Its 1993, you're in charge of the Jag, what do you do?

[sd32]

Yikes!, it had to do everything in software?

I saw some youtube videos of Super Marathon for Pippin, it runs it really well. Seemed pretty comparable with PS1 FPSs from about the same year.

[Gorf]

Yikes!, it had to do everything in software?

I saw some youtube videos of Super Marathon for Pippin, it runs it really well. Seemed pretty comparable with PS1 FPSs from about the same year.

 

The PPC is a pretty powerful chip....perhaps there is some developer only undisclosed hardware support but all the specs

I've ever seen suggest otherwise. Never did take a look at any of the title.

[kool kitty89]

The Pippin's PowerPC603 CPU was 3 instruction superscalar, right? (meaning that 66 MHz CPU would be rated at ~200 MIPS)

 

So pretty damn powerful, but it had to do everything, rather like contemporary PC's (though 3D hardware acceleration was catching on by then, especially with 3DFX's Glide, along with DirectX/Direcrt3d and Open GL).

All older PC games were basicly all done in software. (games like Doom, X-wing, Tie Fighter, Wing Commander 3/4 very fast, high end CPU's for their time to run at full detail settings)

Edited August 2, 2009 by kool kitty89

[Pete5125]

2) Make it CD based to begin with.

I know that the remaining 20% of the thread has been about this, but it's also worth repeating. In addition to the added storage space for FMV (which really DID sell consoles back then, as bizzare as it seems now), CDs had the added cachet of being "futuristic"--cartridges were so last-gen, but CDs represented the shiny new future of gaming. The problems with low drive-speeds could have been worked out in future revisions of the console, or by partnering with another company--say, one that produced CD drives. Which leads us to the next idea...

 

3) Team up!

 

3b) Team up with SNK

At the same time that Atari was developing the Jaguar, SNK was developing the Neo-Geo CD. The Neo-Geo CD had an incredible library, but was weighed down by it's slooooooooow 1x CD drive and a complete lack of any sort of advertising. But what if they'd teamed up with Atari? The two companies teaming up is almost probable; the corporate headquarters of Atari and the US branch of SNK were, IIRC, across the street from one another, and they were presumably on friendly terms. I remember reading somewhere in this thread that the Jaguar was approximately as powerful as the Neo Geo, so there shouldn't have be many problems porting Neo Geo games (or at least the ones that existed ca. 1994) to the Jaguar, and teaming up with SNK would have ensured a steady supply of third-party titles for Atari. The Jaguar's 2x CD drive would have made most of the Neo Geo CD games much less tedious by cutting out the ridiculous load times between levels/fights. Imagine if our hypothetical CD-based Jaguar had launched with, say, The Art of Fighting (as well as Cybermorph) as a pack-in and had an arcade controller available at launch. The fate of the console might have been very, very different... (see 4b)

 

4) Focus on it's strengths

Although it might not be obvious given all the hate it gets on the internet, the Jaguar was pretty good at some things. Unfortunately, none of those things involved polygonal graphics, which is what most of the games focused on and suffered for. There were two main areas that the Jaguar was good at that Atari should have picked up and focused on:

 

4a) First Person Shooters

The Jaguar was good at FPSs. Even detractors admit that Alien vs Predator was a good game. The Jaguar had the best console version of Wolfenstein 3D, and even Doom, while not great in retrospect, was the best console version available at the time (and was the basis for the GBA port). In the years of the Jaguar's existence, a whole slew of well-regarded FPS came out--Heretic, Rise of the Triad, Duke Nukem 3D, Zero Tolerance, Marathon--all of which the Jaguar was powerful enough to handle (I mean, they ported Duke Nukem 3D to the Genesis, fer godsakes). Were I Atari, I would be buying up the rights left and right and starting work on (preferably exclusive) Jaguar ports for all these games. If the original Xbox taught us anything, it's that you can support a console primarily on FPSs.

