Were the Atari ST's big for gaming or just the 8 bit line?

[kool kitty89]

At least in the UK I think a large Problem for Commodore, Atari & maybe even Acorn was that they were focussing there attention on their perceived competition i.e each other when the real threat were the rising dominance of PCs for serious stuff and consoles for gaming. They didn't help by splitting scarce resources to compete in the PC market when what they needed to do is focus on there strengths and weaknesses. In some respects I think Commodore UK towards the end new what needed to be done and it was unfortunate they weren't successful in their buyout proposal.

Yeah, it certainly seems that way, especially with the design of the STe...

 

Though the ST was slow to upgrade in general other than RAM and the BLiTTER in Mega models ~1987, even stranger as the PC obviously would have been the main competition to match or beat if they wanted any chance of competing in the US. The Amiga was even slower to expand in some respects, but it was better off (tech wise) from the start.

 

But yes, I think Atari would have been better off aiming at being better than the PC and Mac rather than following the Amiga. In the US it was niche (more or less) and needed some sort of consistent advantage to hold on as well as enough publicity for people to actually know about it. In Europe it was more or less the de facto mainstream 16-bit computer in the late 80s and in both cases it was simple enough to have serious potential as becoming a broader standard with clones. One question would be whether either the ST or Amiga (with licensing/clones/broader standard) would have become cost effective enough to block out PCs, especially with the CPU issue. (Motorola being rather closed with all post 68000 chips in the 68k architecture thus limiting it in price competitive terms against x86 -with Intel, TI, Cyrix, AMD, NEC, etc -the latter was more of a clone then licensed iirc and only went up to 286 class CPUs) It wouldn't have stopped faster 68000s though (which topped out commercially at 20 MHz with 16 MHz already there in the mid/late 80s -I don't think there's any technical reason that they couldn't be pushed much further) and fast 68000s would have been competitive in the low-end in the early 90s. (68EC020s and 030s would be pretty reasonable for mid-range by then too... but the line died with the 68060 and the Amiga and ST/Falcon had started pushing for PowerPC accelerators towards the end -and the post Commodore Amiga OS4 switched to PowerPC specifically -and Apple went that way too of course, but how soon any architectural switch occurred would depend on a lot of factors assuming either the ST or Amiga had lived on in the mainstream in Europe or niche in the US -commercially, not like what happened with the post CBM Amiga stuff)

 

 

 

 

Technically speaking the PC was the thing the ST had the most universal advantage of on the market and the bar may have been set at the Mac, but the PC truly was the emerging standard (which should have been rather clear by 1985 even). At the start, the ST was better in OS, CPU, sound, graphics, and price (by far) than any mainstream PC and EGA had a couple advantages with 640x200 with 16 CGA default colors (intended to be compatible with CGA monitors) and 640x350x16 colors limited to 6-bit RGB and only for full 256 kB EGA cards: with lower-end 64k would only be able to do 640x350 in monochome (which the ST could outdo with 640x400 mono) and EGA was also expensive up until ~1987 with the release of MCGA/VGA. That's when the ST'g graphics started to get more dated: PCs got 18-bit RGB with 256 indexed color mode (320x200 -outside of "mode X") with packed pixels as well as a 640x480 16 color (planar) indexed mode and expanded high res text modes (and more rows of text) with hardware scrolling added on top of all that. So not only did it add to the maximum color and graphics capabilities in general, but the packed pixels and hardware scrolling made the 320x200x8bpp mode 13h ideal for games and higher performing with similar CPU resource than EGA or CGA and it became common for games far faster than EGA had following its 1984 release. 1987 also marked the introduction of the Adlib card followed by several others (many short lived) in 1988 but also the high-end MT-32 MIDI module that some developers started supporting very quickly (interestingly on the ST as well in a few cases -and obviously a good thing to couple with the ST for actual composers/musicians using the ST with its built-in MIDI interface)

 

The ST got a blitter in '87 but that wasn't standard to any lower-end models until 1989 with the STe (I think at one point they'd planned an upgrade for other models, but it was dropped), and the STe added a 12-bit RGB palette (still a 16 color display) and the 2 channel PCM chip as well but that was all pushing more towards an approximation of the Amiga and was too late in any case. Before any dedicated blitter or such they should have focused on simply adding hardware scrolling to the SHIFTER as soon as possible (probably 1986) and standardizing that for all machines (meaning like with the 360 kB drive only the very early models would have been lacking -and they could have offered upgrades to mitigate that -and floppy drives could be upgraded as it was), and still before pushing for a blitter they should have aimed at providing packed pixel modes and increasing the color depth (even if packed pixel was limited to the higher color modes like with VGA) and probably bumping up the color depth of existing modes -back when they added scrolling, maybe add some other simple acceleration like copy or line fill functions or at least with the higher color/packed pixel modes. (VGA added some other simple acceleration like that to aid software blits -a simple line fill function is useful for clearing frame buffer area and for flat shaded/filled polygon games) To meet VGA they probably should have pushed that by '88 and at that point you've got notable advantages over the amiga as well (no packed pixels, no 256 color bitmap mode). Adding genlock probably would have been significant as well. (cut into the Amiga's video editing niche)

CPU speed should have been offered at higher speeds from the start (higher end models with 12 or 16 MHz CPUs) and later shifting to lower end models being fast by default. If they went that route, they wouldn't really need to heavily focus on a blitter either, but focus on the simpler graphical upgrades and CPU specifically (PCs didn't start pushing blitters until the early 90s and primarily for GUI acceleration and certain drawing applications, not for games until the mid 90s with API level programming becoming common in win 9x and 2D+3D accelerators appearing)

 

Atari had the TT030 in 1990 but that was also too late and way too high-end. (maybe if it had been more like a MEGA STE with TT video it would have done better -still pretty late nonetheless) And one thing that would be important is if they DID push a blitter for the ST and did so after introducing packed pixel modes, they could have designed it to specifically be tuned to packed pixels, and more than that be considerably faster at many things than the Amiga blitter. (they probably didn't need it though)

 

And for audio they also should have pushed much sooner: there was the very clean option of switching to the YM2203 (fully 2149 compatible with 3 added 4-op FM synth channels) and worked on getting some sort of PCM chip out as well. (the 2203 was available in 1985 so that's another thing they could have pushed on higher-end machines from the start)

 

Expandability was one disadvantage with PCs and they really probably should have added a general purpose expansion port rather than the cart slot and internal expansion ports on higher-end machines. (they probably should have pushed the MEGA form factor by '86 at least)

And perhaps pushed for higher-end 32-bit versions earlier as well, but not as dramatic of a jump as with the TT. (like a 12 or 16 MHz 68020 -and regardless of 32-bit, maybe offer FPU sockets for MEGAs as well with optional 68881/2 -another thing the Amiga was lacking in)

[+wood_jl]

Technically speaking the PC was the thing the ST had the most universal advantage of on the market and the bar may have been set at the Mac, but the PC truly was the emerging standard (which should have been rather clear by 1985 even).

