Unlimited Detail Tech blahblah

Sep 1, 2009
573
323
It seems to have a lot of potential, though.

But that potential depends upon many factors and also it would be bad if this increased game development time...you know why.

In retrospect this is just blurring the line between fact and fiction which recent movies have already done countless times. For example look at POTC3. The scene in which one character fights another on top of a mast, you can't tell the difference between the digital camera swing then it's transition into a real camera swing.
 

Da_Man

L4: Comfortable Member
Aug 23, 2009
173
39
As a proof of concept, yeah its cool.

In practicality, why the fuck do you need to render every single grain of sand separately? Does being able to render down to the atomic level really add anything to a game besides a longer loading screen?
 

BrokenTripod

L5: Dapper Member
May 11, 2009
248
65
I wonder if they plan on doing physics at all with this?

I suppose you could do it the same as they do it now with a really simplified collision box, but is the engine capable of making all these atoms rotate easily?
 

What Is Schwa

L6: Sharp Member
Jan 13, 2008
375
445
If this is true, doesn't it also imply an end to the graphics hardware wars? Huzzah!
 

grazr

Old Man Mutant Ninja Turtle
aa
Mar 4, 2008
5,441
3,814
In practicality, why the fuck do you need to render every single grain of sand separately? Does being able to render down to the atomic level really add anything to a game besides a longer loading screen?

If you actually listen to what they're saying, they're saying that this increases performance. On basic machines they're running 20 FPS on a near infinite amount of objects.

Yes, rendering at this level is kind of redundant since even with polygons we get reasonable levels of detail, look at BF3. People know and understand this process and a lot of money, as the video mentioned, went into the development of polygon technology.

But this appears to be the next step in technological evolution. When there's no detriment to performance, why not render details at such an extreme level? There's no negative aspect, so do it.
 

Pocket

Half a Lambert is better than one.
aa
Nov 14, 2009
4,694
2,579
I remember when this was first being discussed. And yeah, I'm very curious how they expect to fit all this into the memory that normal computers have. We have a hard enough time making games without intrusive loading screens now. I don't even want to think about how much more space this stuff will take up. It's nice that this would basically mean no more worrying about optimization for the level designers, though.
 

owly-oop

im birb
aa
Apr 14, 2009
819
1,215
As a proof of concept, yeah its cool.

In practicality, why the fuck do you need to render every single grain of sand separately? Does being able to render down to the atomic level really add anything to a game besides a longer loading screen?

Science isn't about why, it's about why not. You ask: why is so much of our science dangerous? I say: why not marry safe science if you love it so much. In fact, why not invent a special safety door that won't hit you in the butt on the way out, because you are fired!
 
Mar 23, 2010
1,872
1,696
-Tobie "supersandvich" McGuire
 

Freyja

aa
Jul 31, 2009
2,994
5,813
The fact is, if we stop having to buy expensive graphics cards to put in our computers, we can buy expensive 2^some huge number gigabytes of RAM instead, therefore memory won't be a problem!

I like to believe that this is real anyway. Games have to get to this point eventually, and how they explain it works makes sense if you have a processor that can do some decent calculations.
 

T-LOID

L1: Registered
Jul 22, 2011
11
0
Hrm... I'm sceptical about this. The way I could see this work is by having further away atoms collapsing into each other, and the closer you get, the more atoms you will actually see. That way you could perhaps always make sure there's only a limited amount of atoms on screen at any given time, making render times shorter. Sort of how Factor 5 did the dragons in Lair. They made a system whereby, the further away a model was, the less polygons would be rendered. the closer it got, the polycount of the model was dynamically altered. Heck, the commentary states that, when they went around the game with developer cam, they noticed the detail on the saddles of the dragons was amazing.

I'm not an expert though, so don't take my word on this. it's just speculation.
 

StickZer0

💙💙💃💙💙
aa
Nov 25, 2008
664
647
T-LOID, I haven't seen anything to suggest that what Factor 5 did was tessellation, but the concept of merging polygons automatically is why tessellation really needs to take off. Instead of models having LoD models, they just have one super detailed model, and it automatically only uses the polygons that are necessary and it works very well.
I'd have thought that this system makes use of tessellation.

Also, how come nobody made this thread when they released their first video like a year or two ago?
 

A Boojum Snark

Toraipoddodezain Mazahabado
aa
Nov 2, 2007
4,775
7,669
I'll be interested when their island isn't repetitive. I don't care what they say about it being because they "aren't artists", it seems suspicious. Most engines have efficient methods of handling identical content rendered multiple times.
 

Pocket

Half a Lambert is better than one.
aa
Nov 14, 2009
4,694
2,579
Hrm... I'm sceptical about this. The way I could see this work is by having further away atoms collapsing into each other, and the closer you get, the more atoms you will actually see. That way you could perhaps always make sure there's only a limited amount of atoms on screen at any given time, making render times shorter. Sort of how Factor 5 did the dragons in Lair. They made a system whereby, the further away a model was, the less polygons would be rendered. the closer it got, the polycount of the model was dynamically altered. Heck, the commentary states that, when they went around the game with developer cam, they noticed the detail on the saddles of the dragons was amazing.

I'm not an expert though, so don't take my word on this. it's just speculation.

From what I remember of the first video (I think there was a thread, actually; at least I remember seeing the video and I can't imagine where else it would have been posted that I go to), the engine basically maps each pixel to whichever "atom" is directly in front of it, ignoring all the rest. This takes care of the need to not render anything that's outside the line of sight as well as reduction of detail.

This approach would leave them without any way to filter textures to keep them smooth-looking, though. I don't know if antialiasing (which I guess would have to be of the "take a bigger picture and shrink it" variety) would be effective at compensating for this or not.