 

4b) 2D Games

Let's go back to 3b. Suppose that the CD-Jaguar came out with a reasonably arcade-perfect port of Art of Fighting, and received a steady stream of Neo Geo ports. The console attracts a lot of attention from fighting game fans and people who like Neo Geo games but not enough to pay $200 apiece for them. Atari manages to snag some other licenses too; a good looking port of Mortal Kombat II, some more Raiden games, Tyrian, etc.--in addition to the polygonal shooters that seemed to comprise a significant portion of their library.

 

5) Klax

How do you release a console and not port your flagship franchise to it?

 

I like your ideas I would add this....If you added one more category then you would have a success in the US at this point in history....get EA on board make whatever deal even if you have to produce the games for them....you must have Sports Games, even today Madden Football is huge every release still get a big hooha.

 

Atari already owned the rights to RBI Baseball, Wayne Gretsky Hockey, Jack Nicholous Golf, so do like Sega challenge your programers to make the best version of each category, at this point in history no single game company owned the rights to all NFL or MBL so if you tried making a more Arcade style (NFL BLITZ)game then that would be the way to go.

 

Also get ATARI Arcade games ported to the Jag, obviously you own the rights to the games they were all released to the Lynx, get those games then go after Area51...etc.

Edited August 2, 2009 by Pete5125

[triverse]

Atari Games was a seperate company at least that is what I understood (they split years ago for some reason, probably taxes or something so it is feasable that both were owned by the same company/family).

 

EA was not likely to support the Jaguar since it was in direct competition with the system they backed with games and financially, the 3DO (look it up, EA is one of the 4 or 5 companies that came together to get it off the ground).  EA started supporting the PSOne and Saturn when it was clear the 3DO was dead in the water (supporting 16-Bit was no problem since it was not only a money making area for them, but it was also considered a different tier on a different level from 3DO).

 

 

Now Atari teaming up with SNK, that would have been nice to have seen.  What if SNK and Atari did team up and Atari sold the Neo Geo CD system (with a 2x drive under their name?).  That would shoot backwards compatibility in the foot but they would have arcade perfect ports and a fairly powerful system (SNK has proven it was more powerful than we really thought it was with the newer games they were able to release for their Neo Geo hardware).

[kool kitty89]

I like your ideas I would add this....If you added one more category then you would have a success in the US at this point in history....get EA on board make whatever deal even if you have to produce the games for them....you must have Sports Games, even today Madden Football is huge every release still get a big hooha.

 

Atari already owned the rights to RBI Baseball, Wayne Gretsky Hockey, Jack Nicholous Golf, so do like Sega challenge your programers to make the best version of each category, at this point in history no single game company owned the rights to all NFL or MBL so if you tried making a more Arcade style (NFL BLITZ)game then that would be the way to go.

 

Also get ATARI Arcade games ported to the Jag, obviously you own the rights to the games they were all released to the Lynx, get those games then go after Area51...etc.[/b]

 

Please see my previous response to that post (#471), a lot is impractical, some is reasonable though.

 

Yeah, EA, when (as triverse mentions) 3DO was Trip Hawkins' baby... (inless you're taling about getting him to abandon it, in which case, good luck...)

 

 

Atari Games was a seperate company at least that is what I understood (they split years ago for some reason, probably taxes or something so it is feasable that both were owned by the same company/family).

 

Atari Inc. (the original Atari, founded By Bushnell and Dabney), ceased to exist in 1984 when Warner Comunications (then Owner of Atari Inc.) sold it to Jack Tramiel (former president of Commodore), who formed the new Atari Corp. to market what would become the Atari ST computer. Warner kept the arcade portion of the original company and formed Atari Games. Now, as too cooperation between the 2, Warner still had holdings in the new Atari Corp (a decent chunk of stock), and thus, it was in Warner/Atari Games' best interest for Atari Corp to do well. (it went both ways, too, for example Atari Games had the Jaguar derived "cojag" arcade system) Also note Atari Games is the same company to release games under the "Tengen" label. (notorious for the debacle with Nintendo over their unlicenced games using reverse engineered lockout hardware)

 

 

Now Atari teaming up with SNK, that would have been nice to have seen. What if SNK and Atari did team up and Atari sold the Neo Geo CD system (with a 2x drive under their name?). That would shoot backwards compatibility in the foot but they would have arcade perfect ports and a fairly powerful system (SNK has proven it was more powerful than we really thought it was with the newer games they were able to release for their Neo Geo hardware).

