to sprite position list).

 

IMHO, at that time, on screen action trumped colors overall, so I would go for that, and having 64 colors on the screen is a good amount of color, and it's broken down into usable chunks.

 

If shift register type devices were not all that expensive, I would use those for controllers, being sure to have analog channels.

[avanteproject]

CPU- Custom built based on the Motorola 6800 range.

 

24bit data bus

 

GPU- Ability to produce 150 64x64 sprites on screen, 16 bit colour, 10,000 flat shaded polygons on screen, 320x240 max resolution.

 

Sound processor capable of 8 channel sound, digitised voice reproduction, synthesiser.

 

2MB RAM

 

RF, Composite, S-Video, RGB output.

 

Two controller inputs.

 

8MB RAM cartridge storage.

 

Whether it'll be sold at a loss, I don't know. But if I have the money I'll throw in blast processing too. lol

[kool kitty89]

Deffo 6309, even the Moto 6809 had good characteristics. Defender was written on a 6809, BTW.

All of williams classic arcade boards were 6809 based and used a 6808 MCU driving a DAC for audio. http://www.system16.com/hardware.php?id=598

 

Funny, I was reading through this thread thinking of the 63 and 6809 CPUs as good targets for a game console.

 

One really great thing about these chips is that custom hardware doesn't have to interrupt the CPU. That's a huge performance gain. The other thing about them, compared to the 68000 series, is having short instructions. Overall instruction density is higher on the 6309, and can be faster for "read it", "decide on it" kinds of structures.

But it's pretty bus heavy isn't it? (like the 650x) For a shared bus architecture, the 68k works pretty well due to the ability to interleave accesses and take advantage of the bus with (assuming a 16-bit chipset), hence what they did with the Amiga (and ST) for DMA. (at least until you get to FPM DRAM, where interleaved random accesses loose much of their value)

 

Now, if you designed a system for the 6809, you'd probably design it to have a dedicated bank of RAM for the cpu and separate for custom chips.

It all depends on the design philosophy.

 

IMHO, the 64K address space is a bit of a limit, but the powerful addressing modes make relocatable and re-entrant code very doable, allowing for some creative use of memory banking.

There's always bank switching, either implemented by external logic (including an MMU) or embedded in a custom package with the CPU. (as NEC did with their 65C02 derivative)

 

If it were me, I would have put video and audio signals on the cart, along with enough address lines, and the clock, so that additional custom hardware could be used with that native feel. More expensive carts are a better sell than some expansion that not everybody has, and economies of scale over time improve the use of them, while improving the base console.

Yep, that's what contemporaries were doing, Sega, Nintendo, NEC (the latter on the expansion port and the most extensive of the 3, unfortunately never taken advantage of -Supergrafx could have been an add-on), and ironically Sega made their expansion port too limited. (had the CD unit been cartridge mounted, that wouldn't have mattered much though) The famicom did that too to a large extent, though the NES lacked the audio expansion. (rerouted to the bottom expansion port iirc)

 

If you used separate audio/video buses, you could go the route Nintendo did with the famicom, adding connections for multiple buses and enough addressing to facilitate quite a bit of expansion. (numerous mappers, etc) The downside to this is greatly increased pin count and thus cost of the cartridge PCB. (getting worse for newer consoles with wider buses and address space -you end up in Neo Geo territory eventually) With a shared buss system that doesn't matter of course, as a single cart bus would allow direct access by either the CPU or coprocessors. (as with the 7800 and I think A8/5200)

 

-custom sound chip, 4 channel stereo, 8 bit samples on DMA access (remember this can be done with no CPU interrupt on those CPUs), plus some waveform / frequency / distortion options.

So a bit like the Amiga's sound then -don't forget to include volume control: independent stereo control would be better to rather than the fixed stereo of the amiga. (allow any of the 4 voices to output to either L/ or both as the Genesis and SNES did)

 

-custom video chip, 16 color tile graphics, scrolling, base memory pointers, 16 color sprites, 10 bit color specification, 64 color palette registers(16 for screen, 2 banks of 16 for sprites), screen interrupts, display list, sprite list (point to sprite position list).

That's pretty limited. Genesis and PC engine used 9-bit RGB palettes and the limitation shows compared to SNES, 6-bit RGB is way too limited... 12-bit minimum for 1990 I'd say. number of subpalettes (assuming 16-color tiles) is also very important and a huge disadvantage for the Genesis to SNES and PC Engine (the latter even showing its advantage over SNES in better shading if more limited master palette to SNES).

