Consoles that provide(d) the best performance increase from 1stgen to last gen titles

Not at all... just saying that console's are cheap and have a long life cycle. They sometimes need to incorporate more complex architectures to keep competitive later in the life cycle.

It's not complexity that you want, it's flexibility. IMO that's what the problem was with the DC and (to an extent the GCN). They exhibited the same characteristics that DX7 and earlier PC game development did, and that is you were basically molding your design around a bunch of fixed function principles ordained by the IHV (and/or ISV; API constraints. Although this is less of a problem on consoles). Visually, that lead to rather cookie cutter'ish looking games, vs. the more (IMO) creative implementations of the software renderer days. Also, you'd find yourself penalized rather significantly performance wise when you wandered off the optimal path of the hardware. So while they offered easy approachability and granted significant access to the upper echelons of their performance envelope, they also (in various degrees) created several boundaries to creative solutions to work around their limitations.

Where the PS2 has somewhat evaded this, thus extending is life span is that it's design has encouraged a fairly significant amount of creative solutions to what would be deemed at first glance as a significant feature (or lack thereof) roadblock. People looked at the EE and immediately commented how complex and difficult it looked (to be fair, the early devkit programming examples IMO led to some misconceptions on what you could and couldn't do). In reality it's not a terribly complex chip. It's really just a large pile of ALUs, registers and local stores, that places a lot of onus on the programmer to creatively leverage.

The GS has been an even more misunderstood part than the EE. At first sight the most common impression would be that of "ZOMG!!! It's just a a goddamn overclocked Voodoo w/almost no video memory!!!" But really leveraging the GS meant unlearning a lot of preconceived notions on what a GPU is and how best to get performance from it (i.e. *many* practices you learn to avoid doing for performance reasons on most GPUs the has GS gleefully encouraged). Instead of looking at it as a GPU (e.g. an overclocked Voodoo), you'd be better off thinking of is as a simple old 8-bit accumulator CPU (e.g. an overclocked 6502). The frame buffer starts looking more like a massive array of 8-bit registers, and that primitive handful of hardwired features starts looking more like an instruction set (e.g. alpha blending is just your adder and multiplier. Disabling color clamp and you've got shifts and primitive conditionals.). In the simplest terms, it becomes "what can I compute in 16.7ms before the setup engine runs out of poop?"

Bleh, I think just rambled a bit unto pointlessness...
 
It's not complexity that you want, it's flexibility. IMO that's what the problem was with the DC and (to an extent the GCN). They exhibited the same characteristics that DX7 and earlier PC game development did, and that is you were basically molding your design around a bunch of fixed function principles ordained by the IHV (and/or ISV; API constraints. Although this is less of a problem on consoles). Visually, that lead to rather cookie cutter'ish looking games, vs. the more (IMO) creative implementations of the software renderer days. Also, you'd find yourself penalized rather significantly performance wise when you wandered off the optimal path of the hardware. So while they offered easy approachability and granted significant access to the upper echelons of their performance envelope, they also (in various degrees) created several boundaries to creative solutions to work around their limitations.

Where the PS2 has somewhat evaded this, thus extending is life span is that it's design has encouraged a fairly significant amount of creative solutions to what would be deemed at first glance as a significant feature (or lack thereof) roadblock. People looked at the EE and immediately commented how complex and difficult it looked (to be fair, the early devkit programming examples IMO led to some misconceptions on what you could and couldn't do). In reality it's not a terribly complex chip. It's really just a large pile of ALUs, registers and local stores, that places a lot of onus on the programmer to creatively leverage.

The GS has been an even more misunderstood part than the EE. At first sight the most common impression would be that of "ZOMG!!! It's just a a goddamn overclocked Voodoo w/almost no video memory!!!" But really leveraging the GS meant unlearning a lot of preconceived notions on what a GPU is and how best to get performance from it (i.e. *many* practices you learn to avoid doing for performance reasons on most GPUs the has GS gleefully encouraged). Instead of looking at it as a GPU (e.g. an overclocked Voodoo), you'd be better off thinking of is as a simple old 8-bit accumulator CPU (e.g. an overclocked 6502). The frame buffer starts looking more like a massive array of 8-bit registers, and that primitive handful of hardwired features starts looking more like an instruction set (e.g. alpha blending is just your adder and multiplier. Disabling color clamp and you've got shifts and primitive conditionals.). In the simplest terms, it becomes "what can I compute in 16.7ms before the setup engine runs out of poop?"

