I am probably opening myself up for my own Clark Griswold moment, but I took a shot at the FaceLinking in away3dlite. I tweaked the FaceLink.as from away3D the best I know how to get it to work with away3Dlite. I am not an expert programmer, but I *think* it works...
Here is a test swf and some links to faceLinkTest.as and my tweaked for lite FaceLink.as : http://moosemouse.com/face-link-test.html Shawn P.S. - I don't really know the protocols for tweaking someone else's code, so if I have done something uncool please let me know. On Jun 26, 12:15 pm, Fabrice3D <[email protected]> wrote: > thats why the offset is made for, you can pass a negative offset as well, > but this also means that in certain cases the object would look like if its > inverted. > > > When a certain face in an animation faces a certain way (opposite? etc), > > the link seems to rotate the object the wrong way. > > Problem here is more of a visual perception kind than a math one I think. > > > This may be one of the hardest things to get right, since it's pretty hard > > to test. > > actually, it remembers me I started wrote a "face" explorer just for this > purpose a while ago for Prefab for a personal project, but never finsihed it. > may be I will as it seems the class is being used fairly often. > > > The coolest thing about this feature to me is the ability to attached a gun > > or backpack on a moving model for a game. > > I've seen already pretty creative uses of it, but yes, tmp attached items > such as guns while keeping independant object behavior were > the reasons why I wrote it in the first place :) > > Fabrice > > On Jun 26, 2010, at 3:52 PM, Peter Kapelyan wrote: > > > Hmmm, > > > I don't know if to take that as a compliment or if thems fightin' words :O > > > J/k But, here comes the rain again. > > > A (the?) correct Facelink should work on animated models as well like in > > this demo: > >http://www.closier.nl/playground/facelink/facelink2.html > > > However, I don't think Fabrices code works right 100% of the time. When a > > certain face in an animation faces a certain way (opposite? etc), the link > > seems to rotate the object the wrong way.This may be one of the hardest > > things to get right, since it's pretty hard to test. > > > The coolest thing about this feature to me is the ability to attached a gun > > or backpack on a moving model for a game. You can then have tons of > > separate models made that just magically "stick" to where they are supposed > > to be. And there is the problem, it seems more like magic than something > > that is possible :O > > > I have seen this feature in other 3d engines, not sure, but I think I > > usually saw it as attach to vertice, not face. > > > Anyways I wanted to congratulate you for getting stuff to work so far, it > > looks like it could be really handy! > > > -Peter "Party" Pooper :P > > > On Fri, Jun 25, 2010 at 11:09 PM, savagelook <[email protected]> > > wrote: > >http://savagelook.com/blog/away3d/away3dlite-face-linking-take-2 > > > ok, there's my code using an ObjectContainer3D to keep the alignment > > correct. I'm guessing there's a way to do it with upAxis, but it's > > beyond me until I do some 3d math studying. Oh, and I left a little > > something special for you peter ;) > > > On Jun 25, 11:36 am, Shawn <[email protected]> wrote: > > > Yes, that is how I would have done it with parenting :) > > > > On Jun 25, 6:02 am, savagelook <[email protected]> wrote: > > > > > Well, I think your code/math is over my head for now, but I get what > > > > you meant by parenting. I put the source object and the linked > > > > objects inside an ObjectContainer3D and everything works as expected. > > > > This should suit my needs for now. I've got a book on 3d math I > > > > ordered a little while ago that I haven't had time to dive into yet. > > > > Looks like this might be the catalyst for finally reading it. > > > > > On Jun 25, 8:22 am, savagelook <[email protected]> wrote: > > > > > > I'm looking at your code now to see how I can apply it, but what do > > > > > you mean by "use parenting" to solve this problem? I'm open to > > > > > anything. > > > > > > On Jun 25, 3:59 am, Shawn <[email protected]> wrote: > > > > > > > I believe the problem is with the lookAt's upAxis. lookAt first > > > > > > points > > > > > > the z axis at the target and then tries to align a local up axis > > > > > > with > > > > > > the scene's Y_AXIS. So, as your faces move they keep adjusting to > > > > > > match their up axes with the scene's Y_AXIS. > > > > > > > Unless you want to use parenting, the only way I know to fix this is > > > > > > to use an "up object" to aim the up axis at after the lookAt has > > > > > > been > > > > > > done. I have a test of an up object with a link to the code > > > > > > here:http://moosemouse.com/up-object-test.html > > > > > > > I have been using this for something quite different so it may not > > > > > > immediately make sense to your project. But I *think* this is one > > > > > > way > > > > > > to fix your problem. It is late so I am calling it a night, but let > > > > > > me > > > > > > know if you would like further clarification. > > > > > > > Best, > > > > > > Shawn > > > > > > > On Jun 24, 8:08 pm, savagelook <[email protected]> wrote: > > > > > > > >http://www.savagelook.com/demos/normals_align/sandbox.html > > > > > > > > I started on the alignment but I had a problem. For some reason > > > > > > > the > > > > > > > planes keep rotating along the normals even after adding the > > > > > > > lookAt() > > > > > > > call when I'm updating the linked planes. Can anyone who knows 3d > > > > > > > math better than me explain why this is happening, its in the link > > > > > > > above. Thanks, its driving me nuts and I can't figure it out. > > > > > > > > On Jun 24, 12:04 pm, savagelook <[email protected]> > > > > > > > wrote: > > > > > > > > > Shoot, I forgot to mention I didn't do the alignment yet. I'm > > > > > > > > gonna > > > > > > > > make another demo soon using planes to show the alignment part. > > > > > > > > Should look pretty neat with just the planes and an invisible > > > > > > > > source > > > > > > > > mesh. > > > > > > > > > On Jun 24, 10:44 am, Peter Kapelyan <[email protected]> wrote: > > > > > > > > > > Hmm it doesn't seem aligned to the face (those things should > > > > > > > > > be turning too, > > > > > > > > > no?). > > > > > > > > > > On Thu, Jun 24, 2010 at 9:00 AM, savagelook > > > > > > > > > <[email protected]>wrote: > > > > > > > > > > >http://savagelook.com/blog/away3d/away3dlite-normals-and-face-linking > > > > > > > > > > > normals and face linking in away3dlite. The face linking > > > > > > > > > > is not a > > > > > > > > > > rolled up class like FaceLink in Away3D since the > > > > > > > > > > modifications were > > > > > > > > > > minimal. I did make changes to Face.as to include a normal > > > > > > > > > > and center > > > > > > > > > > at creation time, so as long as you don't deform the mesh, > > > > > > > > > > everything > > > > > > > > > > should work. If you do deform the mesh (like with as3mod > > > > > > > > > > or manually) > > > > > > > > > > you would just need to recalculate the normal and center > > > > > > > > > > for each face > > > > > > > > > > affected. > > > > > > > > > > > One note, katopz mentioned that you could access a face's > > > > > > > > > > vertices > > > > > > > > > > through face.mesh.vertices, but I never got around to > > > > > > > > > > trying that. By > > > > > > > > > > doing that and maintaining your own vectors of face centers > > > > > > > > > > and > > > > > > > > > > normals, you could create this face linking code without > > > > > > > > > > making > > > > > > > > > > changes to Face.as, if you were so inclined. > > > > > > > > > > > On Jun 23, 11:37 am, savagelook > > > > > > > > > > <[email protected]> wrote: > > > > > > > > > > > Thanks for the insights katopz. > > > > > > > > > > > > For anyone interested I have just got face linking (on > > > > > > > > > > > steroids) > > > > > > > > > > > working in lite. All credit goes to the original > > > > > > > > > > > FaceLink author as > > > > > > > > > > > 98% is just copied code from there. The rest is just > > > > > > > > > > > minor > > > > > > > > > > > computational changes using Vector3D in the absence of > > > > > > > > > > > Number3D. The > > > > > > > > > > > code I have now requires minor changes to Face.as, but > > > > > > > > > > > I'm going to > > > > > > > > > > > see if I can also do it without changing the source and > > > > > > > > > > > maintain a > > > > > > > > > > > good framerate. I'll post it up once I have it all > > > > > > > > > > > ironed out. > > > > > > > > > > > > On Jun 22, 3:38 am, katopz <[email protected]> wrote: > > > > > > > > > > > > > Hey guys > > > > > > > > > > > > > it's for speed reason, dot access is slower, so private > > > > > > > > > > > > _vertices is > > > > > > > > > > just > > > > > > > > > > > > bypass that dot and acting like referrer, you can saw > > > > > > > > > > > > bypass like this > > > > > > > > > > every > > > > > > > > > > > > where in lite libs > > > > > > > > > > > > > if you really want to access vertices you can use > > > > > > > > > > > > mesh.vertices and > > > > > > > > > > provide > > > > > > > > > > > > faceIndex for each face, or mesh.faces[index] > > > > > > > > > > > > > for normal comment out there is just under development > > > > > > > > > > > > for light > > > > > > > > > > somehow, > > > > > > > > > > > > it's not in my field btw > > > > > > > > > > > > > hth > > > > > > > > > > > > > On 22 June 2010 14:10, Michael Iv > > > > > > > > > > > > <[email protected]> wrote: > > > > > > > > > > > > > > I talked to Rob Bateman about this issue a few weeks > > > > > > > > > > > > > ago in relation > > > > > > > > > > to > > > > > > > > > > > > > another class. He said that some properties became > > > > > > > > > > > > > private in order > > > > > > > > > > to not > > > > > > > > > > > > > be exposed for a regular user .(Still don't > > > > > > > > > > > > > understand a reason for > > > > > > > > > > > > > this).But you can still access most of these using > > > > > > > > > > > > > arcane namespace. > > > > > > > > > > > > > > On Tue, Jun 22, 2010 at 6:56 AM, savagelook < > > > > > > > > > > [email protected]>wrote: > > > > > > > > > > > > > >> While I'm still hoping for answers to my questions, > > > > > > > > > > > > >> I wanted to note > > > > > > > > > > > > >> that I uncommented the normals code in the faces in > > > > > > > > > > > > >> away3dlite and > > > > > > > > > > it > > > > > > > > > > > > >> works well. I also was able to use it to put basic > > > > > > > > > > > > >> face linking > > > > > > > > > > > > >> functionality into lite. I'll post it up on my blog > > > > > > > > > > > > >> soon if any is > > > > > > > > > > > > >> interested. > > > > > > > > > > > > > >> On Jun 21, 11:33 am, savagelook > > > > > > > > > > > > >> <[email protected]> > > > > > > > > > > wrote: > > > > > > > > > > > > >> > Also, why are the vertices of faces now private > > > > > > > > > > > > >> > with no getters or > > > > > > > > > > > > >> > setters? Why were mapping coordinates used to > > > > > > > > > > > > >> > calculate the > > > > > > > > > > normals > > > > > > > > > > > > >> > instead of the vertices like in Away3D 3.x? > > > > > > > > > > > > > >> > On Jun 21, 11:09 am, savagelook > > > > > > > > > > > > >> > <[email protected]> > > > > > > > > > > wrote: > > > > > > > > > > > > > >> > > I would also like to know why the normals are > > > > > > > > > > > > >> > > commented out in > > > > > > > > > > > > >> > > away3dlite, as of the latest trunk > > ... > > read more »
