From: Christopher Sean Morrison <brlcad@ma...> - 2013-08-31 15:11:10
Migrating the technical discussion to the mailing list where it belongs.
On Aug 31, 2013, at 8:32 AM, Harmanpreet Singh wrote:
> On Fri, Aug 30, 2013 at 12:41 PM, Christopher Sean Morrison
> <brlcad@...> wrote:
>> or converting them to boundary representation format (brep command) and comparing volume with the non-brep form to make sure they match (see the sh/cmp.sh script for examples).
> I convert implicit sphere (radius 5, units: mm) object into brep
> version and ran gqa -Av to calculate and compare their volumes. From
> output, I got different results (see undersigned). Are these expected
> results? Or do I need to set any flag to get desired output? Since,
> I'm not familiar with many of these calculations and values so need
> your suggestions.
The results will converge with the right settings.. The problem is in the usage. The sphere you made has a 5mm radius. Since you didn't tell gqa what sample density to start with, it starts with a default far too big for a model that small. It adjusts to something smaller, but not anywhere near small enough.
> mged> gqa -Av sphere
> Trying initial grid spacing: 50 mm
> Using grid spacing lower limit: 0.005 mm
> Initial grid spacing 50 mm does not allow 1 samples per axis.
> Adjusted initial grid spacing to 5 mm to get 2 samples per model axis.
> Using estimated volume tolerance 1 cu mm
> Processing with grid spacing 5 mm 1 x 1 x 1
There it said it merely sampled 1x1x1. It needs to sample something like 256x256x256 at a minimum.
You'll need to either make the sphere much much bigger (like 5m), or tell gqa a grid spacing (e.g., 0.001). There is a manual page that describes the usage (man gqa).