If you went with sprite/tile based graphics with multiple scrolling planes and sprites, 16-color tiles/sprites would be good, but if you opted for a blitter instead, you'd probably want 256 colors (probably with 16-color mode too), or more variable if you opted for planar bitmap opposed to chunky. (then you could get 32 or 64 colors as with the Amiga -very competitive with a much larger number of indexes of 16-color palettes due to redundant colors)

Of course, you could have a hybrid set-up of a heavily sprite+multi-layer scrolling tile set-up as well as a blitter/coprocessor specifically designed to work in tiled graphics. (as Sega did with the ASIC in the Sega CD -albeit the Genesis VDP wasn't well suited to that as things were, but that doesn't mean it's a bad idea in general)

 

IMHO, at that time, on screen action trumped colors overall, so I would go for that, and having 64 colors on the screen is a good amount of color, and it's broken down into usable chunks.

Wait, 64 idexed colors, or a 64-color master palette. The Genesis has 64 indexed colors but a 9-bit RGB palette, though both of those proved quite limiting compared to the SNES (or PC Engine) and took a lot more work to optimize for. If you were going the 16-color tile route, I'd say 12-bit RGB and 8 16-index palettes minimum. However, still going with tiles, but opting for planar graphics (which SNES and PCE did use -as well as NES/SMS) you could use a more flexible set-up, like allow for up to 6 bitplanes and thus 64 indexed colors slectable per pixel (and allowing you to use fewer bitplanes in case where lower memory use is desirable). 4x 4-bit indexed palettes is WAY more limiting that 64 indexed colors or even 32 indexed colors much of the time.

 

It really depends on what you want to do, but if you're going for the PC market, I'd definitely opt for a blitter, probably support a 256 color chunky mode due to VGA starting to become popular, etc. With an Amiga-like set-up, you've got the ability to match a lot fo VGA stuff fairly well with halfbrite, but the set-up it a bit different and requires more optimization and CPU heavy PC games really need to cater to the blitter. (Wing commander was very poorly converted in that respect)

 

Further, there's the non-blitter all CPU option crazyace proposed, using a consolized Archimedes basically. (in that specific case, you'd need to arrange a deal with Acorn as they weren'r openly licencing yet)

 

 

Of course, a consolized Amiga directly would fit the bill pretty well and be fairly competitive with the MD and SNES in spite of its age. It already got a lot of conversions of PC titles, so that's already in a good position (other than some sloppy ports), all it really needed was a cost optimized form factor (with minimal RAM and probably no OS), cartidge based, enhanced controller (at least match Sega's 3 buttons plus start, preferably 4 buttons), alterations made to games that required it (obviously games requiring more RAM would need to be modified, and games could cater to running from ROM mostly, supporting more buttons, etc)

Mouse support wouldn't be a bad idea either.

 

 

If shift register type devices were not all that expensive, I would use those for controllers, being sure to have analog channels.

Why analog? Analog control wasn't very important that generation, and even for flying games, digital controll works pretty well. (I was recently trying Wing commander and X-wing using a gamepad and it worked surprisingly well other than being awkward to use the keyboard -due to lack of button mapping)

 

There's a variety of different logic choices for controllers, nintendo had been using the same system with NES and SNES, while Sega used direct inputs on the SMS (parallel I/O) and utilized a select line to extend that for the Genesis.

 

Using a straight parallel interface with a 9-pin connector you could get 8 inputs plus gnd, though if you wanted provisions for a mouse and possibly some other peripherals, you'd need +5V at least, so 7 inputs: so you could use the conventional pinout for directions and fire then add 2 more buttons for a total of 3. (and stanrt/select or other non action buttons could be read as unused combinations of the directional inputs, like up+dn or L+R, L+R+dn, etc)

 

 

 

24bit data bus

Why???

 

10,000 flat shaded polygons on screen

That's a rather worthless spec, 10,000 polygons per second might be reasonable. 10,000 polygons on-screen could be done with very slow hardware to render, no 10,000 polygons (triangles) per frame and any kind of decent framerate is rather impractical (that's PS1 territory and under 30 FPS). Peak of ~30,000 polygons per second isn't crazy for the time either, possible, that's what flare had planned for the multisystem. (with DSP maxed out, so no sound)

Adding a coprocessor to support such isn't a bad idea, that or building the whole system around software rendering with a fast CPU as crazyace suggested.