 

Bad idea, the Neo Geo CD required over 7 MB of RAM and it's own specific hardware, which would need to be included on top of the Jag's, not a very efficient design to work with. And as to the "power" of the Neo Geo, its greatest strengths came from the massive amounts of memory it used for the time (huge ROM cartridges, compensated for by a rather large chunk of RAM and multiple loads in the CD version), and it had lmited color capabilities compared to the Jag, and don't even think about 3D (or psudo 3D, scaling, or rotation). It was a sprite monster (an all sprite system actually), and impressive for its time, but the Jag could out-do it in 2D, at least when properly utilized (possibly even using the 68k). What the Jag may have had trouble with is sprite animation, too little RAM to fit all the animation tiles, the Saturn managed to adress this problem with RAM expansion cartridges (and in one case a companion ROM cart).

With the cartridge based Jag you might actually be able to get around this without expanding RAM, by streaming ROM off the cartridge; while this is limited (6 MB w/out bank switching, and keep in min ROM is expensive, so you wouldn't want to go much higher anyway), the Jag has the advantage of compression, using one of the RISC processors to decompress it and dtream the data in real time. (IIRC Gorf mentioned something like 14:1 compression being practical, which would allow a 6 MB cart to hold the equivelent of 84 MB of uncompressed data, that's 672 Mega Bits, or close to the size of the largest Neo Geo games)

 

As to the CD based Jag, you could have provisions for RAM expansion, like the Saturn, for virtually Arcade perfect (if not enhanced) ports not lacking any animation.

 

I don't think having any Neo Geo hardware onboard would be a good option at all, very impractical in terms of material costs and board space it would take up, and the large amount of RAM that would be difficult for the Jag Hardware to properly take advantage of (being on a 16-bit data bus, or 8-bit for the Z-80 portions, inless you modified that too, but that could be problematic) Really, just a bad idea.

 

I'm still a bit skeptical of gettting SNK to actually partner with Atari, I suppose the Jag's 2D strengths might be attractive. (maybe even licence it to them as an arcade board)

 

I still think, if anything NEC would be the best choice (though again, any partnership could be rather shakey), in particular they had the Manufacturing capabilities to produce the hardware, so a close partnership could be good, of couse added funding would be great. Additionally, they had a name in Japan, with the successful PC Engine and successive CD unit (2nd to Nintendo in the console market), meaning they could market a rebranded Jaguar in Japan quite well. 1993 would be a good time too, as their current home console was starting toshow its age and they never did get a proper successor out. Having Hudsen Soft as a developer would be a boost too.

But who knows if NEC would have ever been willing to make such an arrangement.

[sd32]

I dont know if it would have been posible, but having NEC support the Jaguar back in the day like you guys are suggesting would be great. Imagine the Jaggy getting "64 Bit" sequels to all the major PC Engine games. Bonk 64, Lords and Gate of Thunder 64, Valis 64, Spriggan 64, Sapphire 64, YS 64, Tengai Makyou 64... i could go on and on. Imagine this games with the Jags 2d capabilities...amazing! Add to this the Atari lineup and well, it still wouldnt have beat PS1 and Saturn but it would have done much better i guess.

Man now i have a new wet dream...

