[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/vr/ - Retro Games

Search:


View post   

>> No.8991848 [View]
File: 113 KB, 622x466, sgi-diddy.jpg [View same] [iqdb] [saucenao] [google]
8991848

>>8991809
>>8991809
>The Mario model alone has more polygons than the entirety of the original game
Well, not exactly. The original renders are made from NURBs models, not polygon mesh models; they're basically spline-based and the final smooth surfaces come from setting a certain resolution for them, which determines the number of "steps" each curve is calculated and visualized as in renders. Anyway, the point is that the SGI render assets and the game assets and completely separate stuff and built from scratch using two different frameworks (and most likely by two different art teams). Consoles didn't even have the proper graphics architecture to use NURBs models to begin with afaik.
For a while NURBs were the standard for high quality non-realtime CGi but eventually the industry set on polygon meshes (subdivision surfaces) because they were proved to be more flexible, versatile for continuous organic surfaces, and worked better for deformation and thus for animation. For example, Toy Story was made with NURBs even if Pixar's cofounder actually co-INVENTED the subdivision algorithm in the 70s, but by 1998 they were moving the whole operation already towards polygon meshes.

>> No.8503089 [View]
File: 114 KB, 622x466, sgi-monke.jpg [View same] [iqdb] [saucenao] [google]
8503089

>>8503019
>>8503037
It's amazing how many times the "SGi" thing gets thrown around in these threads as if it meant anything by itself, but in the end it was just hardware really. The "look" came from the state of the art at that time (read: its limitations), as well as the distinctive look of NURBS surfaces versus subdivision mesh geometry, which would become the standard later. Even Pixar didn't switch to the latter until after Toy Story 2 iirc.

Raytracing didn't really implement bounce lighting until much later in the early 2000s; Global Illumination is what really blew up the whole realism thing open, removing much of the "hand crafted" lighting work that was required until then and which resulted in WILDLY different visual styles depending on the people involved (knowledge, meticulousness and good artistic taste weighted more than the render software doing its magic). In that way, it was a fascinating time for the industry, but there's no tragedy in moving beyond that as the technology advances. That same savvy can be still applied on top of what the technology can do by itself to elevate the whole thing even further and dwelling on "the good old times" is just generalized anxiety about change.

>> No.8285286 [View]
File: 114 KB, 622x466, expand-count.jpg [View same] [iqdb] [saucenao] [google]
8285286

>>8284608
>>8284628
There's actually no "polycount" in this early SGI visuals. They used spline-based models, not meshes. A somewhat simplistic comparison would be vector-art vs bitmap/pixel-art.

Even Pixar didn't cross over to meshes until a couple feature films in iirc, and that eventually became the standard. I don't think this amounts to the clearly noticeable difference, but it could be a factor.

Navigation
View posts[+24][+48][+96]