Re: [Flightgear-devel] Re: [Plib-devel] Re: Vertex Splitting, take two
Andy Ross [EMAIL PROTECTED] wrote: Melchior FRANZ wrote: So there's hope that the bo105 bashing comes to an end, finally? Here is a more recent version with less doubled vertices, but still a rough working copy with enough room for people to moan: http://www.unet.univie.ac.at/~a8603365/bo105.ac.gz Bashing? Don't be silly. It was an excellent test suite. :) and it looks really nice ! Martin. -- Unix _IS_ user friendly - it's just selective about who its friends are ! -- ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Re: [Plib-devel] Re: Vertex Splitting, take two
Melchior FRANZ wrote: Unfortunately, I've now run into a problem: while the model looks OK from outside, the cockpit view looks quite ugly, because all the rounded window frames come out too angular. I would have to use a lot of vertices to make this look good. I'd like to avoid a LOD cockpit for the pilot view. Hmm ... Can you post a screenshot? The model in CVS doesn't (I don't think) have any cockpit geometry. The default sharp angle is currently 46°, which means that an octagonal surface will look smooth after the split. So it should only take one extra edge line per 90° to make a smooth edge; that doesn't sound so bad to me. Maybe there's a bug, or the default is set wrong. In any case, the threshold is tunable per-subobject; we just have to find a way to express that in the model file (the AC3D 4.0 crease parameter being the obvious candidate). At worst, you could simply identify the objects that you want smooth and set their thresholds to 180°. Andy ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
[Flightgear-devel] Re: [Plib-devel] Re: Vertex Splitting, take two
* Andy Ross -- Tuesday 04 November 2003 23:02: Can you post a screenshot? The model in CVS doesn't (I don't think) have any cockpit geometry. No, but this one has (kind of; it's just a flipped copy of the fuselage): http://www.unet.univie.ac.at/~a8603365/bo105.ac.gz (40 kB) Here's the screenshot: http://www.unet.univie.ac.at/~a8603365/fgfs11.jpeg (20 kB) [...] it should only take one extra edge line per 90° to make a smooth edge; that doesn't sound so bad to me. No, that sounds good. I'll try that. At worst, you could simply identify the objects that you want smooth and set their thresholds to 180°. Still better than having to provide a high-res copckpit. m. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Re: [Plib-devel] Re: Vertex Splitting, take two
Melchior FRANZ wrote: Here's the screenshot: http://www.unet.univie.ac.at/~a8603365/fgfs11.jpeg (20 kB) Hrm; is that the right screenshot? The cockpit geometry in that shot just looks black. There's very little lighting at all, so I can't imagine how the vertex splitting code could make it look any different. Andy Ross wrote: At worst, you could simply identify the objects that you want smooth and set their thresholds to 180?. Still better than having to provide a high-res copckpit. Ah, I see the misunderstanding. The new code only affects lighting. It creates new vertices only to change the direction of their surface normals; it won't smooth out the visible outlines of the geometry for you. You're basically stuck having to provide a high-res cockpit; sorry. :) Andy ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
[Flightgear-devel] Re: [Plib-devel] Re: Vertex Splitting, take two
* Andy Ross -- Tuesday 04 November 2003 23:48: Melchior FRANZ wrote: Here's the screenshot: http://www.unet.univie.ac.at/~a8603365/fgfs11.jpeg (20 kB) Hrm; is that the right screenshot? The cockpit geometry in that shot just looks black. Yes, it's dark gray. :-) Still better than having to provide a high-res copckpit. Ah, I see the misunderstanding. The new code only affects lighting. It creates new vertices only to change the direction of their surface normals; it won't smooth out the visible outlines of the geometry for you. I thought that smoothing is automatically done everywhere, and that your code omits this for certain angles. OK then, I'll think about another solution ... it will probably make slower machines crawl ... :-/ m. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel