Yep, already got past the scene rotation issue. It's pretty easy to move the camera. I convert to polar coordinates and then back to get it to act like it is stuck to the outside of a sphere.
I'll bet you are on to something here with the camera. For the time being, I am just using the more common camera and it's working very nicely. I would like to get back to Orthogonal at some point, so hopefully someone will look into it. Cheers, BW On Sun, Sep 12, 2010 at 6:38 PM, Joshua Granick <[email protected]> wrote: > The last I saw the source for Drag3D, it does ask the camera to interpret > the coordinates, so if that step didn't work correctly for the > OrthogonalLens, that could have an impact. I'm using a standard camera, > currently, which I tween to new positions based on the perspective I need. > > The only problems I've experienced with Drag3D is that it doesn't work > correctly when the scene is rotated (I know, I should move the camera, not > rotate the scene) and if you are near the plane you're measuring against, it > reports Infinity instead of -Infinity when you move you mouse out far enough > > > > > On Sun, 12 Sep 2010 14:04:56 -0700, Bob Warfield <[email protected]> > wrote: > >> On Sun, Sep 12, 2010 at 11:07 AM, Joshua Granick >> <[email protected]> wrote: >>> >>> I'm currently using the Drag3D class. It has a getIntersect () method >>> which >>> should return the coordinates in most circumstances. I don't think you >>> need >>> to be dragging an object for it to work. >> >> Joshua, that Drag3D's pretty good! >> >> But, some funny business of various kinds that I wonder if you have >> some thoughts about. >> >> If I just bring my app up and move the mouse around, I am updating >> the coordinates on my status bar and it looks great. I have a visible >> coordinate grid I can use to calibrate off of for these tests. >> >> The one thing that is very odd is I have to divide the coordinate >> values that come back by 8 to get the right answer. That seems weird. >> FWIW, I am running the orthogonalLens for my application, rather than >> the default. I have noticed a pretty fair number of things in Away3D >> are a touch wonky with that lens and I wonder if that is the source of >> the 8 or my other problems? >> >> But things go much more awry if my camera changes position. It's as >> if the coordinates do not reflect the camera moving, which they >> should. >> >> For example, my application has "Top", "Right", "Front" and >> "Perspective" views that I can toggle. In fact, the display is sort >> of a replica of Rhino3D, but looking at one view at a time instead of >> all 4. In all the views except "Perspective", I pan the image when >> you hold the right mouse button down. This is implemented by running >> the camera along as though it were on tracks that stay a fixed >> distance above the objects being viewed. >> >> So, for example, a mouse move in X with the right button down will do >> this in the "Top" view: >> >> camX = setCamCoord( >> camX+moveX ); >> camLookX += moveX; >> >> This just moves the camera's location and where it is looking by an >> amount proportional to how much the mouse moved in X--"moveX". Feels >> good, works great. But after I make a move like that, the coordinates >> coming back are now off by approximately however much I moved. If the >> camera moved, why doesn't Drag3D get updated too? >> >> FWIW, I tried switch to the other camera types. First thing, is I no >> longer have to use that correction factor of 8. Second thing is the >> panning now works properly. So something is wonky with the >> OrthogonalLens. Sounds like a bug to me. I'd like to be able to use >> it, because the perspective effect of the other lenses is a bit much >> for me on this app. >> >> Any thoughts on this would be much appreciated. >> >> Also, I'm going to want a call that does the getIntersect() while >> letting me pass in x, y (pity this one doesn't have default values for >> the parameters of the mouse's position), but I >> have already hacked my copy of Drag3D to work using the default >> parameter style. Seems like others might like that change too, >> Fabrice. >> >> Thanks in advance for helping! >> >> Cheers, >> >> BW >