At the start, the ST was better in OS, CPU, sound, graphics, and price (by far) than any mainstream PC and EGA had a couple advantages with 640x200 with 16 CGA default colors (intended to be compatible with CGA monitors) and 640x350x16 colors limited to 6-bit RGB and only for full 256 kB EGA cards: with lower-end 64k would only be able to do 640x350 in monochome (which the ST could outdo with 640x400 mono) and EGA was also expensive up until ~1987 with the release of MCGA/VGA. That's when the ST'g graphics started to get more dated:

 

It's just REALLY, REALLY, easy to forget - in modern times looking back - how $EXPENSIVE$ computing power was, in 1985. Sure, an EGA PC had awesome graphics for that time, but how much did it cost? How many multiples of a $1000 ST system would it cost for a 286 system with ample RAM, and a 256K EGA card? It also took some time for the PC software to cater to the new graphics modes; when I got my first 286 (with VGA) in 1990, the VGA had been out for a long time, and almost all the games were in CGA/EGA still; it was depressing.

 

Even the price of the RAM was so damn expensive, it's just difficult to fathom today. The Mega 4 ST was like $2400 and the only [major] difference was more RAM.

 

Today, when el-cheapo computers have 4GB of RAM, the specifications of an ST system - and the PC system that would best it - both seem so laughably anemic that they appear similarly so; it's easy to diminish the difference in price they had. Each additional "unit" of computer performance/capability meant a skyrocket in price back then. Also, in the very late 80's to early 90's - PC prices (and "units of performance") plummeted at an alarming rate, probably exponential in nature if it were graphed. So the scene in 1989 looked markedly different than 1988. The "clone" scene took off almost overnight, and that, in turn, lowered prices even more. Nobody could have seen this; it was completely unprecedented. Absolutely. The state of the personal computing had *never* changed so fast. Although it continues to change, it would be surprising to see anything as watershed as the clone movement. Just evolution.

[kool kitty89]

It's just REALLY, REALLY, easy to forget - in modern times looking back - how $EXPENSIVE$ computing power was, in 1985. Sure, an EGA PC had awesome graphics for that time, but how much did it cost? How many multiples of a $1000 ST system would it cost for a 286 system with ample RAM, and a 256K EGA card? It also took some time for the PC software to cater to the new graphics modes; when I got my first 286 (with VGA) in 1990, the VGA had been out for a long time, and almost all the games were in CGA/EGA still; it was depressing.

I was taking all that into account, sorry if I wasn't clear. I blended some of my comments a bit too much I think. The general comparison was of overall capabilities and advantages/disadvantages, cost was one of those things, but remember that there are cases where cost is not the factor, but pure capability is. (if you needed a high-end work station with floating point and high res color capabilities, the PC was pretty much the only way to go at the time -no Amiga or ST models with FPUs introduced until later, or faster 68000s even, and amiga had high res 16 color but flicker fixers didn't appear until later either -and for some applications, monochome would b fine and thus the ST's 640x400 mode would have an advantage over EGA if it was available in workstation form factor -but even when the MEGA did appear it wasn't available with FPU or 16 MHz 68k... a shame Atari didn't put a bit more focus on higher-end or even workstation class ST derivatives earlier, not their main focus but much sounder than the SUPER high-end Transputer )

Such high-end PCs were niche of course, but the niche market was ever getting closer to the only market viable for non PCs in the late 80s. (in the US)

 

But for the mass market users, EGA wouldn't come until later in the late 80s when clone PCs and cheaper clone EGA cards would become common. And note of course than most if not all EGA games ran in the 320x200 CGA color mode and EGA was planar like the ST and wouldn't get any performance advantage there, so games would look worse than on the ST with its nice 9-bit RGB (even with indexed colors EGA was only 6-bit RGB). For graphics design, the ST would obviously be a more attractive option regardless of cost as well due to the color capabilities, but the Amiga would be more attractive still.

From the mass market standpoint that 320x200 16 CGA color mode was the same as what the PCJr and Tandy 1000 offered in 1984 at mass market consumer prices AND with a SN76489 sound chip like the SMS or Colecovision (a bit weaker than the ST/MSX/Intellivision/CPC/Spectrum's AY/YM with a smaller frequency range and no ADSR and less flexibility with noise generation -but PCs didn't get anything at all until 1987 with adlib, not even things like the mockingboard or high-end midi modules -MT32 appeared in '87 as well). I think the TGA graphics also used packed pixels rather than planar graphics, so a boost over EGA or the ST in that sense (there was also the low res 160x200 16-color mode used for some earlier games (several also using artifact colors to approximate that in CGA versions), but the mid/late 80s EGA era stuff generally used the same res and graphics as EGA did) Of course, being limited to Radio Shack in-store distribution weakened the Tandy machines as a common standard but they by far would have been the best consumer level PCs on the market until the end of the 80s on a pure tech/price standpoint, of course it probably would have been hugely popular in the mass market if IBM had designed the PCJr to be like that rather than a funky proprietary set-up with funky keyboard. (ie make it a full ISA PC like Tandy did -albeit a few of the later low-profile console form factor models used proprietary expansion cards)

There actually was the Plantronics Colorplus card which supported a very similar feature set to TGA but it apparently never caught on beyond use for CGA and more video pages with the 32 kB vs 16 kB of CGA.

 

VGA was a massive leap that put color capabilities over the amiga (320x200x256 and 640x480x16 from 18-bit RGB -the latter only for full 256 kB cards) and immediately boosted capabilities for gaming with packed pixel graphics, hardware scrolling, and some other acceleration to aid with software blits, and 64 kB models being all that was needed to support the popular 13h 320x200 color mode (mode X would need more as would using 640x480 above monochrome) plus you'd also get 320x200 and 640x200 16-color EGA modes plus CGA compatibility. (unless you had one of the IBM MCGA cards which I don't think supported the EGA modes)

 

 

 

Even the price of the RAM was so damn expensive, it's just difficult to fathom today. The Mega 4 ST was like $2400 and the only [major] difference was more RAM.