[Crazyace]

NEC did have the PC-FX though - which probally matched the Jaguar in some ways. ( It had 3D plus a lot of sprite h/w )

[minuS]

NEC did have the PC-FX though - which probally matched the Jaguar in some ways. ( It had 3D plus a lot of sprite h/w )

I thought the PC-FX was all high-quality FMV and no 3D...

 

Anyway, the whole NEC idea does sound kind of cool in a fantastic sort of way; market the hardware as Atari in North America and as NEC in Japan...

[kool kitty89]

NEC did have the PC-FX though - which probally matched the Jaguar in some ways. ( It had 3D plus a lot of sprite h/w )

 

Yeah, I forgot about the PC FX, still, who knows how they'd have reacted to the Jaguar hardware, not to mention possible modifications they might have wanted. Still, they'd have to choose between their (then shelved) Iron Man hardware or Flare/Atari's. Atari's would seem equal or better in both cost and capabilities (except Video playback which the PC FX shine in, but I'm not sure whether that was even part of existing Iron Man hardware), though I'd immagine they'd have wanted to make some modifications to the design compared to the existing hardware, probably similar to what Gorf's been suggesting. (though probably different in some areas)

Assuming such a partnership occured some time in early '93, you'd have some time to make hardware revisions, but probably not go with major redesigns, ie work on hardware bugs, use an EC020 host (be it 16.7 or 25 MHz), and double buffer the blitter. (not sure about other modifications to Jerry) You could probably bump the system speed up a bit too, previously ~33 MHz would be practical (with coresponding increase in bandwidth as well), but faster would get into expensive RAM, at those speeds you could also have a 16.7 MHz clocked close to its full speed rather than 13.3 MHz of the current system.

I don't think redesigning the system with dedicated memory blocks would probably not be possible in the timeframe, the shared system bus was one of the key features (and probably in large part to keep cost low), along with the host processor needed to boot the system. Things like increased RAM or CD drive would be other decisions, but not stricly related to development of the hardware. (CD would be definitel with NEC, having already had a CD basedunit as their main console for a while)

 

But, again, with NEC you've got the manufacturing capabilities, so no middle man to deal with for chips or drives, so maybe they could keep cost down even with a 2x speed CD drive, maybe even with 4 MB of RAM. (or sticking with 2 MB but adding provisions for RAM expansion if cost still starts getting high, though you could be selling the Unit at cost, or below, relying on software sales and licencing royaltees for profits)

 

 

 

Also, thinking about the PC-FX's video playback capabilities, while the Jag CD only used Cinepak, you could develop a custom codec specifically talored to the hardware which would be substancially better. (higher compression, and more colors, at least 16-bit, Cinepak only using a 256 color format)

Edited August 3, 2009 by kool kitty89

[sd32]

NEC did have the PC-FX though - which probally matched the Jaguar in some ways. ( It had 3D plus a lot of sprite h/w )

 

Really, the PCFX had 3D capabilities?, i had only heard that it had great sprite and background capabilities and zero support for polygons. And of course it was great at FMV...

It had 3.75 MB of RAM in total, but only 256KB for video it seems. Oh, and it was stuck with the PCEngine sound chip!...WTF!!

I have wondered for a while why the Graphics Man had a PCFX image on his avatar, a fan maybe?

So without the Playstation getting released, i guess the Sega Saturn, Atari Jaguar and NEC PCFX would have duked it out for the crown of best 2d hardware, right?, i would have loved to see that, we needed a 2d super consoles war!. I wonder which had the technical edge in pure 2d...

