On Sunday 29 June 2003 18:19, WillyB wrote: > > A couple of things I noticed though: The blue area of the texture you're > > using has a very slight gradient to it - was this intentional? While > > you're trying to sort out geometry and smoothing problems, using a > > texture with a subtle gradient could make things a bit confusing. > > Yes, that was intentional, so I could see about how it was being mapped on > the model.. also that is the reason for the white square.
A good tool/texture for that is a colour grid - start with a blank square texture and divide it into 8 differently coloured horizontal stripes. Then divide the square vertically 8 times so you have 64 boxes. Use the text tool to put a number in each of the boxes in each stripe, from 1 to 8 so you end up with horizontal colured stripes and eight columns of numbers: 1-8. When you map the texture to a model you'll be able to see where red-3, green-7 or blue-4 is. It'll also show you how the texture is being applied to the object and whether it's being distorted. Generally, when texture mapping, you'll want to use a parallel projection, as opposed to cylindrical, spherical or default(whatever that is for that specific package). This will give no distortion when you look at the texture from the projection point but it will when you look at it from another angle - think of parallel mapping a series of stripes onto a cylinder from a horizontal view and then look at them from a top view. > > > It looks like the texture is being mapped to each poly and not to the > > model as a whole and that's why it seems to be repeating. It's > > interesting because it shows the point order for each of the polys - > > unless you've specifically mapped each poly that way. > > Nope... that was not intentional. Didn't think so;) > > > There's some strange smoothing going on too - most of the polys seems to > > be getting smoothed independently of each other i.e. the wings and left > > tailplane in pic 12 compared with the right tailplane in this pic. > > I think now from reading other replys it's because of the normals not being > right after I shift-D and then S X to mirror or reverse it. Yeah - this could be 'normal' problems. Normals, on their own, shouldn't cause these problems. But they seem to. > > > Smoothing is done when there are two polys joined by a common edge, or > > even a common vertex but the smoothing artifacts you're getting on the > > wings, for example, don't appear to be due to smoothing across the polys > > making up the surface of the wing - instead, it looks like each poly is > > getting smoothed individually. However, this doesn't happen - if you > > have a single poly there is nothing to smooth it with so it'll just > > appear flat. I'd guess that each of these polys has another poly > > attached to it but it isn't the one adjacent to it - two-sided perhaps? > > Hmm.. well.. I started with a circle and then deleted half of it.. then > extruded the rest from that. So blender did what it did to it.. I did not > do anything like purposly making it two-sided or anything. It's strange because each poly isn't simply flat shaded, as it would be if it wasn't connected to another poly. The artifacts are there because the renderer thinks the the triangle is connected to another triangle and the boundary/join between the two needs smoothing. If there are a couple of triangles in a single four-sided face that are orientated upside down to each other then that could be a problem but I can't see how that would have occurred from extruding from a single object as all the normals would be correct to start with. Great fun eh? ;) > > > It's because of the way that this smoothing works that I make the wings, > > rudder and control surfaces in upper and lower (or left and right) halfs > > with separate 'cap' objects to prevent the renderer from trying to smooth > > 'hard' or sharp edges such as the wing trailing edges etc. > > > > Something to watch out for if you do this is that FG doesn't seem to like > > objects consisting of a single poly so if you make a simple upper aileron > > surface using a single four-sided poly, you'll have to manually split it > > into two triangles before FG'll be happy with it. > > > > Another thing to watch out for is using polys with greater than four > > sides. Once the model is loaded into FG all the polys with > 3 sides > > appear to get triangulated and where you've made half of an object and > > then mirrored it, the polys in each half of the mirror can get > > triangulated differently and you end up with a glitch on just one side. > > The only way I've found around this is to identify the polys concerned > > and manually triangulate them. > > > > LeeE > > Thanks for the good points and tips :) > I've added it to my file on how to model for flightgear. > > Re's > WillyB > I've been doing 3d stuff for over ten years now but I've still learn't something new with each of the a/c I've done for FG. I wouldn't say that I'm certain about anything in 3d but I'm certainly less uncertain than I was;) ...I think. LeeE _______________________________________________ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