Bleh, I think just rambled a bit unto pointlessness...

You're right... complexity was a bad choice of words, but flexibility is more along the lines of what I meant... and of course, this flexibility comes at the cost of ease of use, but provide more gains later in the console's lifecycle.
 
On the other hand the DreamCast was one of the worst product to show a difference between first and last gen products (Soul Caliber always looked best to me and was a launch title)... same goes for XBOX1.

Uh... The Dreamcast got killed of after a year more or less. If it would have the same lifecycle and support as the PS2, you would probably end up with something simular.

Same goes for the Xbox, if it had the same support as the PS2, it would show something far more impressive.

Really, it has virtually nothing to do with the architecture, its about how many people are developing for it, there is allways room for improvement. If we spend another 10 years developing on the PS2, we would surpass GoW2, by quite a bit.

When it comes to current gen hardware it seems that both the PS3 and XBOX360 have been designed with significant headroom.

I think they were designed with price vs performance in mind, it definately wasn't a race for "who can make the most complex hardware". In-Order cpu's have better performance vs price ratio assuming you know how to code for it. Thats why the PS3 uses the Cell cpu. If there was a 1 cored OOO CPU with the same performance and price as the Cell, they would have used that one. Nobody goes out of their way to make things harder.
 
I would say quite to the contrary.. the Wii will get pushed the least because it offers the least amount of headroom.

Then that means you agree. By "pushed the most" I mean pushed the most to it's limits. If it's least powerful that means it will be easier to max it out... get it?:p
 
If I designed a console with a bunch of cores cobbled together with crappy software tools initially and had almost 10 years of game development on it from every developer in the industry, I'm pretty sure you'd see a huge difference between 1st generation games and final generation games. I don't think that would make my hardware design the shiznits...:LOL:
 
You will most likely see the Wii pushed the most hardware wise this gen. Some 360 developers aren't even bothering to make games that use more than one core.

The Wii will probably get pushed the most ( especially if it keeps selling like it is) mainly because it's single core.

I disagree, PGR3 and Kameo were using 3 cores from day 1, and the myth of games running on one core is just like ps2 not using vector units.

As for the Wii, I see reasons for not pushing it "Graphics do not matter @ Iwata - E3 2005". add the fact that the learning curve is not as steep as multi core platform.

ps3 is the most likely to be the biggest gap between first and last gen, providing the market accept it.
 
Uh... The Dreamcast got killed of after a year more or less. If it would have the same lifecycle and support as the PS2, you would probably end up with something simular.

2 years between December 98 and January 00.

You can even count 3 years until the release of Shenmue2.
 
2 years between December 98 and January 00.

You can even count 3 years until the release of Shenmue2.

Much of the art (textures, models, whatnot) dated back to around 98, or even 97. A lot of Shenmue 2 is actually very old, and if you look at it is very poor at showing off the hardware (no space allocated for mip-maps, no use of aniso, crazy mid Saturn <-> DC low polygon character modelling - especially in the first game). It's the art direction and business of the scenes (which "free" overdraw no doubt helped) that made the games look so good (not just the brute manpower).

Which leads me nicely on to my real point, which is that art direction and artist skill and familiarity with platform constraints makes a huge impact on how games look. Not to under emphasise the importance of software people, but a lot of what makes game visuals improve over the lifetime of a console is actually nothing to do with the flexibility or complexity of the hardware.
 
Flexibility also owes more to developer support than hardware. Rez's wireframes, Jet Set Radio's cel-shading, and Space Channel 5's fusing of video and 3D represent a wide scope of visual possibilities on DC for such a small amount/time of support.

For all of the variety in Shenmue, the 2-bit per pixel texture compression was also a big technical factor along with the high fillrate for opaque overdraw.
 