Sometimes the price difference doesn't reflect the actual component cost. In some cases it's purely up to the inflated cost from going from more expensive components to final retail price, but in other cases the profit margins are pushed higher for whatever reason. (ie it would be cheaper to buy a lower-end model and upgrade the RAM yourself -if you could upgrade the RAM; I believe that actually happened back with the PET to the point that they removed the RAM sockets from the boards of lower-end models to prevent user expansion)

RAM was an expensive component and there was the odd trend of a jump up in price following 1987 and not settling back down until 1990 (and then prices stagnated from 1992-1996), at least looking at the common mass market RAM prices (ie the vast majority of RAM sold would be lower-end stuff with fast high-end RAM being a very small percentage). The ST and Amiga's DRAM should have fallen well within the mass market price range.

See:

http://phe.rockefeller.edu/LogletLab/DRAM/dram.htm

Of course there's also the added board space used for more RAM until higher density chips become affordable. (in 1987 they'd be sticking to 32 kB chips, so 4 MB would take a lot of board space with 128 individual 12-pin DIP DRAM chips, at some point switching to more expensive higher-density RAM would be cheaper due to sheer space, but in '87 128 kB chips were over 60% more expensive but the next year they were much closer... though due to 32 kB chips jumping back up in price while 128 kB chips only jumped up moderately -both still much cheaper than 1985 prices, but a good deal more than 1986/87) That would also vary depending on the manufacturing region. (I believe the Reagan administration was attempting to block Japanese imports from destroying the domestic DRAM market with price dumping, but all that did was drive up domestic RAM prices all around until the whole thing was dropped anyway)

Once they started using 128 kB chips though, that would cut board space down hugely (for all models -ie 520s would use 4 chips rather than 16, 1040s only 8 chips rather than 32, etc, etc)

From a pure component price though, the difference between the MEGA 2 and MEGA 4 in 1987 should have been in the area of $150 and the 520 and 1040 would be roughly $40, but what the minimum translation was for the retail price is another matter. (aside from changes in profit margin, or keeping profit margin the same but that margin adding more to the end price due to the increase in base cost)

 

 

 

Today, when el-cheapo computers have 4GB of RAM, the specifications of an ST system - and the PC system that would best it - both seem so laughably anemic that they appear similarly so; it's easy to diminish the difference in price they had. Each additional "unit" of computer performance/capability meant a skyrocket in price back then. Also, in the very late 80's to early 90's - PC prices (and "units of performance") plummeted at an alarming rate, probably exponential in nature if it were graphed. So the scene in 1989 looked markedly different than 1988. The "clone" scene took off almost overnight, and that, in turn, lowered prices even more. Nobody could have seen this; it was completely unprecedented. Absolutely. The state of the personal computing had *never* changed so fast. Although it continues to change, it would be surprising to see anything as watershed as the clone movement. Just evolution.

Yes, economies of scale and increased competition will drive costs (and prices) down hugely on top of base hardware getting cheaper. You had the clone manufacturers competing and expansion of the components manufactuers as well along with off the shelf motherboards becoming available and soon following that: full arrays of off the shelf parts for the end user to build their own systems. (and used parts warehouses further expanding lower cost options -the larger the market, the more used parts as well ) You had Cyrix and AMD pushing for lower prices than Intel (and often higher performance, or higher performance upgrades for lower-end systems Intel ignored -ie 286, 386SX, 386SX, or 486DX accelerators, even ones catering to surface mounted chips -like the clip-on 486SLC upgrades)

 

The only way the ST and Amiga could have even remotely met that competition is with corresponding clones and licensed production. The Amiga would have necessitated licensing due to its extensive custom chips, but the ST really was imply enough to attract unlicensed clones like the Apple II, Spectrum, and PC had, and it did happen to an extent, but not enough to make the difference. (plus the lack of a standard expansion system like the Apple II and PC had limited that side of flexibility)

But in Europe they already had a massive presence to build onto, so unlike the US there was far more potential for continued competition regardless.

With clone/licensed manufacturing and gradual evolution of the chipset, economies of scale could have been significant but still fundamentally limited by off the shelf component costs, especially the CPU. The 68000 was extremely widely produced and licensed/second sourced, but even by the 68010 Motorola cut way back on how open they were with that, so following developments ended up slower and increasingly less cost/performance competitive with others on the market (be it x86 or emerging RISC architectures), but they could have gotten a lot of milage out of vanilla 68000s in any case and increasing use of the 32-bit chips as they got cheap enough to be reasonable. (if increased demand also didn't push for faster 68000s -even without licensing later designs the original licensees could build on the original chips and have faster and faster models if not more integrated features competing with Motorola's own chips -expanded address lines to full 32-bit, prefetch if not an actual cache, MMU functionality, integrated MMU, integrated FPU, 32-bit data bus and ALU, etc- faster 68ks alone would have gone a long way though, especially with the common DMA mechanisms allowing considerably slower memory than CPU speed without wait states -68k only needs the bus to be at about 1/3 its clock speed in such cases; and with the plain 68k you had a very small relatively low power core by contemporary late 80s/early 90s standards so the main issue would be boosting the clock speed with reliable yields and shrinking the die to reduce costs, power use, and heat -the Hitachi HD68HC000s were already known for their significantly lower heat and power dissipation than Motorola counterparts -not sure if they used a smaller die to do that or something else -they were both CMOS chips)

[svenski]

The Amiga had the edge graphics and sound wise from the start but history has shown us that the likes of Atari and Commodore were stuck between a rock and a hard place when it came to making changes to their established lines.

 

Atari had it with the 1200XL, Commodore had it initially with the A500+ and then the A600 and A1200. The STE was a I guess a compromise between trying to maintain a high degree of compatibility with the already huge user base and software library whilst moving forward.

 

Commodore's A1200 was a leap forward with its built in hard drive port or you could buy one with hard drive for around £399. I think externally the ST has aged better, the Amigas tend to show their age more.

 

Anyway, the ST was showing its age. The public knew the Amiga was better and when Atari cut the price to £199 it smacked of desperation. The STE extended the life of the platform, but not by a great deal. It was good while it lasted but once the general public knew just how much better the Amiga was, it was downhill from there.