[kool kitty89]

I thought there was supposed to be an add-on for 3D, but it was never released. I don't know about onboard hardware, but the CPU was fast enough (21 MHz/15.5 MIPS RISC) to do simple 3D in software, fairly limited though. (depending on what kind of hardware support there was)

 

Also, as to sound, it looks like they added a second ADPCM channel to the CD/Duo's. (so CD-DA, 6 PCE sound channels, and 2x 31.5 kHz ADPCM channels)

 

 

As to a scenario with no Playstation, who knows what would happen, I kind of doubt the PC FX would have done too much better (the hardware really was kind of dated), though perhaps NEC would have gone for a new, updated design without the threat from Sony, likewise the Saturn might not have been so rushed. (though there are some intrinsic problems with the Saturn in certain areas that are not bug related, still it was sstrong in 2D, and less rushing could have meant better tools/dev kits before launch, though there'd still be Sega's internal problems to considder, not a big problem for the JP market, but really screwed Sega of America)

 

And who knows what would have happened with Nintendo with Sony not beating them down. (though maybe some frustrated developers would leave for NEC or Sega instead of Sony)

Edited August 4, 2009 by kool kitty89

[triverse]

I will check into finding the PC-FX coverage in Gamefan but it was going to get some 3D games, one of which was an updated Star Soldier type shooter that was sprite based from what I understand with a view similar to Axelay (with the angled vertical scrolling levels).

[Crazyace]

I dont know too much about the PC-FX , just basing the comment on the HuC6273 ( 3D chip ). It doesn't look like many PC-FX games were actually 3D ( apart from things like Super star soldier, which never came out )

 

Having NEC support the Jaguar would have helped with software though

[triverse]

I agree with you Crazyace, there wasn't anything really to show that the system could do 3D released.  The shots they showed in Gamefan could easily have been mock ups.  A system like PSOne would not have been able to produce those graphics in real time so I am sure a system like PC-FX would have been experiencing trouble with it.

[kool kitty89]

Interesting to note that the Virtual Boy used the same CPU as the PC-FX.

[sd32]

I did some research on some PCFX forums and it indeed didnt have hardware support for polygons. I found some comments where it was said that it could do 3d like Sega 32X did, trough RLE. I am not a technical guy, so i dont know what that means...

Still, suppousedly its very impressive in 2d, better than the Jag?

[kool kitty89]

"like the 32x" would mean doing 3D in software on the CPU, but judging by the CPU, it'd probably be closer to 3D on the Game Boy advance than the 32x. (though probably a bit better thna the than the GBA, as I think the CPU is a bit faster and has a cache unlike the 3DO or GBA) PC's up to this point were the same way, as were MAC's, it wasn't until ~1995 that 3D acceleration hardware started to become popular. (as mentioned above the Pippin was also all software, granted with a very powerful CPU)

 

The 3DO didn't have hardware support for 3D either I beleive (or at least not for the 3D calculations), and its CPU was actually a bit slower than the GBA's, but it had hardware support for tecture mapping and the GPU would draw the polygons with the CPU handeling the 3D calculations. (the Saturn was kind of like this too, but with much more support and the disadvantage of using quads instead of triangle polygons, with much more CPU power of course, the PlayStation had the Geometry Transformation Engine coprocessor to handel all the 3D calculations, so no load on the CPU) The Jag would do the calcultions on one of the RISC processors (Tom or Jerry), but had to do texture mapping in software as well (unlike the 3DO), there was just hardware support for gouraud shading, which I'm not sure the 3DO had. (still a nice feature to have though, a lot better than flat shading) I think there was also hardware support for z-buffering.

[Crazyace]

I did some research on some PCFX forums and it indeed didnt have hardware support for polygons. I found some comments where it was said that it could do 3d like Sega 32X did, trough RLE. I am not a technical guy, so i dont know what that means...

Still, suppousedly its very impressive in 2d, better than the Jag?

 

It's a strange system. I looked at the source code for an emulator ( as there's little real technical info ) - and that's where I found the 3d chip spec. Maybe it was broken/missing in the machine though - as the games definitely don't seem to use 3D.

