What is the 7800 really capable of?

[atarian63]

The tramiels started of good but really became a collection of horses asses

the way they handled many things. The ST is a perfect example. Great

start to a machine already more capable than a PC and a third the price

WITH 1 meg and a color screen. Then they completely dropped the ball after

that. No support, thinking the word of mouth nonsense would do the work

for them. I think they have plenty to be blamed for.

 

Wake up. IBM-compatable PC's had one major aspect that lent itself to getting a foothold across the entire consumer market: Open-ended architecture. No single company could compete with that. Apple tried, so did Commodore. Both of them were way ahead of Atari for product support.

 

Passing blame for their financial woes to their own user base didn't help Johnny-come-lately any.

We were a dealer for both Atari and Commodore at the same time. 87-93. Atari support wasn't great but it sure beat the crappy attitude and lack of support at least to the dealer from commodore.

[supercat]

It's also capable of displaying high res bitmap graphics with restrictions.

 

With a few slight changes, the 7800 would have been a lot more powerful. Some things I would have liked to have seen:

 

-1- Selectable 228 vs 227.5 chroma clocks per line, and selectable 262/262.5/263 lines per frame. Different combinations have their own advantages and disadvantages.

 

-2- Change the write-mode-2 transparency logic so that writing a pixel value of xyy00 would not be transparent unless yy were both zero.

 

-3- Add a read mode which would use the upper bit of each pixel to select between a 12-color lo-res mode and a 3- or 6-color high-res mode (the 6-color mode would be like 3-color mode, but with separate color registers for the left and right halves of each pixel)

 

-4- Change the display-list logic so that a "width" value over 24 bytes would trigger a 5-byte display list entry (instead of requiring a value of 32). That would add an extra 3 usable bits to 5-byte display list entry; those could be used to allow Kangaroo and Holey modes to be switched on and off on a per-entry basis.

 

-5- Delay processing of palette writes by an 8Mhz pixel clock, to avoid display artifacts that occur when writing them even in low-resolution mode.

 

-6- Make address $3E of MARIA always read $4C and address $3F read the scan line counter, AND'ed with $FC. This would allow one to store at RAM address $40 the MSB of the address of a jump table and point the NMI vector at address $003E. This would allow a rapid dispatch of display-list interrupts.

 

None of those changes would be major, but they would have been quite handy.

[Inky]

None of those changes would be major, but they would have been quite handy.

 

 

The problem I have with these type of discussions is that the tech stuff "sounds cool," but what does it mean in layman's terms? What would have joe gamer experienced?

[supercat]

The problem I have with these type of discussions is that the tech stuff "sounds cool," but what does it mean in layman's terms? What would have joe gamer experienced?

 

Using 227.5x263 chroma clocks/frame would have improved the appearance of colors on the screen and reduced some types of color smearing. Some of the other changes would have made it practical more games to use high resolution mode for at least some of their graphics. Note that transparency effects would all be based upon the 160-wide pixels, but objects' interiors could be drawn in 320 mode. For example, the external contours of a character's face would have to be drawn using low-resolution pixels, but the interior portion could be drawn with twice that resolution.

 

Switching Kangaroo mode within a display list would allow games to have some objects drawn with an opaque rectangular background while others are drawn transparent. Delaying the effect of writes to palette registers would avoid some slight screen distortions that can occur if games change palette registers during a frame (so as to display more colors simultaneously).

[Gorf]

... games change palette registers during a frame (so as to display more colors simultaneously).

 

 

Changing colors on a frame? Do you mean changing colors on a particular raster?

[Gorf]

The tramiels started of good but really became a collection of horses asses

the way they handled many things. The ST is a perfect example. Great

start to a machine already more capable than a PC and a third the price

WITH 1 meg and a color screen. Then they completely dropped the ball after

that. No support, thinking the word of mouth nonsense would do the work

for them. I think they have plenty to be blamed for.

 

Wake up. IBM-compatable PC's had one major aspect that lent itself to getting a foothold across the entire consumer market: Open-ended architecture. No single company could compete with that. Apple tried, so did Commodore. Both of them were way ahead of Atari for product support.

 

Passing blame for their financial woes to their own user base didn't help Johnny-come-lately any.

 

 

I assure you that I am wide awake. I am also convinvced that Atari's fall was based on seriously bad decisions.

Marketing word of mouth for instance. That might be fine for a start-up but not for a large company like Atari.

Open ended was a big advantage but not one any of the other three you mentioned could not come up with.

 

In the PC you HAD to have open ended architechture as it was the only way to add all the features you already

had inside an ST, Amiga or an Apple. The Tramiels are hated at just about every mom and pop shop and especially

at music stores all throughout New Jersey. Why? Because they too feel like me, that Atari could have been dominant

if they only stood behind their products. The open ended thing was not the appeal to most casual users and only us

real power users gave a crap about. I cant tell you how many shop owners and store managers told me that Atari

would have ruled the roost if they we better at support and really had little to do with open ended achitecture.

 

The St was a very complete computer out of the box at under $1000. A $3000 PC was pretty useless out of the box.

