Jim Wilson

 
> It occurred to me that our 3-D modelers might not all be 
> running plib cvs. 
> Andy came up with an excellent improvement to the ac3d 
> loader/optimizer (new optimizer actually) that incorporates 
> logic to automatically split vertices in two so that they go 
> into seperate normal calculations (averages).  Briefly, the 
> plib change gives us sharp edges automatically.
> 
> For the many aircraft models this change really helps.  But 
> for the models what were already designed to work around the 
> problems with the earlier plib, they end up looking a little 
> worse.  I've gone through a few and for the most part they 
> have square or sharp leading edges on the wings now.
> 
> I'm bringing this up now because this is the version of plib 
> that the next release of FlightGear will released for.  
> Looking at my own models,  this has proven to be a very 
> difficult problem to deal with.
> 
> Basically what you'll see is that a sharp edge is created 
> whenever the angle between the normals of adjacent surfaces 
> is greater than 46 degrees.  Another way to describe the way 
> this works is that a cylinder object will appear round if it 
> has 8 or more sides.  7 sides will have sharp edges.
> 
> This creates a serious problem for ac3d/blender modelers 
> because you will need at least 4 surfaces to make up a 
> leading edge.  In practice you actually need 6 or more 
> because wing edges are not perfectly round.
> 
> At some point simgear and plib will have to be modified to 
> support the ac3d 4 format or some similar scheme for setting 
> the angle.  Either that or we'll finally be able to move to a 
> file format that actually supports normals "normally".
> 
> For the time being though,  I propose that we try to get this 
> default angle to a higher value.  From my testing 61 degrees 
> seems to be a good number.  61 degrees means that a 6 sided 
> cylinder will be smoothed.  Anything less (e.g. 4 or 5 sides) 
> will be sharp sided.  Many of the aircraft models will still 
> need some modifications to recover their rounded shapes at 61 
> degrees,  but the process is much easier, with far fewer 
> vertices, etc.
> 
> Best,
> 
> Jim

We seem to have jumped into something without really thinking it through.
Pity. We need to implement something like the AC3D 'crease' sooner rather
than later. 

Regards

Vivian 



_______________________________________________
Flightgear-devel mailing list
[EMAIL PROTECTED]
http://mail.flightgear.org/mailman/listinfo/flightgear-devel

Reply via email to