[kool kitty89]

Crazyace, some articles (including the wiki one) mention the PC-FX had a planned 3D expansion module to plug into an expansion port, but ended up never being released, perhaps that's what's being refered to. (the HuC6273 being onboard the module)

 

Edit, found this article which mentions the 6273 being on the board. http://translate.google.com/translate?hl=e...%3Doff%26sa%3DG Translated though so it's a tad confusing,

Edited August 6, 2009 by kool kitty89

[kskunk]

The 3DO didn't have hardware support for 3D either I beleive (or at least not for the 3D calculations), and its CPU was actually a bit slower than the GBA's, but it had hardware support for tecture mapping and the GPU would draw the polygons

 

Actually, the 3DO does have a fixed point matrix math coprocessor. This is pretty similar to the Playstation's Geometry Transform Engine. It does all the 3D math at speeds similar to the Jag RISC.

 

The 3DO is often criticized for being 'not 3D enough' due to its lack of a triangle based rasterizer. Instead it uses quads that can be scaled and warped, just like the Sega Saturn. This type of hardware can be used as a sort of rectangular rasterizer and offers performance benefits at the cost of a lot of extra difficulty in creating models.

 

The Jaguar's hardware can only rasterize sprites (unrotated rectangles), not any kind of polygon (triangle or quad). Of course software is used to turn the blitter into a proper rasterizer, but that cuts into overall performance. As Gorf points out, handling rasterization in software opens up unconventional types of rasterization, such as "voxel engines" (height map rasterizers).

 

The 3DO can gouraud shade like the Jaguar, but not as quickly. Also the 3DO has no Z-buffer.

 

I wouldn't say one is clearly better or worse than the other at 3D, they're just different. The 3DO is much better at texture mapping and a few games like Need For Speed really show that off. The Jaguar is much better at gouraud shading as shown by the really great frame rates achieved by games that go light on the texturing.

 

In the end both the Jaguar and 3DO suffered from bandwidth-hungry CPUs that starve their graphics chips of their true potential. Both had nice tech demos that turned into (mostly) choppy games. Neither the Saturn or Playstation had this problem because both used CPU caches and in the case of the Playstation, a texture mapping cache as well.

 

Bandwidth aside, the ARM in the 3DO was pretty nice. It was a lot faster than a 25MHz '020 or a Jag RISC running out of main. It could even hold its own against a Jag RISC running in on-chip RAM. And of course it was very C-friendly.

 

- KS

Edited August 7, 2009 by kskunk

[Crazyace]

Crazyace, some articles (including the wiki one) mention the PC-FX had a planned 3D expansion module to plug into an expansion port, but ended up never being released, perhaps that's what's being refered to. (the HuC6273 being onboard the module)

 

Edit, found this article which mentions the 6273 being on the board. http://translate.google.com/translate?hl=e...%3Doff%26sa%3DG Translated though so it's a tad confusing,

 

That's cool - but I think it's talking about the PC card rather than an add-on for the console.

To me - it seems like the PC-FX is completely unbalanced. For 2D it has the PCE VDP - giving sprites and a single scroll plane, and then a second graphics chip that only gives 4 extra scroll planes. So there are no extra sprite capabilities at all. ( Even the older duo had better sprites due to duplicate VDPs ) It only makes sense if the 3D part was there to provide sprites as well ( in the same way that the Saturn VDP1 provides 'sprites' to be merged with the VDP2 backgrounds )

It would be interesting to see hear from Hudson about what happened ( and what went wrong ) - although I guess that's really unlikely

[Crazyace]

Luckily the Jaguar didn't have to match the 3D0 - it was in a completely different price bracket - and for some of the most popular pc games ( Doom style raycasters ) the Jaguar could hold it's own. ( And for 2D it was equal to anything present in home or arcade )

[Gorf]

The 3DO didn't have hardware support for 3D either I beleive (or at least not for the 3D calculations), and its CPU was actually a bit slower than the GBA's, but it had hardware support for tecture mapping and the GPU would draw the polygons

 

Actually, the 3DO does have a fixed point matrix math coprocessor. This is pretty similar to the Playstation's Geometry Transform Engine. It does all the 3D math at speeds similar to the Jag RISC.

