View Full Version : 64-bits color depth??????

Coffe fan
25th Jul 2001, 10:16 PM
I saw that in an articacle saying about DirectX 9. Not that much difference between 16 and 32-bits, I can only notice some lack of color depth when zooming near a shadow on a wall. I had only seen over 32-bits when talking about photorealistic scanners.

What´s up with 64 if 24 is called True Color? Any 3D card up to date support this color depth?

26th Jul 2001, 03:58 AM
What's the point? The human eye can only distinguish a few million colors, much less than 24-bit is capable of showing. It's just excessive redundant technology.

27th Jul 2001, 12:12 PM
Not sure how it works out but 32-bit as I've seen looks way cooler but could also be because it ususally meant a more beefy system
No card to date supports 64-bit I think

27th Jul 2001, 12:24 PM
You'd need 64-bit color for simultaneous hardware additive and subtractive alpha blending. No card to date supports it though... you see 24 bits of color, the extra 8 are for alpha-blending(either addivive or subtractive). With 64 you could do additive and subtractive at the same time.

Coffe fan
27th Jul 2001, 01:38 PM
Hmm.... I was thinking on that, this color depth can only be about something related with effects, layers, such stuff.

Errr.. 24 is True Color and the extra 8-bits are used for alpha bleding but what #### is addictive and subtractive? Any site or pics showing what´s the difference?

27th Jul 2001, 01:42 PM
Additive is blending light, subtractive is blending liquid, that's the best example I can think of.
With additive, 0,0,0=Black, you add colours to make a new one. With subtractive, 0,0,0=White, add colours to darken it.

27th Jul 2001, 05:27 PM
ph34r the future of color depth: there would be channels for RGB, for the transparency, for the mirroring, for the height of the pixel if displacementmaps will be supported, maybe even for things such as the strength of the pixel: will or will not be destroyed when you shoot it, etc...

or do I have to "dream on" :D

28th Jul 2001, 03:26 PM
I just need 16 colors and im fine :D

28th Jul 2001, 07:08 PM
Dream on.
You'd have pixel-based collision detection, but deformation must stay on a per-polygon basis. However, you could deform the polygons based on per-pixel data.

28th Jul 2001, 09:33 PM
WildfireX2000, do you know the UT source or C++?

You really know your stuff.

28th Jul 2001, 09:54 PM
No, I'm programming an engine...
The reason I know that is that I have a very good intuitive understanding of how something would work. None of this is possible in the current, 436 incarnation of UT, but will probably be present in Unreal 2.

Coffe fan
29th Jul 2001, 12:42 AM
Do NURBS models have any advantage over polygon models? I saw that in rumors about DX9 too.

29th Jul 2001, 02:19 PM
heya WildfireX2000
mind if i could take a look at what your engine is like, do you have a site?
and could you sum up what the features are?

29th Jul 2001, 07:36 PM
There's no site, but we might release a tech demo in the next few months.
However, it features NURBS, per-pixel lighting, and all sorts of wonderful things(not to mention a completely realistic system, with a true blood vesses system, meaning a knife wound to the wrist would be more lethal than to the hand, and none of this is scripted). It also has, for some mysterious reason, dynamic multiview LOD...

Wait for the tech demo...

And the point of NURBS is a true LOD, meaning whatever you set it to, the model can deform to it. NURBS is also good for faster model loading...

Coffe fan
30th Jul 2001, 12:06 AM
What a @@, for what are programming a entire new game engine? Follow Unreal engine paths?

I´m wondering if U2 could show things like real bullets, really hurting, making holes on the body and with blood coming out.