[Bf-committers] Bevel function
Sun, 7 Dec 2003 18:54:08 -0800 (PST)
Yeah, using the kind of previsualisation that Goof
used would be neat. The shrinking algo in itself only
deals with vectors, and the part that iterates on the
face and shrink them could probably be easily be
duplicated to draw gl polygons instead of actually
shrinking the face. No need to hack transform for
that. For one thing, it would be pretty and it's not
--- Matt Ebb <firstname.lastname@example.org> wrote:
> Perhaps it's a little premature for these sorts of
> suggestions (heh),
> but I have one. Manually entering the size value in
> blender units could
> get very cumbersome. When modelling, I often turn
> the grid right off,
> so I often have no concept of units in my mind. It's
> also difficult for
> people to track if they have scenes that are scaled
> up or scaled down
> further than usual. I can imagine it would take a
> lot of trial and
> error to get it right. If you could change the
> amount of bevel
> visually, by moving the mouse (think of Alt S
> shrink/fatten), it would
> be much faster and more painless.
> It would be great if there is some sort of visual
> interactivity to it,
> a la Wings. Disclaimer, I don't know anything at all
> about how this
> bevelling code works, but anyway, I'm prepared to
> look like an idiot if
> there's a chance it may be useful :)...
> It seems in goofster's edge cut and select that
> there's some way to
> previsualise an action, and then display it in
> yellow without actually
> enacting it. Could this be possibly done during the
> 4. 'shrink' step,
> or just previsualising the 4. shrink step before you
> commit the action
> (which goes and runs through all the steps again,
> with the values that
> it remembered from the previs stage). If the shrink
> was somehow made
> interactive (like hacking it in transform()? (could
> get nasty..), could
> this somehow be possible?
> Anyways, just some ideas. Excellent stuff so far.
> On 8 Dec 2003, at 4:59 AM, Alexander Ewering wrote:
> > Should there be any questions about the algorithm
> > here's the basic way it works (developped in a
> > night :)
> > 1. Split all faces so that they are "floating"
> > 2. Search for edges that lie exactly on top of
> each other.
> > Make a quad between them.
> > 3. Search for more than 2 vertices which lie
> exactly on
> > the same point. Make lists of lists of these.
> > 4. Shrink all original faces (i.e., excluding the
> new quads)
> > using an algorithm that moves a face's edge
> towards the
> > face's center independantly of the shape of the
> > This algorithm has originally been inspired by
> #math on
> > Dalnet, theeth has later integrated a simpler,
> > version.
> > 5. Take the list of lists of vertices, and make
> n-gons between
> > each of the "clusters".
> > n-gons are made by calculating the Boundbox and
> > all outer edges towards the center, merging in
> the middle.
> > Any questions on this wonderful algo? ;-)
> > | alexander ewering instinctive new
> > | ae[@]instinctive[.]de
> > _______________________________________________
> > Bf-committers mailing list
> > Bfemail@example.com
> Bf-committers mailing list
Do you Yahoo!?
New Yahoo! Photos - easier uploading and sharing.