 

The 3DO is often criticized for being 'not 3D enough' due to its lack of a triangle based rasterizer. Instead it uses quads that can be scaled and warped, just like the Sega Saturn. This type of hardware can be used as a sort of rectangular rasterizer and offers performance benefits at the cost of a lot of extra difficulty in creating models.

 

The Jaguar's hardware can only rasterize sprites (unrotated rectangles), not any kind of polygon (triangle or quad). Of course software is used to turn the blitter into a proper rasterizer, but that cuts into overall performance. As Gorf points out, handling rasterization in software opens up unconventional types of rasterization, such as "voxel engines" (height map rasterizers).

 

The 3DO can gouraud shade like the Jaguar, but not as quickly. Also the 3DO has no Z-buffer.

 

I wouldn't say one is clearly better or worse than the other at 3D, they're just different. The 3DO is much better at texture mapping and a few games like Need For Speed really show that off. The Jaguar is much better at gouraud shading as shown by the really great frame rates achieved by games that go light on the texturing.

 

In the end both the Jaguar and 3DO suffered from bandwidth-hungry CPUs that starve their graphics chips of their true potential. Both had nice tech demos that turned into (mostly) choppy games. Neither the Saturn or Playstation had this problem because both used CPU caches and in the case of the Playstation, a texture mapping cache as well.

 

Bandwidth aside, the ARM in the 3DO was pretty nice. It was a lot faster than a 25MHz '020 or a Jag RISC running out of main. It could even hold its own against a Jag RISC running in on-chip RAM. And of course it was very C-friendly.

 

- KS

 

I'd believe is would beat the 020 but I bet its ARM is not faster than a J-RISC out in main. Remember the J-RISC's are

more than twice the clock speed in the Jaguar.If you use to many tight loops the ARM can beat it but we have run code

in main and in local that run almost the same speed. It depends on how much you rely on registers/internal memory vs

external memory and how tight your loops are. Tight loops should stay in the local.

 

I do like ARM's nice conditional instructions and the highlevel aspect, but I'd still go with the J-RISC at twice the speed,

as primitive as they may be in high level ability and a few other things.

 

With the right code i think the Jag will outdo the 3DO even in T-mapping. Most games Ive seen for the 3DO are 256 color

modes tend to be lower than 30 in frame rate. Take the HoverStrike CD code and moev most of the 68k crap to GPU main

and I bet it will run very fast and with all the effects it uses, I dont recall any 3DO games that can pull of the same lighting

and shading. 16 bit cry would probably hurt the 3DO.

[kool kitty89]

Actually, the 3DO does have a fixed point matrix math coprocessor. This is pretty similar to the Playstation's Geometry Transform Engine. It does all the 3D math at speeds similar to the Jag RISC.

 

The 3DO is often criticized for being 'not 3D enough' due to its lack of a triangle based rasterizer. Instead it uses quads that can be scaled and warped, just like the Sega Saturn. This type of hardware can be used as a sort of rectangular rasterizer and offers performance benefits at the cost of a lot of extra difficulty in creating models.

 

Hmm, that's interesting, I wonder why they went with quads, perhaps for the Same reason the Saturn did (for sprite tiles), seems like a bad decision though. Could the coprocessor still be used if software rendering was done with triangles? (I'd assume the texture mapping and maybe even g-shading wouldn't work coupled with software rendering, so that would hve to be done in software too, so that would be pretty unattractive, especially with the issues with the current CPU/bus)

 

 

 

In the end both the Jaguar and 3DO suffered from bandwidth-hungry CPUs that starve their graphics chips of their true potential. Both had nice tech demos that turned into (mostly) choppy games. Neither the Saturn or Playstation had this problem because both used CPU caches and in the case of the Playstation, a texture mapping cache as well.

 

Bandwidth aside, the ARM in the 3DO was pretty nice. It was a lot faster than a 25MHz '020 or a Jag RISC running out of main. It could even hold its own against a Jag RISC running in on-chip RAM. And of course it was very C-friendly.