[+wood_jl]

kool kitty89, I really enjoy your posts. They're quite informative. Such enthusiasm for this "sport" is entirely refreshing!

[OldAtarian]

The Amiga had the edge graphics and sound wise from the start but history has shown us that the likes of Atari and Commodore were stuck between a rock and a hard place when it came to making changes to their established lines.

 

Atari had it with the 1200XL, Commodore had it initially with the A500+ and then the A600 and A1200. The STE was a I guess a compromise between trying to maintain a high degree of compatibility with the already huge user base and software library whilst moving forward.

 

Commodore's A1200 was a leap forward with its built in hard drive port or you could buy one with hard drive for around £399. I think externally the ST has aged better, the Amigas tend to show their age more.

 

Anyway, the ST was showing its age. The public knew the Amiga was better and when Atari cut the price to £199 it smacked of desperation. The STE extended the life of the platform, but not by a great deal. It was good while it lasted but once the general public knew just how much better the Amiga was, it was downhill from there.

 

The open architecture of the PC with it's expansion slots was what really made the difference. If you weren't happy with the graphics and sound that came with your PC, you could upgrade them later simply by plugging in a card. With the Amiga and ST, what you got was pretty much what you were stuck with unless you were able to install some of the expensive third party mod kits that typically required fine soldering skills and that could really mess up your computer if you put a drop of solder in the wrong place and shorted something out and even if you could afford them and had the skills to do the installation, there wasn't much software support for them since only a small subset of users would have any of them. With the PC, you just installed the card and drivers and software instantly supported it.

[+wood_jl]

Heh heh. This reminds me of the frustration (in the DOS days) of games not finding the Soundblaster, etc. Remember starting a game and not having sound? Then you'd have to reboot to DOS and edit the autoexec.bat and/or config.sys files, etc. That was a rude awakening, after the days of the 16-bit home computers, where "everything just works."

 

Remember "dos=high, umb" and running Quarterdeck Expanded Memory Manager (QEMM) and all that stuff? Ah, those were the days. (cue up the theme to All In The Family.)

[svenski]

Heh heh. This reminds me of the frustration (in the DOS days) of games not finding the Soundblaster, etc. Remember starting a game and not having sound? Then you'd have to reboot to DOS and edit the autoexec.bat and/or config.sys files, etc. That was a rude awakening, after the days of the 16-bit home computers, where "everything just works."

 

Remember "dos=high, umb" and running Quarterdeck Expanded Memory Manager (QEMM) and all that stuff? Ah, those were the days. (cue up the theme to All In The Family.)

 

I remember the first PC that came into the store - had a price of around £1000 and this was when the Amiga and to a lesser extent the ST were still going strong. We used to get loads of hassle selling PC games as people never used to read the huge list of spec requirements and even when they did you always got cases where a game just wouldn't work on someone's PC despite it meeting the spec.

 

We used to spend hours with the PC playing Wing Commander and the LucasArts games - when the store was quiet (obviously ) .

 

Part of the fun was definitely the challenge involved in tweaking settings just to get a game to load.

[frogstar_robot]

 

The open architecture of the PC with it's expansion slots was what really made the difference. If you weren't happy with the graphics and sound that came with your PC, you could upgrade them later simply by plugging in a card. With the Amiga and ST, what you got was pretty much what you were stuck with unless you were able to install some of the expensive third party mod kits that typically required fine soldering skills and that could really mess up your computer if you put a drop of solder in the wrong place and shorted something out and even if you could afford them and had the skills to do the installation, there wasn't much software support for them since only a small subset of users would have any of them. With the PC, you just installed the card and drivers and software instantly supported it.

 

 

You really didn't see the full benefit of that until standard multimedia APIs became commonplace with Windows 95. Prior to that, games were implementing half the driver model themselves. Come to think of it, many gaming houses were buying a handful of standard sound programming frameworks. Many is the time I installed J. Random Game and heard that same little test ditty that was played first out one speaker then the other.

 

So yeah, the Gravis, Soundblaster, and Adlib were sorta semi standardized but it DOS gaming took a bit of dedication for a working setup. Heck, many games recommended you make a custom bootfloppy just to get the memory layout the way that particular game liked it. At one point, I even had menus in my autoexec.bat for doing that.

[kool kitty89]

The Amiga had the edge graphics and sound wise from the start but history has shown us that the likes of Atari and Commodore were stuck between a rock and a hard place when it came to making changes to their established lines.

 

Atari had it with the 1200XL, Commodore had it initially with the A500+ and then the A600 and A1200. The STE was a I guess a compromise between trying to maintain a high degree of compatibility with the already huge user base and software library whilst moving forward.

What do you mean on the 1200XL? The 1200 was one of the bigger screw-ups in the A8 line... the 800XL was more the direction it needed. (albeit too late and exacerbated by other delays, then the split and liquidation of Atari Inc)

 

I'm not sure about that comment on the STe... I don't think it was so much to maintain compatibility as simply being too modest of an upgrade too late and aimed in the wrong direction (it definitely looks like they were aiming at pushing against the Amiga rather than the ever emerging PC standards -by 1989 you had VGA becoming more common and getting support in most games -as well as a significant performance boost from the packed pixels, scrolling, and modest hardware acceleration). They had the BLITTER back in '87 but it was only on the MEGA and really was the wrong direction to go in: faster CPUs (could have been pushed on some models in '85), better sound (better synth -like the compatible YM2203- and PCM support), and focus on getting more modest hardware acceleration out even sooner and standardized (ie rather than the blitter, a SHIFTER II adding hardware scrolling and maybe even chained bitplane packed pixel support) and later adding 8-bit packed pixel modes with 256 indexed colors and at least 12-bit RGB (they didn't even need to bump all the way up to the TT SHIFTER, but something a bit less and sooner, like maybe just 640x200x16 color plus 320x200 256 color and maybe bumping 640x400 to 4 color and/or 4-shade grayscale).

