Writing your own QEM routine is not actually terribly difficult. The
difficult bit is burning through the mind-bending maths notation used in the
papers (well, I have trouble with it anyway) and finding underneath it a
very simple routine.
Alternatively, there's source code of a very simple QEM implementation
sitting in my Games Programming Gems 2 article, based on the Garland and
Heckbert paper. It generates a VIPM collapse thingie (which is cool by the
way, and dead simple to use), but if you really don't want continuous LOD,
it's pretty simple to rip out the QEM part of the code and use it with
whatever mesh data structures you have.
I have no idea what license the GPG2 source is published under, but if I
have any say in it, it's published under Share and Enjoy as long as you
email me saying "thanks". Just to be courteous.
Email me offline if you want further specific help on this bit of code.
Tom Forsyth - purely hypothetical Muckyfoot bloke.
This email is the product of your deranged imagination,
and does not in any way imply existence of the author.
> -----Original Message-----
> From: Ben Discoe [mailto:ben@...]
> Sent: 09 May 2002 00:35
> To: GDAlgorithms-list@...
> Subject: [Algorithms] open source for TIN decimation?
> I'm looking for an unencumbered implementation of an algorithm to
> create simplified versions of a TIN heightfield. Does anyone
> have any
> The situation: i'm loading an enormous indexed TIN terrain,
> and need to
> render it in real time (not the 0.8 fps i get from passing it
> to OpenGL).
> Continous LOD is not necessary, just a way to create simplified
> versions of the TIN to use at a distance.
> So far i've found:
> * Michael Garland's "Scape". It assumes the input is a regular grid
> heightfield, not a TIN.
> * Michael Garland's "QSlim". It handles arbitrary models, but the
> current version is GPL'd so it can't be used on the project i'm on
> which is not GPL'd.
> * Hoppe's "Mesh Optimization" work. It's not free for commercial use.
> Algorithms which do arbitrary models are probably overkill,
> since this
> is simple heightfield data.
> Any help??
> Ben Discoe
> Have big pipes? SourceForge.net is looking for download
> mirrors. We supply
> the hardware. You get the recognition. Email Us:
> GDAlgorithms-list mailing list