 

2MB RAM

Too much for 1990. 1 MB maybe if you used the cheap/common DRAM types of the time (512 kB is much more likely though, for cheap DRAM)

 

RF, Composite, S-Video, RGB output.

Sounds good, probably use a standard DE-15 port for that (with stereo sound).

 

8MB RAM cartridge storage.

You mean ROM, right?

 

Whether it'll be sold at a loss, I don't know. But if I have the money I'll throw in blast processing too. lol

No company prior to Sony launching the PlayStation in the US (and after the '83 crash) ever sold at a loss AFIK, perhaps when dumping surplus/overstock, but otherwise it was sold for slim profit most of the time, perhaps at cost in some cases. (and sony took a much bigger hit than planned due to stagnating RAM prices)

Edited April 17, 2010 by kool kitty89

[malducci]

Yeah, 1megabyte of ram doesn't do much for consoles of that date. Unless you mean 1megabit (128k), which is what the snes had.

 

And 10,000 polygons per screen? Screen as in a single frame instance of movement? 60fps, that's 600,000 polygons a second. Hell, even at 30fps that 300k polygons a second. Pretty unrealistic. Unless you mean 10k polygons per second. But that's extremely low (333 polygons per frame as 30fps).

 

I like the 6309 idea. That is one badass CPU. Of course, it needs to be matched by clocked speed of around 7-8mhz and back up by at least 8k logical segments for external memory mapping. DMA doesn't have to happen at interleaved memory access. If the video system is "port" based like SNES and Genesis, then a fast DMA that stalls the CPU until finished... is just fine (it's what happens on both the SNES and Genesis).

Edited April 17, 2010 by malducci

[kool kitty89]

Yeah, 1megabyte of ram doesn't do much for consoles of that date. Unless you mean 1megabit (128k), which is what the snes had.

It depends on what you're aming at. With a framebuffer based system (blitter or CPU rendering), you need a fair chunk of RAM for that, and you also tend to need more RAM if you want to facilitate ports of PC games. (Wolf 3D must have been hacked to hell to fit on the SNES -and no wonder it didn't make it to the Genesis -surprising there's no Sega CD release though)

DRAM is relatively inexpensive though compared to what was in a lot of consoles (Genesis is a mix of SRAM, PSRAM, and VRAM, SNES has a lot of SRAM and a chunk of DRAM, and I think TG-16 is all SRAM), 512 kB in 1990 is pretty reasonable. Here's an interesting reference on the topic: http://phe.rockefeller.edu/LogletLab/DRAM/dram.htm

 

Again, it depends on what you want to do: I personally think moving away from sprite+tile based systems and going for a more flexible (blitter or CPU) approach is a good direction to move in.

 

And 10,000 polygons per screen? Screen as in a single frame instance of movement? 60fps, that's 600,000 polygons a second. Hell, even at 30fps that 300k polygons a second. Pretty unrealistic. Unless you mean 10k polygons per second. But that's extremely low (333 polygons per frame as 30fps).

10,000 polygons per second isn't bad at all for the early 90s especially if that's practical performance and not peak. That's equal or better than what the Super FX coprocessor was managing on the SNES or Sega's SVP chip (custom packaged Samcung DSP) used in virtua racing. Flare projected the DSP in the Flare 1 (implemented in the 1989 Konix Multisystem) could manage around 33,333 polygons per second peak performance, but that was without sound as the DSP also drove audio. (so much more than practical in a game)

If you included a DSP coprocessor for such (for 3D math, helping with rendering, etc -ray casting, affine rendering, scaling, rotation, polygon rasterization etc) that could be practical. (it depends on how powerful the DSP is -I do wonder if Nintendo's DSP would have been useful for any such had Argonaut not opted for the Super FX)

Of course, a fast CPU is also useful for such, again going back to the Archimedes derivative.

 

And 30 FPS is far higher than such games ran... usually it was more like 12-20 FPS (Virtua Racing runds at a rock steady 15 FPS and Star Fox runs at 15 FPS with a fair bit of slowdown at point -below 10 FPS sometimes)

The 3D games running on average 1990 PCs, Sega Genesis, Amiga, or Atari ST were much slower than that generally. (like LHX, Hard Drivin', or F-15 Strike Eagle II)

So with 10,000 polys per second at 15 FPS that's 666 polygons per frame, not too shabby for the time.