It would have made things a lot easier if they'd used packed pixel from the start though. I think that was more a design choice than anything related to ease of development. (there were trade-offs, but without a blitter -and especially without hardware scrolling- packed pixels would have been a huge performance boost for games and some other graphics operations -monochome wouldn't be any different though) Hell, using packed pixels probably would have made it easier to add a 160x200 256 color mode (8bpp) with the original shifter. (that probably would have been very popular for games and some graphics programs) It would only need to be simple pixel accumulation with packed pixels used, so instead of 2 4-bit pixels it would be 1 8-bit pixel. (just as the A8 does for GTIA modes with 2-bit pixels accumulated to 4-bit)

 

But faster CPUs standard would have been the simplest and most foolproof standard upgrade later on. (something both Atari and CBM missed out on -at very least the MEGA and A2000 should have used 16/14.32 MHz CPUs and possibly offered optional FPUs)

 

Commodore's A1200 was a leap forward with its built in hard drive port or you could buy one with hard drive for around £399. I think externally the ST has aged better, the Amigas tend to show their age more.

The 520ST had the ACSI DMA port for Hard Drive back in 1985. (the HDD wasn't sold by Atari until 1986 though iirc, but demoed back in '85)

 

 

 

 

The open architecture of the PC with it's expansion slots was what really made the difference. If you weren't happy with the graphics and sound that came with your PC, you could upgrade them later simply by plugging in a card. With the Amiga and ST, what you got was pretty much what you were stuck with unless you were able to install some of the expensive third party mod kits that typically required fine soldering skills and that could really mess up your computer if you put a drop of solder in the wrong place and shorted something out and even if you could afford them and had the skills to do the installation, there wasn't much software support for them since only a small subset of users would have any of them. With the PC, you just installed the card and drivers and software instantly supported it.

Sort of, except it took a very long time for much of that to become affordable or available at all, and very rarely was it "plug and play" type stuff. I agree that expansion would have been important to include though, the ST should have had a general purpose expansion port rather than the bare bones cart slot and multiple slots in the box form factor models. The Amiga (sans the 500) DID have a very comprehensive expansion bus (mainly video on the A1000, far more on the A2000 including expansion boards adding PC ISA slots) and numerous 3rd party video boards, but that was all high-end stuff.

It took until the late 80s (~87) for EGA to become cheap enough to add to common PCs or include standard on new machines while sound remained lacking on most machines into the early 90s with the exception of the Tandy 1000 (and canceled PCJr) which had a sound chip a bit weaker than the ST (SN76489 like the TI99/4, CV, SMS, etc) years before any PC sound cards were available and later added an 8-bit DAC DMA channel up to 48 kHz. It took ages for any sound cards to appear at all, Adlib was the first in 1987 and it wasn't even until that year that high-end synth/midi modules appeared (like the IBM music feature card and MT-32) with nothing even on the level of the PCJr sound or Apple's Mockingboard being offered from 1981-1986 (Covox did release a simple resistor DAC using the parallel port in '86 -later becoming a common DIY sound option supported for PCM in some games -the Disney Sound source was a version of that as well, all very CPU intensive but far better than PWM using the PC speaker). In 1988 you saw the arrival of several other sound boards but none long lasting up to the Sound Blaster in '89 which roughly matched Adlib's price, was adlib compatible, added a built-in IBM gameport, and an 8-bit DAC DMA channel sampling up to 23 kHz (plus a rebranded Intel 8051 series MCU offering ADPCM decoding).

 

So you had EGA games supporting adlib appearing in 1988 and looking inferior to ST games but generally sounding better (with examples of similar quality effort), but soon after VGA came rushing in as a mass market standard and games quickly supporting it (unlike EGA which took years to reach that stage) with some games supporting VGA/MCGA even by 1988 and more comprehensive VGA 256 color support (actually taking advantage of the faster rendering with packed pixels and some hardware acceleration and scrolling) by 1989 and increasing rapidly from there on. And alongside adlib you had all the higher-end games also supporting MT32 and often some other options that would later fade. (so if you had the money and/or were getting the module for actual professional purposes you could get some awesome music/sound -same with several ST versions of those games except you could plug straight into the midi port) Though support for the MT32 started waning by the mid 90s. (general midi and some sample synth cards got more popular)

 

However, after that things stalled again, you got sound blaster support with the DAC used more and more from 1991 onward (and eventually cases of the Pro/16 specific features being used -more often for the DAC than the OPL3 and even some cases like Doom didn't really offer SB16 or pro specific support -no stereo no 16-bit PCM -not even sure if the SB2.0's 44 kHz 8-bit mono was supported or just the 23 kHz output of the 1.x, and no OPL3 support either in spite of full general midi and several other sample based cards supported -no MT32 either) and you had general midi and some other cards and the odd CD game using CD-DA or wav files, but VGA was the common standard for many years with SVGA specific stuff (640x480 or higher 256 color and highcolor) not supported until the mid 90s (you had several 1995 releases starting to use 640x480), you had a ton of games sticking to the plain limited 320x200 mode 13h VGA mode possible with MCGA and all 64kB VGA cards but some later games (prior to and alondside SVGA games) pushing mode X stuff offering 320x240 with square pixels and double (or technically even triple) buffering in VRAM (with 256k cards) or other various resolutions like 320x480, 400x300 (could be double buffered), 512x384, up to 640x400 (max possible in 256k). Double buffering in VRAM was important particularly with slow ISA (or even EISA) cards as you'd have trouble copying the entire next frame from main RAM in vblank alone and end up with tearing.

 

 

 

 

However, even after the fact, the ST still could have offered such upgrades as well with kits and clip on chips minimizing (or eliminating) any soldering needed. To avoid voiding warranties Atari could have offered upgrade services at licensed service centers (I'm not sure if they even offered the RAM upgrade that way or if that was all "unofficial" homebrew/hack stuff). That should have been very reasonable for most SHIFTER, BLiTTER (honestly the BLITTER shouldn't come in until after other SHIFTER upgrades for scrolling, packed pixel, and higher color), sound, and RAM upgrades, but CPU might be a bit tougher depending how you'd manage the different system clock. (I don't think a simple async double clock mode would work without timing issues unless the DMA manager already addressed that -on a bare bus connected to DRAM it would definitely be an issue, but the simple latch mechanism used for DMA might have solved or at least partially address the issue of higher CPU speeds, especially if the system allowed fast page mode accesses -with pure random accesses I think you'd get wait states beyond ~12 MHz on the 8 MHz ST bus) Not an issue with full new machines though, just modify the system to run at the higher clock rate and allow the CPU to make fast page accesses. (note that that would no longer make interleaved DMA efficient as in 8 MHz and serial bus accesses become more efficient)

 

 

 

 

 

 

You really didn't see the full benefit of that until standard multimedia APIs became commonplace with Windows 95. Prior to that, games were implementing half the driver model themselves. Come to think of it, many gaming houses were buying a handful of standard sound programming frameworks. Many is the time I installed J. Random Game and heard that same little test ditty that was played first out one speaker then the other.