 

Wouldn't the 020 still have the advantage of the cache though? (so wouldn't it have similar benefits as using it in the Jag?)

 

Still, had the ARM CPU had cache, that problem would have been resolved, but looking at the models of ARM processors around, it looks like the ARM610, which would be a bit faster as well, though I have no idea on the cost. (others like the ARM600 would probably be too expensive, and have unnecessary features, and older chips lie ARM3 would probably be expensive too, and using older manufacturing processes and probably generating significantly more heat)

Or they could have designed the 3DO to have a dedicated video bus, but all these options (excpt the 020) would have added to cost (of an already expensive machine, but that's related to other issues).

I seem to remember that the 3DO used VRAM, so that probably added a lot to cost too, so maybe they could have saved money using a dedicated video bus with less VRAM (something like 512 kB, though that would probabably force them to drop the 320x480 mode, inless there was an 8-bit color mode), then keep the 2 MB of main but use cheaper DRAM instead.

 

edit: I just noticed that the 3DO actually already has 1 MB of VRAM separate from main, so why would it still have the bus shaing problems similar to the Jaguar? (except for the general 50 MB/s bandwidth limitation) http://www.diehardgamer.com/index.php?opti...s&Itemid=66

 

A previous post you made seems to imply a shared bus too, so is the 1MB VRAM thing wrong?

I've managed to dig up some technical docs on the 3DO, but they're hard to find. It's an interesting counterpoint to the Jaguar because it meets the market requirements of a next gen system -- it's just really underpowered. On the other hand, the Jaguar is far more powerful, but of course Atari marketing was a bunch of tossers and had no idea what the market wanted. We all knew that.

 

The 3DO was designed to be programmed in C and texture mapping is a core feature, two things the Jaguar missed out on. However, the 3DO's clock rate is slower compared to the Jaguar, and the custom chips are on an older process technology, which makes the chipset bigger, less powerful, and more expensive than Tom and Jerry.

 

The 3DO's ARM is weaker than the Jag RISC, but has the advantage of an excellent good C compiler and a bug-free memory interface, making large programs a cakewalk to make.

 

Like the Jaguar, there is no cache, so main bus bandwidth is a severe restriction. While the 3DO is texture mapping the ARM CPU is frozen.

 

This led to awesome graphics demos that turned into choppy games, since gameplay logic killed graphics performance.

 

 

Luckily the Jaguar didn't have to match the 3D0 - it was in a completely different price bracket - and for some of the most popular pc games ( Doom style raycasters ) the Jaguar could hold it's own. ( And for 2D it was equal to anything present in home or arcade )

 

Yeah, that was probably the biggest mistake with the 3DO, much more than any technical limitations (which it definitly had an advantage over the Jag for ease of programming), it was the sales philosophy that killed it, along with some other mistakes. (some site the overemphesis on the M2) The Jag had eve worse problems in this respect, on top of the technical issues, and the managementproblems do tie inot some of the technical problems as well.

 

But the 3DO didn't use the typical razor+blade markeiting which is pretty much standard for the console market, sell the base unit as cheaply as possible, and make money from game sales. 9either first party or licencing/royaltees) The 3DO's licencing fees were virtually nonexistant, which was the big problem, plus the licenced production made things worse, with competing manufactures and nonstandardization. (licenced hardware would have been good had they used the standard razor/blade marketing, but they still would have been best ro standardize the design so as to aviod confusion, with the usual cost reduced form factor progression any console goes through)

Still, a single manufacturer would have simplified things, just using the partnered Pannasonic and th e"razor-blade" strategy, selling the hardware at, or close to cost (or even take a loss), and focus on profiting through the software sales. (they probably should have set up more 1st party stuff as well, maybe merged with a developer like Sony did with Psygnosos, EA would kind of be a natural choice)

Edited August 8, 2009 by kool kitty89