Truth is, software and support make or break a machine. If not the 2600 should have been pounded by the Astrocade,

which was by far a truly open ended video gaming console.

 

Perception beats out hardware features every time.

Edited November 2, 2008 by Gorf

[Atarifever]

None of those changes would be major, but they would have been quite handy.

 

 

The problem I have with these type of discussions is that the tech stuff "sounds cool," but what does it mean in layman's terms? What would have joe gamer experienced?

The problem I have with these discussions regarding the 7800 is that none of it would have mattered. The system was barely ever even close to pushed with what it could do. A few added features wouldn't have changed that. It would have just been more stuff not to be used, or stuff to make it even easier to make a mediocre game with even less effort.

[Crazyace]

Yes, the nice thing with demos is to show what could have been...

[DracIsBack]

Yes, the nice thing with demos is to show what could have been...

 

That's one thing I've not see a lot of on the 7800 relative to computers like the Atari 8-bit and Commodore 64. There are cool demos (non-playable) that push those systems graphically and sonically. On the 7800, there's only a couple.

[Crazyace]

A lot more people program on the computers ( compared to any console ) I guess...

 

Even the 2600 had the supercharger to allow easy programming - On early consoles making eproms is a bit more of a hassle.

[shadow460]

... the 2600 should have been pounded by the Astrocade, ...

 

ROFL

 

I needed a good laugh.

[CV Gus]

Oh, I don't know `bout that- the NES did pound the superior SMS in sales, didn't it? Don't forget how much effort into expanding the 2600's abilities there was, but the other systems from that time? Not much.

 

But the X-Box 360 under the control of the Tramiels likely couldn't have beaten the RCA Studio 2. So what chance did the 7800 have, even if (this occurred to me years later) the whole 1983 7800 idea wasn't doomed from day one?

[Tanegashima]

Hardware is meaningless sometimes...look at the 3DO. It technically outclassed ALL of it's 16-bit rivals in 1993 but the software support for the system was a joke. Besides the excellent Amiga ports, 80% of the software was low budget junk. The Neo Geo is another example, despite SNK's 24-bit claims (it really is a supped up 16-bit system) the system still was technically far superior to both Nintendo and Sega's machines at the time. Yet it's extreme price point, expensive software and scant releases kept it from ever being anything more than something for those who could afford it. It was supported for a long long time and had a following but probably no more than 10% of the entire market share, if even that. I guess in many ways it really never was competing at all.

 

Hell, look at Sega and Nintendo's later consoles. Neither the Saturn nor the N64 reached the hardware or software sales the Genesis and SNES did.

 

Proof that hardware isn't always where it's at, software is far more important. Good marketing helps too.

Edited November 4, 2008 by Tanegashima

[nathanallan]

When I was a kid and the Nin came out(okay, old kid), all I remember all over the place was Nintendo this and Nintendo that. Marketing is where Nin had it. I can barely remember seeing any Atari/Neo Geo/3D0 commercials. Nintendo poured it on thick, though. So, agreed.

Edited November 4, 2008 by nathanallan

[Alison DeMeyer]

What is the 7800 really capable of?

 

Playing video games, no better or no worse than all the others out there.

[Crazyace]

Oh, I don't know `bout that- the NES did pound the superior SMS in sales, didn't it? Don't forget how much effort into expanding the 2600's abilities there was, but the other systems from that time? Not much.

 

But the X-Box 360 under the control of the Tramiels likely couldn't have beaten the RCA Studio 2. So what chance did the 7800 have, even if (this occurred to me years later) the whole 1983 7800 idea wasn't doomed from day one?

 

The most brilliant thing about the NES was the PPU bus on the cartridge - That allowed so many cool h/w tricks to be implemented in the mapper chips Although the SMS had better graphics it was bottlenecked in some ways by the fixed VRAM.

If the 7800 had been released in 1983 with a 'big launch' it would have dominated - the software launch was a lot better than the 5200 - and the H/W was cheap to make - and more powerful than the competition ( It would have been interesting to see a 'computer' version to compete against the C64 )

[malducci]

Yes, the nice thing with demos is to show what could have been...

 

Well... not really - if you're using the term demo in the way I think you are. Demos don't really show practical effects in general. A lot of demos exploit *extreme* situations that are just not practical for real use - huge precalculated LUTs and animations, near maxed out cpu resource intensive code, the 'infinite' sprites onscreen effects, etc. Demo coders are more like illusionists. A lot of Amiga demos are misleading in what people think the system is actually capable of (talking about the 7mhz A500 series, not those 030's with AGA chips, etc).

Edited November 7, 2008 by malducci

[Crazyace]

All game coders are like illusionists

 

I find one of the things about demos is that way more time is put into a small bit of code than most commercial coders have the luxury for. But the results are there - and sometimes , an effect that appears as a demo first may make it into a game somewhere. Othertimes 'demo' effects are produced in house, and games can be made around them ( Rescue on Fractalus )

 

Anyway, an extreme situation is still using the H/W - ( isn't scrolling on CV where theres no hw scroll just one extreme situation )

 

For the 7800 one thing is nice that memory isn't limited - so it would be practical to put in 128 or 256K of rom if it were needed to get some amazing effect...

