|
 |
Over all I like the image. However in my opinion image maps are not the way to
go to get realism. The problem with them is that you either have to make a
REALLY high resolution image maps or you have to repeat the image map and when
you do this your eye picks up on these repetitions, even if only subcontiously.
If you use a lot of them your image can start to look like a Quake/Doom/Duke3D
level. I would have used standard textures and just tweaked them a lot until
they looked super real. The extra processing time could have been because of the
image maps, though I'm not expert on this. Anyone else agree with me about the
image maps making the scene look like a Quake level? Another thing is if you are
looking for hyper realism as you said you were then you need to round out some
of these objects. You won't find many books which are mathematically perfect
rectangles.
Ps: I like the glass texture on column of the lamp post.
Johnny Smith wrote:
> This image took over 55 hours on a 166 pentium (with 64 meg mem)
> using pov for windows 3.1E and a render priority of "highest". Nothing
> else running concurrently. (no background processes of ANY KIND)
> Radiosity settings were at the highest quality. (according to the
> radios.inc include file)
>
> Trying to create this image was a nightmare from the start. Quite a
> few of the objects in the scene were from 3rd party sources, and I had
> a hell of a time scaling/translating them as a result.
> I was trying for MEGA realism but as you can see below, I failed
> miserably. I just could NOT make the damn thing look REAL! There were
> supposed to be an actual set of open doors in the back wall leading to
> the outside. But I got so digusted
> with this picture I simply gave up by the time I got around to trying to
> create them.
> With the exception of the pillars and the desk, (and the obvious glass
> objects) everything you see here is a product of image maps used for
> textures. Without them, it wouldn't have even looked as real as it
> does.
> Anyways, enough rambling, Please let me know what you think of this.
> What you would change, etc. Maybe I'll work on this some more down the
> road sometime.
>
> ----------------------------------------------------------------------------
>
> Pixels: 312640 Samples: 781952 Smpls/Pxl: 2.50
> Rays: 22231340 Saved: 6623 Max Level: 8/8
> ----------------------------------------------------------------------------
>
> Ray->Shape Intersection Tests Succeeded Percentage
> ----------------------------------------------------------------------------
>
> Box 314511423 104826168 33.33
> Cone/Cylinder 165374433 28754828 17.39
> CSG Intersection 4139925844 24599418 0.59
> CSG Merge 49710803 24477307 49.24
> CSG Union 29965953 10157572 33.90
> Height Field 83102206 50684302 60.99
> Height Field Box 83102206 83102205 100.00
> Height Field Triangle 256728217 50932168 19.84
> Height Field Block 246779047 125384425 50.81
> Height Field Cell 1904852430 155420269 8.16
> Plane 32591814895 18825854942 57.76
> Prism 2750596 927879 33.73
> Prism Bound 12710107 8865686 69.75
> Quartic/Poly 589877000 6365363 1.08
> Sphere 5482155 2778557 50.68
> Torus 20817165 6627363 31.84
> Torus Bound 20817165 7377714 35.44
> True Type Font 1784419 520124 29.15
> Bounding Box 5279555424 2637925898 49.96
> Light Buffer 38495272 17899582 46.50
> Vista Buffer 93000566 82257263 88.45
> ----------------------------------------------------------------------------
>
> Roots tested: 599306367 eliminated: 3828796
> Calls to Noise: 2342534 Calls to DNoise: 56984996
> ----------------------------------------------------------------------------
>
> Shadow Ray Tests: 102898401 Succeeded: 39544033
> Reflected Rays: 167004 Total Internal: 25976
>
> ------------------------------------------------------------------------
> [Image]
Post a reply to this message
|
|
