Hi,

QUESTION 1
what would improve the render speed of my scene: many small difference
holes (boxes) inside a blob, or one large difference hole (box)? I need
to understand better, how I can improve the render speed of my scene.
Thanks.

QUESTION 2
And what renders faster: a shape made of many boxes of different sizes,
or a mesh resembling the same shape? Thanks again.

Post a reply to this message

On 1/31/2016 2:11 PM, Sven Littkowski wrote:
> Hi,
>
> QUESTION 1
> what would improve the render speed of my scene: many small difference
> holes (boxes) inside a blob, or one large difference hole (box)? I need
> to understand better, how I can improve the render speed of my scene.
> Thanks.
>
> QUESTION 2
> And what renders faster: a shape made of many boxes of different sizes,
> or a mesh resembling the same shape? Thanks again.
>

Have a look at Mike Williams site. There are lots of good tips there.

http://www.econym.demon.co.uk/holetut/index.htm

-- 

Regards
     Stephen

Post a reply to this message

Am 31.01.2016 um 15:11 schrieb Sven Littkowski:
> Hi,
> 
> QUESTION 1
> what would improve the render speed of my scene: many small difference
> holes (boxes) inside a blob, or one large difference hole (box)? I need
> to understand better, how I can improve the render speed of my scene.
> Thanks.

Well, you /could/ achieve that better understanding by experimenting
with it ;)

In general, you want as few hole objects in any given other object.
For instance, for your hangars you're most likely better off to just use
one large box for the hole, and then add walls to partition that hole
into multiple hangars - even if that means adding many more objects.


> QUESTION 2
> And what renders faster: a shape made of many boxes of different sizes,
> or a mesh resembling the same shape? Thanks again.

I have no idea. You will indeed need to experiment with this one.

My personal expectation would be that the many boxes render faster
(provided that there are no intersections or differences involved) as
you need up to 12 triangles to replace one box, and the render engine
could no longer exploit the fact that the opposite sides of the shape
are parallel. On the other hand the mesh-specific bounding mechanism
might be powerful enough to compensate for this.

Post a reply to this message

Thanks to both of you.   :-)

All these questions came up just during the long render of the partial
scene. No way to experiment, as the scene rendering still goes on (7
days 23 hours 33 minutes = 19%). But as soon as that render of which I
expect the result so eagerly is done, I am doing a lot of restructuring
(minimizing amount of differences) and testing the meshes versus boxes.

Especially the red hangars cost time. Boooo!

I will experiment with Sketchup, if I somehow can use its meshes inside
POV-Ray with some converting.

I am also checking out (right now) that website proposed by Stephen.

Post a reply to this message