Ok, WIP and LIP (Learning In Progress)

[salf]

Hey, i decided to post this WIP (also LIP), Ive been working on it in the office, to learn about Radiosity and Fakeosity.

 

Modelled in AutoCAD and rendered with VIZ 4.

PC specs:

IBM

PV 1.8

1 Gb RAM

 

Theres 3 images, im gonna explain them:

 

1- First image is the whole model, with no materials, (well...only the glass ones) and is using Fakeosity, the usual dome of spot lights and 1 direct light for the sun.

 

2- The second image is a little closer, is using the same light system as the previous one, but i also added a whole of omnis to get more of those "ambient light" shadows.

 

3- The last one is using only 1 direct light as light source (sun) and with Radiosity, however the render you see isnt at 100% quality, cuz i stopped to radiosity calculation at 44% (it was taking tooooooooooo long, almost like 40 or so minutes) so the image you see isnt at its best!

 

The settings for the radiosity are this:

Initial quality 90%

Refine Iterations 6

Filtering 4

Radiosity Meshing Parameters 0.5 m

Log. Exposure Control

 

C&Cs are welcomed!.....actually more than welcomed they are needed!

[salf]

Ok, this is with a 100% radiosity calculation, other materials, background, etc. DUnno why those doorknobs are black, they should be white.

 

 

And this is another view of the model, is just a test, so it need way more details.

 

[William Alexander]

Salf,

 

Mesh processing at .5m is killing the processing times. 1/2 a meter with the entire building in view is to small, at least for a work in process. Try a mesh size at 15% - 25% of the structure, to set up the materials and detials. Then render at less.

 

The handles have not recieved any illumination-yet, or your material needs adjustment. For example using a metal material, the glossiness, specular and ambient work on a wierd curve causing the light conditions to render handles dark.

 

Overall it's a good start. A little flat, but thats the nature of the radiosity beast in viz. Consider pumping up some of the specular aspects of the materials and dynamic range of scene.

 

Keep working at it....and when you start dreaming in your sleep about it, you've made a break through.

 

rgrds

WDA

[salf]

Thanks a lot, i guess you mean that when rendering an exterior image, i should set the mesh at a bigger values, and when rendering an interior image, i should reduce the mesh size?

 

When i did that first try i was thinking more in a "smaller mesh = longer proccess = more perfect image" type off way.

 

Also, im using Raytrace shadows, but what type of shadows do you reocmmend?, i know is a try-and-choose proccess, but over you experience, what type of shadows do you use the most for exterior.

[salf]

ok, heres an update....and to tell you the truth im not that excited about it.

 

 

 

Im using only the daylight system, radiosity meshing is set to 10 m, but some layers (like floor and walls) are set on 4 m.

 

-Initial quality 75%

-Refine Iterations 4

-Filtering 2

Log. Exposure Control:

 

Physical Scale 92800 (same as the sunlight)

 

Now, i only have checked the DAYLIGHT option in here, so i had to turn down the BRIGHTNESS to 10, however i did a try checking also the EXTERIOR option, it makes darker the image so i turned up the brightness, it looks almost the same, no big improvement.

 

I kinda like more the looks on my first radiosity attemp, on the image with the number 3 above.

 

[ February 05, 2004, 08:06 AM: Message edited by: salf ]

[Crazy Homeless Guy]

not necessarily a comment on the rendering quality, but a question about the technique. the materials of the building, besides the floor, do not try to imitate materials in real life. so why does the grass need to be green? i think it would make the last round of renderings stronger if you made the grass a different color... gray? a gray blue to compliment the sky? a gray tan to imitate chipboard? ..the green you currently have draws a lot of attention, it is competing with the building.

[salf]

ok, i added the radiosity settings on that last one, notice that the only "big" difference between the last one and my first radiosity attempt, is that i reduced the meshing amount, but the results are wayyyyyyyyyy worst, is there a way to improve them?, without reducing the meshing again?, the VIZ will crash, i think i cant go any smaller in teh meshing size that what im now.

 

CrazyHomeless, youre right......but i feel sad because im actually looking for a more realistic rendering, with real materials, not the results im having, the grass color you see is just for the rendering test, the radiosity calculation was run on a diferent grass, a grey color so my building wouldnt look all green. If i ever get to the quility i want Ill use a realistic grass.

 

The estructure is metalic, painted in white, so im using the basic GLOSS WHITE material for them, (with ajusts on the reflection level), the mesh in the doors and windows is just provisional, and the walls are also gloss white, but with almost 0 reflection.