 

( As one point - it would be easy to have a game like Xmen vs SF on the 7800 - as the h/w could support massive sprites - the only limitation would be memory )

[Classic Pac]

The truth is we don't know. For one Atari never published a game that put the system to a real test.

[Crazyace]

If you know the hardware - then you know the limits quite well.

[jbanes]

Who blames the Tramiels for the crash (1983 btw)?

I do!

 

Well, not so much the entire situation as much as lighting off the tinderbox. Things were already precarious with an oversupply of games on the market. But when Tramiel dropped the price on the Commodore 64, he lit off a Roman Candle that had every retail store practically giving away its console games so they could make room for inexpensive (and high margin!) computers. He freaked out the Commodore board so badly that they fired him!

 

Gotta love ol' Jack.

 

As for the NES vs. 7800 comparison, both Supercat and Drac are on the money. Supercat pointed out several aspects that would have helped and Drac's point that the 7800 was never actually pushed is correct. However, there is also the aspect of looking at the machine from a programmer's perspective. One of the reasons why the machine was never pushed is because the design was not as practical as the NES's design.

 

On paper, the 7800 looks nothing short of incredible. You can push massive numbers of sprites at high colors and resolutions. By any stretch of the imagination, it should be able to outperform every console of its day.

 

But if you actually sit down and attempt to program it, that's where frustration sets in. It's not that those paper promises are impossible. (Though they are a bit overstated.) It's just that the architecture is so convoluted. Just to display a sprite on the screen requires a Masters Degree in Computer Science with your thesis done on Advanced Data Structures! i.e. The Display List List which ran down every scan line to run the Display List for that line. That sounds like a cool solution until you realize that simple programs like moving sprites vertically are really tough. In result, how much power can be squeezed out of the system depends on how clever you are with manipulating data structures in a limited number of processor cycles.

 

But don't worry! It gets worse.

 

Accessing the 2600 bus significantly drops the performance of the CPU. Which shouldn't be a problem, save for the cheapness of the system's design. Since there is no POKEY chip, I/O routines like joysticks and sound have to be handled through the TIA/RIOT chip. Which means that your precious cycles intended for data structure manipulation are getting lost to odd downshifts in processor speed. Furthermore, making use of the DMA controller results in the processor being halted while the DMA chip does its work. This makes the DMA transfer only useful if you have a significant task to offload to it.

 

Broadly speaking, all these issues can be worked around. But they create a huge barrier of entry, especially for many game programmers who were not former CompSci students. They had to deal with a rather steep learning curve that Tramiel just wasn't willing to pay for.

 

The NES was far less powerful on paper, but its internal design was much simpler and easier to use. Backgrounds were structured around character-mode tiles, scrolling support was more or less built into the PPU, the designers took flicker as a design-necessity into account, and sprite handling was straight forward task despite the system's limitations. The end result was that the NES was pretty easy to program for. Once there was a strong base of programmers, Nintendo was able to push out various hardware expansions that could be included in cartridges to make the system perform even better.

 

Thus the 7800 may technically be more powerful, but it's the NES that was practically more powerful.

 

Unfortunately, Atari would make the exact same mistake of technology over practicality with their next console: The Jaguar

[Crazyace]

It's easy to program for, especially if you've just finished working on some 2600 games

[jbanes]

It's easy to program for, especially if you've just finished working on some 2600 games

Not sure what gives you that idea. The 2600 isn't that hard to program for if you're making straight-up use of its facilities. It's only when you push it beyond its designed facilities that things get hard.

 

The 7800 on the other hand? Dear God in heaven. Amazingly convoluted. You can get used to it, but that doesn't mean there isn't a massive learning curve. Especially if you're used to controlling sprites on the 2600.

[malducci]

Othertimes 'demo' effects are produced in house, and games can be made around them ( Rescue on Fractalus )

 

Actually - I'm not a fan of that method. I've seen plenty games designed that way that are pretty shallow. I'll take a game designed from the ground up as a game first, and with high production value, over any tech/demo game.

[Zap!]

Hardware is meaningless sometimes...look at the 3DO. It technically outclassed ALL of it's 16-bit rivals in 1993 but the software support for the system was a joke. Besides the excellent Amiga ports, 80% of the software was low budget junk. The Neo Geo is another example, despite SNK's 24-bit claims (it really is a supped up 16-bit system) the system still was technically far superior to both Nintendo and Sega's machines at the time. Yet it's extreme price point, expensive software and scant releases kept it from ever being anything more than something for those who could afford it. It was supported for a long long time and had a following but probably no more than 10% of the entire market share, if even that. I guess in many ways it really never was competing at all.

 

Hell, look at Sega and Nintendo's later consoles. Neither the Saturn nor the N64 reached the hardware or software sales the Genesis and SNES did.

 

Proof that hardware isn't always where it's at, software is far more important. Good marketing helps too.

 

The Neo-Geo is a different case. It was never meant to compete with Nintendo or Sega, nor was it meant to sell millions. It was like a high-class car. Rolls Royce and Bently know they won't sell more that a Maxima or Explorer. And that's fine by them. Neo-Geo outlived all its competitors.