That was one of the issues with the BLITTER in the ST as well, developers taking less interest in rewriting engines for both software renderers and the blitter (more so due to the late entry on standard units -or rather they never did that as the vanilla ST still existed). The same happened with PC: you didn't see significant use of 2D or 3D acceleration until after APIs became common in the mid 90s. Prior to that you saw 2D acceleration mainly used by GUIs and some applications made specifically for windows.

 

So yeah, the Gravis, Soundblaster, and Adlib were sorta semi standardized but it DOS gaming took a bit of dedication for a working setup. Heck, many games recommended you make a custom bootfloppy just to get the memory layout the way that particular game liked it. At one point, I even had menus in my autoexec.bat for doing that.

Gravis US wasn't that big of an overall standard either, it was there for sure but sort of a flash between SB/Adlib+MT32 and general midi becoming the overwhelming onboard synth standard with GM appearing alongside (if not before) the GUS in games and persisting to today (having windows supported soundcanvas based GM software drivers was part of that) and of course streaming audio appearing about the same time. (and you had some custom sample based MIDI drivers as well)

 

And of course you had MOD players using SB or various other cards for the DMA PCM channels alone.

Edited November 23, 2010 by kool kitty89

[svenski]

What do you mean on the 1200XL? The 1200 was one of the bigger screw-ups in the A8 line... the 800XL was more the direction it needed. (albeit too late and exacerbated by other delays, then the split and liquidation of Atari Inc)

 

The main reason the 1200XL failed was it represented a significant shift away from the 800 and offered less possibilities when it came to upgrading. The 800XL was/is an improvement over the 1200XL but *if* the 800XL had come first I'm guessing it would have gotten a lot of flak too.

 

 

The 520ST had the ACSI DMA port for Hard Drive back in 1985. (the HDD wasn't sold by Atari until 1986 though iirc, but demoed back in '85)

 

I'm aware of the ACSI port however the Amiga 1200 and it's ATA connector made adding a hard drive quick and easy and at a much lower cost. The 1200 also had PCMCIA. Sadly Commodore went belly up before the A1200 could shine.

Edited November 24, 2010 by svenski

[Christos]

 

 

The 520ST had the ACSI DMA port for Hard Drive back in 1985. (the HDD wasn't sold by Atari until 1986 though iirc, but demoed back in '85)

 

I'm aware of the ACSI port however the Amiga 1200 and it's ATA connector made adding a hard drive quick and easy and at a much lower cost. The 1200 also had PCMCIA. Sadly Commodore went belly up before the A1200 could shine.

 

Amiga 1200 came in 1992. The falcon also had internal ide, while before it the MSTE and TT had internal scsi. Your point is kind of mute.

 

Oh, and 20MB of 2.5" ide used to cost about as much as 100MB of 3.5" IDE.

[svenski]

 

 

The 520ST had the ACSI DMA port for Hard Drive back in 1985. (the HDD wasn't sold by Atari until 1986 though iirc, but demoed back in '85)

 

I'm aware of the ACSI port however the Amiga 1200 and it's ATA connector made adding a hard drive quick and easy and at a much lower cost. The 1200 also had PCMCIA. Sadly Commodore went belly up before the A1200 could shine.

 

Amiga 1200 came in 1992. The falcon also had internal ide, while before it the MSTE and TT had internal scsi. Your point is kind of mute.

 

Oh, and 20MB of 2.5" ide used to cost about as much as 100MB of 3.5" IDE.

 

I don't need you to tell me the Mega ST had a built in hard drive, I'd used a Mega ST4 with Calamus DTP well before I got my A1200 ... And there had been hard drives available for the Atari 8-bit for years lol BUT my point was the 1200 was designed for the mass market, unlike the Mega ST or the TT, and for £399 you got your 1200 with hard drive. Heck even the A600 had an option for an internal HD. Yes the 600 and 1200 came after the STE but Atari had pretty much given up by then so yes, you are right, my point is mute.

[The_Laird]

That would be MOOT not mute

[svenski]

That would be MOOT not mute

 

lol - you got me

[The_Laird]

Its ok you were not the first person to use the incorrect term, I guess you just copied it

[kool kitty89]

 

 

The 520ST had the ACSI DMA port for Hard Drive back in 1985. (the HDD wasn't sold by Atari until 1986 though iirc, but demoed back in '85)

 

I'm aware of the ACSI port however the Amiga 1200 and it's ATA connector made adding a hard drive quick and easy and at a much lower cost. The 1200 also had PCMCIA. Sadly Commodore went belly up before the A1200 could shine.

 

Amiga 1200 came in 1992. The falcon also had internal ide, while before it the MSTE and TT had internal scsi. Your point is kind of mute.

 

Oh, and 20MB of 2.5" ide used to cost about as much as 100MB of 3.5" IDE.

 

I don't need you to tell me the Mega ST had a built in hard drive, I'd used a Mega ST4 with Calamus DTP well before I got my A1200 ... And there had been hard drives available for the Atari 8-bit for years lol BUT my point was the 1200 was designed for the mass market, unlike the Mega ST or the TT, and for £399 you got your 1200 with hard drive. Heck even the A600 had an option for an internal HD. Yes the 600 and 1200 came after the STE but Atari had pretty much given up by then so yes, you are right, my point is mute.

Atari was still very active up into the early 90s, but remember it was the European market that really mattered, not the US.

[+wood_jl]

You guys are saying the Mega ST had an internal hard drive? You're not talking about the Mega STe???? I have never seen either, 'cept for pictures. I did not think the Mega ST had internal hard drive, but what do I know? <very little>

[svenski]

You guys are saying the Mega ST had an internal hard drive? You're not talking about the Mega STe???? I have never seen either, 'cept for pictures. I did not think the Mega ST had internal hard drive, but what do I know? <very little>

 

 

Well I can't remember if it was a MEGA STE or not , but it would have been 1989 as I left the printing company and Spain to return to the UK in 1990. I think the hard drives were optional? I do remember the Calamus DTP package and the Atari Laser printer was amazing stuff. The print quality was brilliant and we could go straight to print plates. I hate to think what my boss spent on the Atari setup as Atari Spain weren't known for being exactly cheap.

