Інсталювати Steam
увійти
|
мова
简体中文 (спрощена китайська)
繁體中文 (традиційна китайська)
日本語 (японська)
한국어 (корейська)
ไทย (тайська)
Български (болгарська)
Čeština (чеська)
Dansk (данська)
Deutsch (німецька)
English (англійська)
Español - España (іспанська — Іспанія)
Español - Latinoamérica (іспанська — Латинська Америка)
Ελληνικά (грецька)
Français (французька)
Italiano (італійська)
Bahasa Indonesia (індонезійська)
Magyar (угорська)
Nederlands (нідерландська)
Norsk (норвезька)
Polski (польська)
Português (португальська — Португалія)
Português - Brasil (португальська — Бразилія)
Română (румунська)
Русский (російська)
Suomi (фінська)
Svenska (шведська)
Türkçe (турецька)
Tiếng Việt (в’єтнамська)
Повідомити про проблему з перекладом
And while Novalogic's ray-casted voxels were old school one thing I could never figure out is how they achieved 6 DOF with ray-casted voxels. You cannot use the DOOM algorithm with 6 DOF which is why in DOOM you cannot look straight down or up b/c the ray casting algorithm is 2D and falls apart fast.
Novalogic ditched the voxels with the introduction of Joint Ops and they remade Delta Force and others without voxels once the hardware could support it.
But voxels have not died and are the future of graphics. But as of right now you cannot hardware accelerate voxels so instead you must cheat and hardware accelerate polygons that emulate voxels and are textured like voxels but aren't actually voxels in the purest sense. Nvidia has some very impressive demos about voxel worlds and they pull off some pretty amazing scenes in the demos but all modern video cards are still focused on triangles as opposed to voxels. Voxels add to the mix a bunch of issues with hierarchical animation, etc. Mixing voxels and hierarchical meshes is somewhat problematic although solvable but animating a voxel mesh has proven to be very complicated and until hardware supports volumetric rendering I would imagine we won't see that in any future games.
Voxels are used in the medical field for MRIs, brain scans, etc. and they are extremely detailed but again those are for the purpose of a one time scan and then 3D browsing of the object. A game has many more requirements so comparing the two is not fair.
There is one company bragging about infinite level of detail via voxels but so far all they have produced is barely functional herky jerky demos and honestly all of Novalogic's voxels back in the day look better than what that company has produced to date.
But also little that you can't already do exponentially quicker with current accelerated poly rendering, and a group of talented artists.
Voxel-based deformable materials are a useful tool, as are voxel-based ambient occlusion techniques, and so on... but that's the point - they're useful individual tools, when applied to specific use-cases where they actually offer a benefit.
They're not, and never will be, a valid start-to-finish rendering pipeline... however advanced directly rendered voxels get, shaded triangle rendering will have already advanced further.
That will always be the case, simply because shaded triangle meshes are effectively a cheaper and quicker hack, while voxel representations are literal and less abstracted. Triangle mesh tech only concerns itself with surface detail and throws all other information away (/never defines it in the first place), but that's fine, because in the vast majority of cases, all you ever see of an object is its surface, so lets spend the CPU/GPU cycles instancing 100000 trees to create a lush looking forest, rather than use the same resources to create one tree that we can chop through and see the rings... (an effect that could also be faked far quicker using polygon rendering and some clever code)
It even extends to content development.
How does a texture artist meaningfully create an object with full 3D interior detail?
There's no way you can paint that... the only way to do it is with an entirely procedural, highly programatic approach... and well, the vast majorty of texture artists prefer to paint textures using a tablet, you won't find many willing to spend years learning procedural content generation systems, just so they can paint up some creases on a guy's shirt.
- Again, like voxel tech itself, there are situations where procedural modelling approaches are highly useful, and highly time-saving... and those situations are becoming ever more frequent as technology develops, but there will never, ever come a time where it completely replaces the traditional creative process. For every handful of developers creating a deformable terrain system, there will always be a thoudand artists creating a hundred conventionally modelled/textured/animated characters to walk around on it.