exterior radiosity tutorial...WTF???

[plastic]

refering to the current CGArchitect monthly column/tutorial:

http://www.cgarchitect.com/upclose/article8_TB.asp

 

are you kidding?

i don't see any sense in this tutorial other than misguiding/confusing beginners.

 

what is the point in simulating early 80's 3d look by using advanced rendering techniques?

'how to make a scanline rendering even more ugly while increasing the calculation time significantly'?

 

apart from the obvious i'd like to point out following things:

 

-) using 3d trees with radiosity is total madness.

 

-) did you ever see color bleeding leaves? max uses the ambient color for color bleeding, viz the diffuse. its not nessecary to use 'radiosity override' BUT it's important to watch the color swatches, even when textures are used.

 

-) using per object radiosity flags results in a big mess. there is hardly a reason to do so, except in special situations.

it's better to use a global meshing size setting for everything but the ground plane, NOT right around.

 

-) whats the purpose of the non-photometric omni light under the house? lighting the background skydome?? hint: material self illumination and bitmap output values.

 

-) the physical scale in the exposure settings isn't needed for daylight, it doesn't affect anything as long as there are only photometric lights, as it should be.

 

The scene does not look good ...

thats right, and NOT because

 

The problem with this model is that it was not created specifically with Radiosity in mind and, for example, the vertices of the roof do not match up with the vertices in the walls nor do the foundation vertices line up with landscape vertices.

VIZ4 and max5 are very indulgend towards 'dirty' models, as opposed to lightscape.

the problem with this scene is that the radiosity parameters and colors are totally off.

0-255-0 type of green simply doesn't work for grass. i didn't measure, but i'd say that most of the colors in this scene are out of the physically working range. (reflectance, transmittance)

 

sorry for my very negative view, but i'm quite shocked and urgently recommend to overwork/pull this tutorial.

 

[ December 02, 2002, 08:32 AM: Message edited by: plastic ]

[seismograph]

hey! someone have read it!

shame on you!

[Craig Ramsay]

The light tracer would be a much better way to light an exterior scene!

 

Exterior radiosity eek!!!

 

[ December 10, 2002, 09:51 AM: Message edited by: Craig Ramsay ]

[seismograph]

Viz4 doesn't have lighttracer, but you can use vray instead, which is much faster than lighttracer.

I wonder why viz4 still don't have lighttracer.

 

I personally think that a well designed skydome with 4-5 lights can be a pretty nice and fast solution for animations but not together with radiosity.

[plastic]

actually radiosity works quite well for exteriour scenes. in combination with max 5's undersampled regathering.

the advantage over GI is that it's still faster, especially when many different views are needed.

with viz4 vray free is most likely the better choice.

[id_ivan]

the advantage over GI is that it's still faster, especially when many different views are needed.

with viz4 vray free is most likely the better choice.

I don't think it's absolutely right. Try to add reflection to objects which is facing each other then you'll get a whole night for one view with GI. With LS i calculated appx. 1.5 hr and render only 45 mins each view at the size.

[plastic]

Originally posted by id_ivan:

quote:

---

the advantage over GI is that it's still faster, especially when many different views are needed.

with viz4 vray free is most likely the better choice.

I don't think it's absolutely right. Try to add reflection to objects which is facing each other then you'll get a whole night for one view with GI. With LS i calculated appx. 1.5 hr and render only 45 mins each view at the size.what i was trying to say is: RADIOSITY is still faster than GI, especially when many different views are needed.