pouët.net

just a little remark on this "Iam an xbox developer so i know" bullshit : how come no xbox game are doing anything really cool? i seriously doubt that as an xbox developer you automatically know how to push a nv20, more than you can put really disgusting bumpmapping where it really shouldnt be.

"btw, as you allow dx9 in the demo/introcompo, and you only have DX8 hardware, isnt that kindof stupid? i mean why force users to download a hughe download?

but , then, can a dx8 demo use the size of dx9 as data files instead?"
dx9 has other components than d3d9, you know? besides, d3d9 exposes features that were already available on dx7 cards but not accessible in d3d before.

backing up kb/ryg on this. you guys are right. but stefan promised something that would push the nv20 to it's limits, (and as a xbox developer i know what it's capable of ;-)), so let's wait and find out!

see you guys there.

oh fuck

so i lost my post.
anyway, there are several reasons why we know some more about the nv20 than most developers do. the primary reason might be the direct usage of it applied at crucial points, another might be some of the fancy sdk documents describing techniques, limitations and features quite closely.

fact is that a lot of "lazy" developers more or less ignore the possibilities. this has several reasons. for starters, the PS2 is very popular. a lot of large gamefactories develop their game with a PS2 in mind, and even use PS2 test boxes to let their artists test freshly made content on. what does this mean for the XBOX? indeed, that it's not going to be pushed to it's limits, and that most people won't bother researching them, since it's damn easy to keep up with the performance of a PS2 (in 95% of the situations). a little more indepth research i had to do on kicking the XBOX GPU's arse (whilst bypassing Direct3D) has teached me quite a lot about what's possible and what's not.. and i guess this goes for quite a few others too. there's your answer.

if you don't like BP's compo-pc just visit another party. or hunt for easter eggs.

still that has nothing to do with the quote "iam an xbox developer so i know" since as you said yourself, most xbox developers dont even try to push the hw >P

it was meant to underline the fact that we were given the possibility to do some nice research on the subject, and the fact that it's being used as an argument, makes the assumption that thus has been done totally safe.

my mother has a possibility to do gf3 researches to

blabla confidentional info blabla etc, grow up.

I'm looking forward to the heap of ultra-über-mega-31337-dx9-demos which seem to be right around the corner... *cough*

Scalable effects are most definately possible and not that difficult if you know what you are doing. Trying to retrofit scalability into your effects when you haven't planned for it from the beginning is like trying to retrofit proper 'const' usage into a huge body of C or C++ code -- its nearly impossible.

Like I said, I would be impressed by scalable effects because it shows coding skills. Coding to a lowest common denominator means that we are doomed to never coding any effect that a TNT2 (or pick any old card here) can't render, which is pretty limiting.

If you code to high end cards that are available now, its true that people with low end cards might not be able to run your demo. (Although if you did scalable effects -- just like the games routinely do -- then it wouldn't matter.)

However, which is a more likely scenario for the future: people downgrade their cards to TNT2 class cards and give up on running fancier demos, or people upgrade their cards to DX8 or DX9 class cards and are able to run fancier demos.

If you can't grok how to do scalable effects, then code for a higher end class of card and eventually people will be able to see your demo in full glory.

Or if you're a "HW is killing the scene!" bigot, then go code software effects.

You can still make an impressive demo either way, but demos that use HW acceleration and restrict themselves to TNT2-class hardware (already several generations out of date) are unlikely to be impressive to other coders because they know what's possible using better HW.

Also, games still are playable when you turn off all the fancy graphics-options your card can't handle. When you that with a demo, what's the point of watching it?

legalize, why would anyone want to impress coders?

to quote netochka, "any fool can program and most do"

legalize: i'd say doing scalable effects, as you describe them, would take the reason why i do demos away :) Like i don't code to make readable, commercial-standard-code -- i do it to make something on the screen, if it looks okay then it will stay that way.

We want to impress all..

Its harder to impress your Mother/brother/sister than it is to impress a person in the demoscene. We where impressed bu simple linevectors on the c64 and our parents just looked at us and thought we where stupid throwing away our sparetime watching and creating that stuff.

Now a days we can do things that out parents atleast can watch without complaining about headache when the first picture is shown :)

Looking at the dx9 part of 3dMark makes me think that we even might get a 'thats quite nice' from a non scener if we work hard and can get the resources needed..

With more support from the HW we can get better output and maybe impress more people. Still we need all the other coding skills, its not only the material that makes the effects.. but it helps in the end. A cloth is more impressive with lit refractive bumps than just green polygons.

The only thing I conclude from this thread is that Stefan has a gift for stirring up shit.

as I'm the only one that's made a dhtml demo (don't say haujobb) I've demanded that this year's gathering compomachine runs at at least 3ghz so that I can shade 100 polygons..I just need that much power.... you just have to comply.

"let's push things forward" - the streets

Uhhh... that's how it all started. Crackers started competing with better and better intros on their cracked games. Pushing the hardware beyond its limits (sprite multiplexing, disabling borders etc).
Eventually these intros got so sophisticated that they were released separately. And they kept growing, into the megademos, which spanned an entire floppy, or more. Competition was the driving force here. And yes, it was all code. gfx and music were just icing on the cake.

So well, if you have to ask why anyone would want to impress coders, I guess you just don't get it. Making demos is a coder's ego-trip. He wants to be the best at what he does.

thom: sorry there aint any gaypr0n pics here

I'm personally tired of not seeing new technology taken advantage of. While it's one thing to exploit what you've already got to the fullest, it's another to completely disregard the power of today's hardware. Some things can be done more efficiently on newer APIs and hardware anyway, right? (I'm no demo coder, but it seems logical to me)

Also, I think the best way to specify limitations on videocards is by DirectX version. If the limit on a graphics card in the compo machine is DX8.1 compatible, then only install DX8.1 on the machine. This way seems ubiquitous and fair to me. This way minimizes ATI vs Nvidia, etc. I think a fair rule of thumb might be to rewind 1 year and see what version of DX was public at the time and make this the ceiling (similar to what someone else suggested). As far as DX inconsistancies between different cards of the same generation, that's a different story.

geforce-2/3/4 is fine with me =) and use dx9 since the interface changed to the better =)

Lepper: installing dx8.1 instead of dx9 is retarded, sorry.