[kool kitty89]

What do you mean on the 1200XL? The 1200 was one of the bigger screw-ups in the A8 line... the 800XL was more the direction it needed. (albeit too late and exacerbated by other delays, then the split and liquidation of Atari Inc)

 

The main reason the 1200XL failed was it represented a significant shift away from the 800 and offered less possibilities when it came to upgrading. The 800XL was/is an improvement over the 1200XL but *if* the 800XL had come first I'm guessing it would have gotten a lot of flak too.

The problem was the same as it had been from the start and twofold: 1 not aiming enough at the cost-cut low-end market (no replacement for the 400 -ie no 600 in '82) and lack of expandability (and technically better expandability in the A8 due to the CPU card), the PBI and 1090XL expansion system would have been critical differences (the other changes for added consolidation were indeed not possible until later).

The OS compatibility issues were there and if they really were too troublesome they could have waited for a later revision to add that even if it meant cutting to 48k (which could also have meant reduced cost and removal of the MMU). Given the amount of software speficically catering to 48k and less that wouldn't have been a huge issue, especially with a general purpose expansion port and socketed OS ROMs.

In the case of the 1200XL, the side position of the cart slot was perfect for the PBI for a side mounted expansion unit (much more convenient than rear mounted), or for that matter they could have used the ECI set-up of the XE machines for dual purpose.

 

They had 2 directions to go in with the lower end/mass market C64/VIC/TI99/CoCo etc (and universally beating out all but the C64) and pushing against Apple and higher end systems. Tactfully they could have done both with different ranges of the same line of machines (let alone the 68k machines -granted rather excessive in the high-end workstation layouts the prototypes showed, but probably lots of potential to be cut down to a more high/mid-range machine).

 

The European market was separate and they lost out for different reasons including not studying the market needs. (emphasis on low cost but significant amounts of RAM, tape drives as main format, lack of strict RFI regulations, emphasis on 3rd party software development and home programming support) They could have started pushing that way from the beginning with the simplest steps being to remove all aluminum castings and replace them with simple support struts and rather than the 800, release the normal 400 and mid-range models adding a full keyboard and monitor port. (including 32k and 48k models to compete more against the higher-end BBC Micro initially and 16k providing a middle ground between the super low-end ZX80/81, VIC-20, and higher end BBC Micro then shifting to 600/1200 derivatives with no shielding and perhaps a bit more compact -the 600 form factor was very nice and a plain 48k and 32k machines using the older OS and no MMU in the 600 form factor could have been very significant for 1982-84 with an existing software base to build on giving an edge over the C64 even at similar prices -and note that against the Speccy, a 32k A8 would be closer in graphics space to a 48k spectrum due to the hardware character modes vs a hefty chunk for the framebuffer in the Speccy -15 kB with double buffering- let alone the generally more capable graphics -with the exception of ease of programming in some case -the latter more dependent on whether 3rd party programming was better supported or not, multiplatform A8/Speccy games probably would have used the highres character modes on the A8)

 

 

 

 

You guys are saying the Mega ST had an internal hard drive? You're not talking about the Mega STe???? I have never seen either, 'cept for pictures. I did not think the Mega ST had internal hard drive, but what do I know? <very little>

 

 

Well I can't remember if it was a MEGA STE or not , but it would have been 1989 as I left the printing company and Spain to return to the UK in 1990. I think the hard drives were optional? I do remember the Calamus DTP package and the Atari Laser printer was amazing stuff. The print quality was brilliant and we could go straight to print plates. I hate to think what my boss spent on the Atari setup as Atari Spain weren't known for being exactly cheap.

MEGA STE wasn't until 1991... unfortunately. (would have been great launched alongside the STe in 1989m though a lower end and mid-range model including the TT shifter would have been really nice as well -ie STe/MEGA STe in 1990 with TT SHIFTER) Though some of that would still be a bit late and not pushing in the ideal direction. (TT SHIFTER or something closer to VGA with packed pixels and scrolling before the blitter and sound upgraded before the blitter and faster CPUs from very early on)

Edited November 24, 2010 by kool kitty89

[svenski]

The 800XL was perfectly positioned to go after the C64 but I think Atari UK were more to blame than anything else as to why they missed the boat in the UK. The 800XL was voted computer of the year and at one point Atari UK claimed they were flying the things in by the plane load as demand was so high and they couldn't wait for the slow boat from China.

 

The XL being delayed, the takeover by the Tramiels, the ST in the wings etc all were factors although initially I'm sure I read somewhere that even Jack Tramiel had at first had big plans for the 800XL.

 

Also you have to be careful not to dismiss the European model as low cost and penny strapped. Atari had sold quite a lot of Atari 400s and 800s in the UK at least - people were spending hundreds just so they could play Star Raiders.

 

As for the tape vs disk model, again there were plenty of Atari owners in the UK who had disk drives. Part of the problem was, as I've said before, Atari not keeping up with demand and the big gap between the 1050 being discontinued and the XF551 being launched. Atari 8-bit users were then for the most part, as they are now, fiercely loyal and devoted to their Atari and whilst they could have jumped ship at any time most stayed with it and spent their cash accordingly.

 

More importantly, and this isn't just Atari related, the tape model was hugely significant as it was a low cost medium and because of it the UK became a game producing powerhouse and we had the budget game pricing model that saw the likes of Mastertronic etc make huge sums of money whilst keeping us well fed with constant new releases across the different platforms.

 

It didn't matter if you had a disk drive or not, when you could buy a great game for £1.99 you'd always be reaching for your tape unit.

[+DarkLord]

You guys are saying the Mega ST had an internal hard drive? You're not talking about the Mega STe???? I have never seen either, 'cept for pictures. I did not think the Mega ST had internal hard drive, but what do I know? <very little>

 

 

Well I can't remember if it was a MEGA STE or not , but it would have been 1989 as I left the printing company and Spain to return to the UK in 1990. I think the hard drives were optional? I do remember the Calamus DTP package and the Atari Laser printer was amazing stuff. The print quality was brilliant and we could go straight to print plates. I hate to think what my boss spent on the Atari setup as Atari Spain weren't known for being exactly cheap.

 

The Mega ST, Atari STacy, the Mega STe, TT030, and Falcon 030 all had internal hard drive options.

Although with the earlier Mega ST, it wasn't quite as "plug 'n play" as the 4 later machines.