 

The Flare 1/multisystem itsself seems like it could have been a pretty solid design for the time, what Konix did with it was unfortunate. (too bad Flare hadn't ended up at Atari earlier; though Amstrad could also have used it... much better than the outdated tech in the GX-1000)

I'm not sure a 6309/6809 wouldn't have fit in that design either, granted the original Z80 is a pretty different architecture (they did switch to an 808x, but that's somewhat closer to the Z80 than the 680x is).

 

I like the 6309 idea. That is one badass CPU. Of course, it needs to be matched by clocked speed of around 7-8mhz and back up by at least 8k logical segments for external memory mapping. DMA doesn't have to happen at interleaved memory access. If the video system is "port" based like SNES and Genesis, then a fast DMA that stalls the CPU until finished... is just fine (it's what happens on both the SNES and Genesis).

Don't forget you need fast enough RAM to manage speeds like that (and ROM as well), the PC Engine had to use SRAM for main memory to manage that speed in 1988 (thus limited to 8 kB), the SNES limited its DRAM bus to 2.68 MHz due to that as well (and ROM to 2.68 or 3.58 MHz), though I think in the SNES's case the 65816 requires memory to be 2x as fast as the CPU due to the way the multiplexed bus works, so a plain 6502/C02/6800 should have been able to be 2x that fast with similar DRAM. 68k based systems don't have that problem as the 68ks memory accesses are much slower, Z80 is a bit like that too (compared to 650x and 680x accessing once per cycle).

 

You could do what the SNES did and allow variable speeds, with slower DRAM and faster ROM (PC Engine used very fast ROM for the time too). I'm not sure if FPM DRAM was available (or well priced) yet, but if it was, that could allow significantly improved performance. (I think the Sega CD used 80ns FPM DRAM in 1991, but I'm not sure)

 

Again with the interleaved comment, I was talking about a System with shared memory, with dedicated memory, that's much less of an issue. Though DMA limitations could get pretty tight on the SNES, and especially Genesis -on the Genesis the limit would have been substantially reduced by using a second bank of VRAM -which the VDP supports --or configuring the main VRAM to allow 1x 64kB or 2x 32 kB banks depending on the situation. (former would double the VRAM cost, latter would simply mean using dual ported 8-bit DRAMs rather than 4-bit ones and some simple mechanism to select between the 2 modes)

The Genesis has a few odd things which make you wonder, like why doesn't the YM2612 have its IRQ line connected to either CPU (which would greatly facilitate PCM playback timing), or why did they use PSRAM for 68k memory and PSRAM or SRAM for Z80 memory when DRAM shoudl have been fine and MUCH cheaper. (allowing more RAM to be used and still save on cost -the SNES did that, but at a performance cost the Genesis wouldn't have taken -plus the SNES needed the cost cut more due to having 128 kB of SRAM and the advanced sound subsystem -as well as dual PPUs)

 

Anyway, it all depends on the design philosophy. I's still partial to Crazyaces idea on the Archimedes derivative, assuming his statements on how fast 2D rendering could have been done in software is accurate. (3D would be great for the time and good flexibility overall, 8 DMA sound channels isn't bad either)

Edited April 17, 2010 by kool kitty89

[potatohead]

@kool_kitty: I think the bus on 63 and 6809 chips is shared, leaving the graphics chips to fetch a lot of data without slowing the CPU. That's what the CoCo 3 appears to do.

 

You are right about the specs being too modest in the color and GFX department. I was still kind of retro, or had moved to PC in the 90's, so I'm a bit out of touch on what was possible console wise.

 

Moving to bitmap / blitter would have been an interesting direction, and with that in mind, I think I would have done chips that blended the two. One feature I think would be necessary is lower resolution bitmaps, with higher resolution sprites. That way, lots of motion, walls, etc could have some potential to occur, without losing the detail. One could stack the sprites up for hud displays and such, while presenting some 3d to the user at a level of detail that keeps action high.

 

Were large capacity serial EEPROM and RAM devices cheap then? If so, adding the pins for dealing with one in the cart, and having a RAM on the system board would be a very interesting addition. Very large scale projects would be possible, loading in key bits over time.

Edited April 17, 2010 by potatohead

[avanteproject]

I don't have much knowledge of specs so I was just throwing out numbers pretty much. Make what you want with them. lol

[potatohead]

Just got done reading a bit on the Archimedes... Yeah, that's a compelling idea for sure.

 

At TV graphics resolution, that kind of machine would be bad ass. Doing stuff in software gives a lot of flexibility, and where custom chips provide some good experiences, so can libraries. The geeky fun hardware aspect isn't as sexy though.

 

For geeky hardware on something like that, stuff it in the cart!!