How to choose the best camera type!?

starbase1

starbase1

May the sauce be with you
I'm really not at all sure when to use which camera type, and what the requirements are for each. (I did look in the manual for 9, but it didn't help much, now on 9.6).

When I start to change this sort of thing I often seem to get unexpected effects, such as needing to make surfaces double sided to avoid lit parts appearing in unlit areas where normal maps are, or jagged borders at the edge of the lit areas, (like smoothing has turned off). Or lens flares go wierd, or they go weird with motion blur.

Is there a guide anywhere on how to choose appropriate combinations of camera type, camera settings, and render settings?

I'm always getting the feeling I'm not using these things optimaly - wasting render time, or generating cockups.

The particular case at the moment is for rendering my Earth model for HD animation, (which gave all the issues aboove at some time or another, in various combinations!). The object is basically 3 tightly concentric spheres, the inner one is opaque with a normal map, the next is the cloud layer, with a transparency and normal map, and the outer one is an airglow / haze layer, with additive transparency.

But a genreal; doc or tutorial would be even more handy.

Nick

Nick
 
Still stuck on Classic camera *sigh* but perspective camera seems to be the new daddy ! , render setting and lighting variables are endless ,just deepends on the scene !
 
Perspective camera if you need a regular, well, perspective. If you need something special, then it depends on what you need.

The 'real lens' thing is a bit pointless, because you can always distort and add vignetting in post-production anyway.
 
Captain Obvious said:
Perspective camera if you need a regular, well, perspective. If you need something special, then it depends on what you need.

The 'real lens' thing is a bit pointless, because you can always distort and add vignetting in post-production anyway.
Click to expand...

Yes, but that also requires larger renders. If the camera can be matched so you do not have to resort to compositing to do distortion and vignetting it does add to the saved buck, so to speak.
 
starbase1 said:
I'm really not at all sure when to use which camera type, and what the requirements are for each. (I did look in the manual for 9, but it didn't help much, now on 9.6).

When I start to change this sort of thing I often seem to get unexpected effects, such as needing to make surfaces double sided to avoid lit parts appearing in unlit areas where normal maps are, or jagged borders at the edge of the lit areas, (like smoothing has turned off). Or lens flares go wierd, or they go weird with motion blur.

Is there a guide anywhere on how to choose appropriate combinations of camera type, camera settings, and render settings?

I'm always getting the feeling I'm not using these things optimaly - wasting render time, or generating cockups.

The particular case at the moment is for rendering my Earth model for HD animation, (which gave all the issues aboove at some time or another, in various combinations!). The object is basically 3 tightly concentric spheres, the inner one is opaque with a normal map, the next is the cloud layer, with a transparency and normal map, and the outer one is an airglow / haze layer, with additive transparency.

But a genreal; doc or tutorial would be even more handy.

Nick

Nick
Click to expand...

Hi Nick.
Since you have gotten **** squat for answers here I will try and chime in.

The classic camera in Lightwave is excellent however its excellent at the sacrificing of speed compared to the advanced camera types (perspective, real lens etc) in certain cases.
One of the biggest issues some people have with the perspective camera - which I run into all the time when dealing with bad models is non-planar polys that used to render okay (no holes in the mesh) - rendering all frakked up with massive holes in them.
The reason why you are probably seeing this more and more is that most people couldn't "see it" way back when before the ACT were introduced like the perspective camera. So if you have models that were built in LW 8.5 for example, that render fine in 8.5 or in 9.x with classic but frak up in perspective camera - that should be the first thing you look at is non-planar polygons.

the other thing is intersecting polygons or polys directly on top of each other. While it used to be okay to do this in LW pre9.0 and in some cases was super handy (decals on stuff for example) - in some cases depending on what is going on - Lightwave can't tell if something is over or below another poly and it will start to twitch. Expecially if rendered over a network. One machine will say "its above" the other will say "its below". Insane!
But it makes some sense, but it can be fixed on newtek's end. Still, the problems you are getting with lensflares I couldn't tell you about unless I saw a scene file or example image.

Lithium.Kat
 
djlithium said:
So if you have models that were built in LW 8.5 for example, that render fine in 8.5 or in 9.x with classic but frak up in perspective camera - that should be the first thing you look at is non-planar polygons.
Click to expand...
More specifically, non-planar n-gons! Any quad polygon will pretty much always render fine in any Lightwave camera, no matter how distorted it is.

If you model out of four-sided polygons, you basically don't need to worry about how planar they are, unless they're *incredibly* distorted.
 
Thanks for the suggestions DJ lithium, much appreciated!

The flare thing seems to be that vector blur hates flares, but I intended it more as an example of a general issue where effects or options don't play nice with each other...
 
You must log in or register to reply here.
Back
Top