AC3D Forums - View Single Post

Dennis · 29th June 2007, 06:57 PM

I performed a couple of fixes to correct a couple of calculations, but have overall bad news on the volume calculator.

I did some research and found a couple more articles that use the divergence theorem to calculate 3D volumes. I created an implementation that takes a different approach (using the same theorem) from the one I use, and it consistently comes up with the same result as my existing algorithm, inaccuracies and all.

Both implementations also have the unfortunate effect of changing the results with a simple Divide on the object, which, one would think, should have no impact on the volume.

Further investigation yielded that this is a known issue with many volume calculation implementations based on this theorem. There are supposedly solutions that address this, but it's not something I would have the time to research/implement.

Meanwhile, I've left the page up with the volume calculator, and have added a disclaimer at the top:

"The resulting calculated volume using this algorithm known to be inaccurate, and any results from this plugin should be taken as an estimated volume at best. For some geometry, testing has shown a +/-25% discrepancy in calculated volumes from actual volumes."

I'll leave it out there for now, but this is still, unfortunately, not a working solution...

29th June 2007, 06:57 PM	#5
Dennis Senior Member Professional user Join Date: Jul 2003 Posts: 899	Re: Calculate the Volume of an Object I performed a couple of fixes to correct a couple of calculations, but have overall bad news on the volume calculator. I did some research and found a couple more articles that use the divergence theorem to calculate 3D volumes. I created an implementation that takes a different approach (using the same theorem) from the one I use, and it consistently comes up with the same result as my existing algorithm, inaccuracies and all. Both implementations also have the unfortunate effect of changing the results with a simple Divide on the object, which, one would think, should have no impact on the volume. Further investigation yielded that this is a known issue with many volume calculation implementations based on this theorem. There are supposedly solutions that address this, but it's not something I would have the time to research/implement. Meanwhile, I've left the page up with the volume calculator, and have added a disclaimer at the top: "The resulting calculated volume using this algorithm known to be inaccurate, and any results from this plugin should be taken as an estimated volume at best. For some geometry, testing has shown a +/-25% discrepancy in calculated volumes from actual volumes." I'll leave it out there for now, but this is still, unfortunately, not a working solution...