Markus Maki
Newcomer
Sorry if I don't have time to check in every day, I'll try to do it once or twice a week. Being pretty busy, especially now that part of the team is at E3 this week, showing The Fall of Max Payne.
Whoa, a long question, in the Reverend style
Radiosity, Global Illumination we already see calculated to light maps. Real time -- we're still some way off from that. But there are some cool lighting techniques under investigation that work with shaders, real time. It's just that now they work in real time for
a) static geometry only
b) for small sets of geometry (one object = 30-60 fps)
Post processing effects like motion blur, dof etc. we're seeing already. Mostly on Playstation2, even Max Payne (PS2) used motion blur to cover the fact that the frame rate was always not optimal (I know that's bad) but lots of games use post-proc on PS2/XBOX. These are becoming viable when PS1.1 cards are common enough on the PC too (pretty much now). The only thing is that the additional buffers take a lot of memory, and stuff like depth of field or too much blur might not add any value to the game, just be annoying to the gamer. So it needs careful thought.
On the advanced materials, I think once we get rid of this DX6 style multitexture, we'll be there. I'd guess in about 2 years we'll be pretty much completely rid of those, and the dominant standard will be VS2.0/PS2.0. A lot of it is up to when we get PS2.0 to low-end and to the integrated graphics solutions, unfortunately.
Although raytracing hardware is being investigated, it's not yet practical and I don't think if anyone knows whether it will ever be. Search some Siggraph papers if you're interested.
---
theHunted asked:
Some developers may have such bad tools (as we had for our first game, Death Rally) that even we didn't want to touch them EVER again
And your second question is "intentionally left unanswered"
---
It may be better if I keep my fingers off the days' hot topic. Still, I'll leave you with a thought:
Would you rather have 100's of different, non-comparable benchmarks, or one forward looking "standard" to compare, although a standard too popular so that it's critical for manufacturers to perform best at it?
Whoa, a long question, in the Reverend style
Radiosity, Global Illumination we already see calculated to light maps. Real time -- we're still some way off from that. But there are some cool lighting techniques under investigation that work with shaders, real time. It's just that now they work in real time for
a) static geometry only
b) for small sets of geometry (one object = 30-60 fps)
Post processing effects like motion blur, dof etc. we're seeing already. Mostly on Playstation2, even Max Payne (PS2) used motion blur to cover the fact that the frame rate was always not optimal
(I know that's bad) but lots of games use post-proc on PS2/XBOX. These are becoming viable when PS1.1 cards are common enough on the PC too (pretty much now). The only thing is that the additional buffers take a lot of memory, and stuff like depth of field or too much blur might not add any value to the game, just be annoying to the gamer. So it needs careful thought.On the advanced materials, I think once we get rid of this DX6 style multitexture, we'll be there. I'd guess in about 2 years we'll be pretty much completely rid of those, and the dominant standard will be VS2.0/PS2.0. A lot of it is up to when we get PS2.0 to low-end and to the integrated graphics solutions, unfortunately.
Although raytracing hardware is being investigated, it's not yet practical and I don't think if anyone knows whether it will ever be. Search some Siggraph papers if you're interested.
---
theHunted asked:
I can't say for every developer, but it's the developer and the publisher together who decide. Usually it's good to release the tools, I certainly haven't regretted for a second we did that for Max1.
Some developers may have such bad tools (as we had for our first game, Death Rally) that even we didn't want to touch them EVER again
And your second question is "intentionally left unanswered"
---
It may be better if I keep my fingers off the days' hot topic. Still, I'll leave you with a thought:
Would you rather have 100's of different, non-comparable benchmarks, or one forward looking "standard" to compare, although a standard too popular so that it's critical for manufacturers to perform best at it?
