[Radiance-general] Frozen Octree instances in a different location

[Greg Ward]

Hi Terrance,

Glad you sorted it out.  Another trick that works and is slightly more efficient is to create a "partial octree" containing your scene except building, then add to it, i.e.:

	oconv -f materials.rad Scene1.rad > octrees/frozen-site.oct
	oconv -f -i octrees/frozen-site.oct Building1.rad > octrees/building1-site.oct
	(make a second version of building)
	oconv -f -i octrees/frozen-site.oct Building2.rad > octrees/building2-site.oct

The above will work without the -f option, also, if you want your octree files to be a little smaller and you don't plan to move them about.  Adding geometry to an existing octree can be much faster than rebuilding the whole thing from scratch.  The one caveat is that your new geometry must fit entirely within the bounding box of the original geometry.  If it will not, you can use the -b option in the first oconv to specify a bounding volume that encompasses everything.

Cheers,
-Greg

> From: Terrance McMinn <t.mcminn at ocpms.com.au>
> Date: August 23, 2016 7:55:18 AM PDT
> 
> I have a Sketchup model with two layers: Building and Site
> 
> Due to the export time (su2rad) of the Site layer, I figured I could switch off the building layer and export the site once create a frozen octree (oconv -f materials.rad Scene1.rad > octrees/frozen-site.oct) then instance the site back into the building project with a simple 3 line file:
> 
> void instance site_instance
> 1 octrees/frozen-site.oct
> 0
> 0
> The unexpected result is that the site is offset to the building. I did not expect to need to add transform arguments (X, Y, Z) to the instance definition:
> 
> void instance site_instance
> 5 octrees/frozen-site.oct -t X Y Z
> 0
> 0
> Am I missing something fundamental and just need to experimentally determine the offset?
> -- 
> Regards
> Terrance McMinn
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.radiance-online.org/pipermail/radiance-general/attachments/20160823/52b7b3a6/attachment.html>