Well i would say the C64 would easily make the top. But in this case it would the PS2, the E3 demoes became reality and then some and MGS2 was in realtime!
 
I really don't understand why so many are hating on the Xbox1 to be fair..

Are you saying there were no significant differences between games like The orignal splinter cell, Halo and PGR, to games like Chronicles of Riddick, Halo 2, Splinter Cell: Chaos theory, Ninja Gaiden and so many amazing looknig titles that properly push the hardware until it was bursting at the seams (literally in the context of Halo2's texture streaming issues..)..?

I agree that the hardware was still pretty "conventional" and so possibly could have been a reason as to why it was so easy to leverage.. But it did recieve alot of leverage over the course of its life (possibly more in the visual side than anywhere else though..)
 
I really don't understand why so many are hating on the Xbox1 to be fair..

Are you saying there were no significant differences between games like The orignal splinter cell, Halo and PGR, to games like Chronicles of Riddick, Halo 2, Splinter Cell: Chaos theory, Ninja Gaiden and so many amazing looknig titles that properly push the hardware until it was bursting at the seams (literally in the context of Halo2's texture streaming issues..)..?

I agree that the hardware was still pretty "conventional" and so possibly could have been a reason as to why it was so easy to leverage.. But it did recieve alot of leverage over the course of its life (possibly more in the visual side than anywhere else though..)


Personaly I did not see that much of a gap between first and last gen titles on the XBOX. Even the difference between Halo1 and Halo2 did not seem like that great of a leap at all. If you compare that to the PS2 nicest looking lauch title SSX when compared to God of War 2 - NO COMPARISON!!!
 
Last edited by a moderator:
Personaly I did not see that much of a gap between first and last gen titles on the XBOX. Even the difference between Halo1 and Halo2 did not seem like that great of a leap at all. If you compare that to the PS3 nicest looking lauch title SSX when compared to God of War 2 - NO COMPARISON!!!
Compare Conker L&R to Xbox launch games. NO COMPARISON!!!1one either;)
 
Hmm, I remember an interview with Gengi 2 developer. He saed that they didnt use SPE-s, cos the game was in fact just a ps2 port with upgraded graphic. [but the grafics are awesome IMO, in some cases really jaw dropping]

Ill try to dig up that interview...
 
Hmm, I remember an interview with Gengi 2 developer. He saed that they didnt use SPE-s, cos the game was in fact just a ps2 port with upgraded graphic. [but the grafics are awesome IMO, in some cases really jaw dropping]

Ill try to dig up that interview...

How is that related to the thread topic? :???:
 
He probably means he expects massive improvements in future PS3 titles.;)

...For MASSIVE damage too no doubt. ;)

/offtopic

Seriously, this is a nobrainer. Ps2 1st gen titles didn't just look "humble", they looked horrible! Half res graphics that looked worse than the "lowly" Dreamcast launch titles. :LOL: It would be like ps3 games looking like Wii games. BTW speaking of Wii games. If people use Wii sports as a benchmark to start from, Wii may take the crown four years from now. ;)

The developers on this system worked magic to turn out titles like GoW and Black.

Every system has shown improvement from inception to death, but when you start that low, the only place to go is up. Add the best-devs-in-the-biz+time and viola. ;)
 
Seriously, this is a nobrainer. Ps2 1st gen titles didn't just look "humble", they looked horrible! Half res graphics that looked worse than the "lowly" Dreamcast launch titles. :LOL: It would be like ps3 games looking like Wii games.

That's a bit of an over generalisation!

How about Timesplitters? Dead or Alive 2? Tekken Tag Tournament? The Bouncer??

These were all 1st gen titles and all of which had audio engines which could easily stand up to some of the later titles..

...For MASSIVE damage too no doubt.

/offtopic

To add to the topic deviations, I remember when they were talking about Genji 2 being set in real life historical battles and then laughing at the giant enemy crabs..

Then a few days ago I was reading on wikipedia about the Japanese Spider Crab which live at the bottom of the ocean off the coast of Japan and have a leg-span which can grow as wide as 4m (the length of one & a half small cars) :oops:

& we all thought they were talking sh**.. :cry:
 
Back
Top