jhl@tty3b.UUCP ("Jonathon Luers" 97320) (12/20/84)
<superfluous bug repellent> I have a basic question about how high-quality shaded color graphics, like you see on a SIGGRAPH poster, are generated. I assume the programmer starts with a 3D mathematical model of the object, and then calculates the reflections from a light source off the surface of the object to whatever viewing position is selected. Now it seems to me that a package like GKS or CORE is ideal for describing the object and the viewing angle, but how are reflection, shading, anti-aliasing, etc. handled? Aren't those calculated pixel by pixel? But that doesn't fit in the GKS model, since the user program doesn't know anything about the device, such as resolution, etc. Any GKS or 3D rendering experts out there care to comment? Thanks. Jon Luers AT&T Teletype Corporation ihnp4!tty3b!jhl
jwp@utah-cs.UUCP (John W Peterson) (12/23/84)
> "Is GKS useful for 3D shaded graphics"?
NO.gwyn@brl-tgr.ARPA (Doug Gwyn <gwyn>) (12/26/84)
> > "Is GKS useful for 3D shaded graphics"? > NO. I'd like to add that, like most panaceas, GKS and CORE are "lowest common denominator" approaches. High-quality computer graphics of any kind requires attention to detail that device-independent packages cannot adequately provide.
greg@vecpyr.UUCP (Greg Millar) (12/29/84)
> > "Is GKS useful for 3D shaded graphics"? > NO. If you are doing ray tracing, or anything else where you are calculating individual pixels you can use the GKS Cell Array. This buys you device independence (which is one of the big deals of GKS), although depending on the GKS it may have some performance cost. If you just want to pump out polygons of different colors to a device GKS also comes in handy. GKS does not do 3D shaded graphics, but it is useful with 3D shaded graphics. Seems that too many people want graphics standards to do more than they were designed to, but that doesn't mean they should not be used.