GDC 2003 Jason Rubin ( he mentions something interesting ;)

Hmm i'm expecting about 500million polygons a second . Thats with about 10 lights and most effects used . Lets hope he is right though .
 
anyone with a transcript? or at least the HOTTEST points he makes....

i really like the guy, apart from being a bit of a Sony lapdog (which is understandable really, given all the resources Sony give to him), he is one of the best developers in the world. what he's getting out of old 2000 hardware is pretty amazing. of course he has the backing of Sony's wallet and all the time he wants though...

Carmak who?
 
His basic premise was that Naughty Dog had worked on the successful principle(25 million games sold) of making each of their games have better graphics than the rest of the games in that genre and to make it fun. It was not to innovate, but that the graphics alone made the games better expierences ie, going from "2d -> early 3d - > High res 3d models" had a positive effect on gameplay.

And that the problem he's sees coming is that no longer will games be able to make better gaming expierences with better graphics. That we are reaching a plateau in which graphics will no longer wow people. He gave some pretty good evidence that making an innovative game is a very large risk, and that licsenes were becoming more necessary to make money from games.
 
I'm not so sure about graphics plateauing... What wow:ed me more than anything else in MP was the graphics.

Sure the sound is nice, monsters growl and yell and die nicely and they have their own sound when they walk and stuff, and ambient sounds are nice too and the music is kinda neat and gameplay is fun (though the controls are less than ideal).

Graphics though is simply OUTSTANDING. The little cube is flexing some serious muscle with rooms like the Frost Cave and Hunter Cave in Phendrana, or Hall of the Elders and Gathering Hall in Chozo Ruins.

Still there is lots of room to improve. GOOD bumpmapping (not dot3 which makes things too shiny), even more geometry for extra details and smoother curves, fsaa, advanced pixel shading techniques, realtime lights and shadows etc... The wow-factor isn't going to go away for at least another two generations I say. Hehe!


*G*
 
Boohoo, making games costs a lot of money today ...

However much they dislike it, the gamer will when given the choice between 2 games which are mostly the same on other levels still pick the one with the best graphics. Developers will continue have to deal with the fact that graphics make up part of their competetiveness, to say it is becoming less important is reasonable ... to say it will become irrelevant in the near term is stupid.

We arent at the point of photorealism yet, not by a long shot.

Marco

PS. dot3 is an operation, not a specific lighting model. It will be used in any reasonable bumpmapping scheme (I seriously doubt anyone on the cube is going to use embossing).
 
Grall said:
GOOD bumpmapping (not dot3 which makes things too shiny)

Dot3 bumpmapping does not necessarily make things too shiny. Diffuse dot3 bumpmapping doesn't make things shiny at all. Specular dot3 bumpmapping only makes things as shiny as the art guy chooses to make them. But when specular dot3 bumpmapping is used, it usually end up being overused.

The reason for the overuse of specular dot3 bumpmapping is because it seems that people doesn't spot bumpmapping at all unless it is overused. A good example of this was some time ago when some Doom3 screenshots without specular bumpmapping where posted (I don't remember if it was here or somewhere else). Because the graphics where not shiny, people claimed that no bumpmapping was used in the pictures at all. This was entirely wrong; every single pixel in those shots had diffuse dot3 bumpmapping.

It's quite strange really. Some people thought that those screenshots were not bumpmapped at all because of the lack of specular highlights, while to me on the other hand it was quite obvious that the pictures were bumpmapped, but I didn't notice that the specular bumpmaping was missing at all until the claims where made that the pictures had no bumpmapping what so ever.
 
to say it will become irrelevant in the near term is stupid.

His point was actually different: while on PSX a mediocre engine could seriously hamper the gameplay, on PlayStation 2 ( as you will see on PlayStation 3 ) a game powered by the Renderware engine, less efficient than the one Naughty Dog has conceived, has reached a roaring success even if the game was powered by not a top of the line engine, it did still open new gameplay while loooking acceptable enough...

The problem is that the difference between visuals produced by acceptable enough technical performance is not that much different from the visuals put out by the top of the line engine which spent years in R&D ( and tons of money )...

He says that Naughty Dog's formula was to make a game with Better Graphics than the competition and make it fun... now they are saying that the first part of that formula is becoming less and less important...
 
MfA said:
PS. dot3 is an operation, not a specific lighting model. It will be used in any reasonable bumpmapping scheme (I seriously doubt anyone on the cube is going to use embossing).

They can also use environment maps.
 
Jason is a smart guy. His ability to spot such trends in game development and adapt in time puts him above many, in my book.

We are already starting to see the beginning of this - it's getting increasingly difficult to make great looking games if you don't sit on a solid budget. Conversely, everyone but us, graphics whores, don't really care about graphics that much anymore (just look at GTA3/VC sales). If something is good looking *enough*, people are satisified. Just as Joe Regular doesn't care how beautiful cinematography or photography particular movie has, the same is starting to be applied to games.

It will probaby matter even less in the next generation, when graphics in every game will reach such a solid feel that few can complain about them, and when making something outstanding looking will be restricted only by budget, talent and development time.

I also think some Japanese developers would do good for themselves if they start spotting such trends (not necessarily this particular trend), or they might get stuck with making games noone is buying, or great games which they try to market to wrong groups of people.
 
Well... he obviously knows everything to know about PS3 at this point being a key developer and all. Maybe this is his hint of to what PS3 can do?
 
1 BILLION polygons per seconds

dr.evil.jpg


Jeebus, its hard to estimate how it would looks like.. any estimates on how much toy story or FFTSW have in their scenes? I know its hardly comparable with an in-game performance number but still..
 
slightly off topic, but is there a way to convert real player streams into something more convinient (ie. mpg, mov etc)? I don't have real player but really need to get a hold of that video... help.. please... pretty please.. :(
 
Phil said:
slightly off topic, but is there a way to convert real player streams into something more convinient (ie. mpg, mov etc)? I don't have real player but really need to get a hold of that video... help.. please... pretty please.. :(

The prob this one is strictly streamed only ... otherwise there are ways:

rm -> avi [raw == big, bigger, biggest]

http://www.divx-digest.com/articles/rm2avi.html

avi -> whatever

not freeware
mpeg1/mpeg2, tmpgenc http://www.tmpgenc.net/

freeware [mpeg4]
xvid, http://www.xvid.org/ + virtualdub http://www.virtualdub.org/

.... enjoyjoy
 
Woah, best of thanks ChryZ. I'll just give it a try... :) BTW; Sorry for not replying to your PM yet - I'll get it done shortly. :D
 
Oops yeah forgot about that, still its only a LUT ... the lighting model it is replacing would still have had dot3s in its calculations.
 
Just some quick calc's- at 60 fps, that's 16 mil polygons per frame! On a 640x480 screen, that's literally 50 polys per pixel. Take it up to a basic HDTV resolution and that's still nearly 14 polys per pixel- or alternately, polys potentially 1/14 the size of a single screen pixel. I think the "n-gon" look and "cube finger" affliction will be pretty well decimated at that point.

...or you could look at this way- you could conjure up a graphics engine that works at a paltry 7% poly throughput efficiency and still come up with 1 poly per pixel (comparable to a 30 mil poly/sec game from today). That's a lot of resources left out to devote to lighting, effects, and filtering!
 
A single hard shadow casting lightsource can generate as many poly's as the unshaded scene itself.
 
Back
Top