 

Just to clear things up folks.

[Crazyace]

I think upgrading the ST is a mistake - the ST was a fixed platform and any improvements should have been in the original model. Otherwise software would just ignore it ( look at how most games were on single sided drives )

That's why I push scrolling - it gives the best 'bang for the buck' and really doesn't involve very much change to shifter at all. Making the cpu speed 10MHz rather than 8MHz would be another possible for the launch.

Then in 1990 release a replacement machine ( cut down TT? ) with TT style graphics + VGA support, maybe with 16/20Mhz 68030, as a low cost ( rather than high end ) machine, emphasising compatibility with all ST software - so people upgrade to that machine. ( or even 2 models - low end 16/20Mhz 68000 , high end 33-40Mhz 68030+expansion )

[kool kitty89]

I think upgrading the ST is a mistake - the ST was a fixed platform and any improvements should have been in the original model. Otherwise software would just ignore it ( look at how most games were on single sided drives )

That's why I push scrolling - it gives the best 'bang for the buck' and really doesn't involve very much change to shifter at all. Making the cpu speed 10MHz rather than 8MHz would be another possible for the launch.

Then in 1990 release a replacement machine ( cut down TT? ) with TT style graphics + VGA support, maybe with 16/20Mhz 68030, as a low cost ( rather than high end ) machine, emphasising compatibility with all ST software - so people upgrade to that machine. ( or even 2 models - low end 16/20Mhz 68000 , high end 33-40Mhz 68030+expansion )

As I said though, there were practical ways to break the closed box design even after the fact, if Atari Corp was willing to do so. (simplified modules with minimum soldering to clip onto the board and installed at Atari Service centers) But otherwise it depends on how common the feature becomes and how easy it is to use/add to a game. (if it's a standard across the board feature and a significant advantage in several aspects, that's important)

 

But yes, scrolling would have been a big thing early on, but might have compromised their tight release date... though I think packed pixels vs planar was probably more a general engineering decision than time related. (in the lack of a blitter especially, Packed Pixels made a lot more sense and probably would have made a 160x200 8bpp 256 color mode more practical as well via linear pixel accumulation like GTIA used) The 9-bit RGB palette was OK for the time (the MD and PCE pushed that years later and into the early/mid 90s). And they could have tweaked audio a bit too: there was the YM2203 in 1985 already, but it might have been pricier than what they were willing to spend at the time (would have extended the life a fair bit if they had though) and on the in-house side some simple DAC/PCM mechanism would have been really nice: preferably a good bit more than just a bare 8-bit DAC (though that would at least have advantages over hacking though the YM) and probably more than a small resistor array as well, perhaps a very basic DMA audio set-up like the early Macs had, or conversely a small IC with embedded DAC and simple FIFO buffer for loading PCM maybe even mixing multiple channels to a single 8-10 bit DAC and perhaps some rudimentary volume control as well, plus a timer input to set the sample rate and interrupt generation along with full/empty flags for the buffer. (either polling full/empty flags or triggering loading with interrupts, but cutting down resource significantly compared to loading every single sample manually with the CPU -even using the stack pseudo DMA trick it was heavy on the 68k -though on CPUs like 650x you could get away with interrupt driven PCM due to the very fast interrupt handlers: hence the PCE managing a 7 kHz channel with 5% CPU resource of a 7.16 MHz 65C02)

Such changes would also have made it a far more attractive game console design too. (in that case you could have more options for tweaks regardless -there was the doubled SHIFTER hack you suggested for dual graphics layers, and definitely more options for off the shelf sound chips)

 

The more custom hardware used (no matter how rudimentary), the more that could be consolidated later on to save space. (that could include some hacked discrete component stuff early on that was later consolidated -perhaps rather quickly- which may have been the case for something like the DAC/PCM circuit above)

 

The jump in design for a more VGA competitive machine in '89/90 would be nice too, but I think from the beginning (sans other upgrades) they should have offered faster CPU models (12 MHz if 16 MHz wasn't readily available in '85) thus developers could avoid timing sensitive routines and allow games to take advantage of the higher CPU speeds when available. In fact, even without a video upgrade a packed pixel display with 9-bit RGB and 320x200x4-bit (16 color indexed) and 160x200x8-bit (semi-indexed or direct RGB) along with hardware scrolling would already put it close to competitive with VGA and much more competitive with the Amiga. (still lacking the added color capabilities of VGA and possibly some of the added acceleration -iirc VGA added simple copy/line fill operations and such to aid software blits, unless the original ST added that as well)

 

But yeah, a proper upgrade with a full successor would be nice, with everything in-between mainly being higher-end models (MEGA and such) and gradual RAM expansion and CPU speed boosts. (probably push faster CPUs as standard than higher RAM capacity -especially given what happened with RAM prices) If they were pushing faster CPUs from the start, probably design the DMA controller to allow fast page accesses from the CPU as well. (unless that's not an issue as it is)

 

For '89/90 something between a STe, TT, and MEGA STe would be nice for sure, but a lot of different routes to potentially take.

The DSP of the Falcon was an interesting addition, and especially could have meant some neat possibilities for 3D if pushed for such. (vertex calculations, aid in rasterization, or for pseudo 3D with scaling effects, ray casting, or even affine texture rendering) Was the 56k a significantly cheaper component than a 68881/2? (it seems like a far smaller chip -and package- and for general use could offer a good deal of acceleration that an FPU wouldn't be that useful for -FPU could be more useful as an option and/or for higher end machines specifically and obviously for various workstation uses -which could lead back to the mid 80s MEGA models probably meriting FPU sockets)

Edited November 24, 2010 by kool kitty89

[Christos]

 

 

For '89/90 something between a STe, TT, and MEGA STe would be nice for sure, but a lot of different routes to potentially take.

The DSP of the Falcon was an interesting addition, and especially could have meant some neat possibilities for 3D if pushed for such. (vertex calculations, aid in rasterization, or for pseudo 3D with scaling effects, ray casting, or even affine texture rendering) Was the 56k a significantly cheaper component than a 68881/2? (it seems like a far smaller chip -and package- and for general use could offer a good deal of acceleration that an FPU wouldn't be that useful for -FPU could be more useful as an option and/or for higher end machines specifically and obviously for various workstation uses -which could lead back to the mid 80s MEGA models probably meriting FPU sockets)

 

Both the MSTE and Falcon have an option for an FPU. It's standard on the TT.