 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
I have a town model with the horizon in the background and a high camera
angle. However the horizon is simply an infinite plane. Is there a way
to simulate the Earth's curvature? Can I use an actual sphere, or is
povray not precise enough to do that?
Mike
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Le 15-08-20 16:24, Mike Horvath a écrit :
> I have a town model with the horizon in the background and a high camera
> angle. However the horizon is simply an infinite plane. Is there a way
> to simulate the Earth's curvature? Can I use an actual sphere, or is
> povray not precise enough to do that?
>
>
> Mike
Quick and dirty: Use a large disk for your ground. In some cases, this
may be good enough.
Use a large flattened sephere. This can prevent, or at least reduce,
some problems with large scale differences between objects.
Use an actual sphere of the appropriate radius translated down by it's
radius so that the area of interest can be centered around the origin.
In this case, having milimetric scalled objects in your scene can cause
some problems: the ground could disapears or have holes, or some small
objects can go missing, get shifted around or have serious shading isues.
Hard core: Compile your version using tripple or quadruple precision FP
numbers. Reduce the epsilon value accordingly to the increased
presision, and very probably have the render time sky rocket.
Alain
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On 8/20/2015 7:15 PM, Alain wrote:
> Quick and dirty: Use a large disk for your ground. In some cases, this
> may be good enough.
>
> Use a large flattened sephere. This can prevent, or at least reduce,
> some problems with large scale differences between objects.
>
> Use an actual sphere of the appropriate radius translated down by it's
> radius so that the area of interest can be centered around the origin.
> In this case, having milimetric scalled objects in your scene can cause
> some problems: the ground could disapears or have holes, or some small
> objects can go missing, get shifted around or have serious shading isues.
>
> Hard core: Compile your version using tripple or quadruple precision FP
> numbers. Reduce the epsilon value accordingly to the increased
> presision, and very probably have the render time sky rocket.
>
>
> Alain
Can you think of a good equation to calculate how big the disc needs to
be? (Maybe overkill.)
Mike
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On 8/20/2015 8:03 PM, Mike Horvath wrote:
> Can you think of a good equation to calculate how big the disc needs to
> be? (Maybe overkill.)
>
> Mike
Actually, that's not going to work because I am using a checker pigment
on the plane and it will look funny on the disc.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Am 20.08.2015 um 22:24 schrieb Mike Horvath:
> I have a town model with the horizon in the background and a high camera
> angle. However the horizon is simply an infinite plane. Is there a way
> to simulate the Earth's curvature? Can I use an actual sphere, or is
> povray not precise enough to do that?
As long as we're talking about reasonably large scenes (the whole scene
- sans earth - measuring meters and upward), just go ahead with the huge
sphere approach.
Make sure to keep the coordinate origin near the camera. Some people
have modeled entire solar systems with the sun at the center, and were
surprised when "portraits" of Deimos (or was it Phobos? one of the Mars'
moons anyway) showed artifacts. This is because absolute precision drops
as the absolute magnitude of coordinate values increases.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
