+- LDraw.org Discussion Forums (https://forums.ldraw.org)
+-- Forum: General (https://forums.ldraw.org/forum-12.html)
+--- Forum: Official File Specifications/Standards (https://forums.ldraw.org/forum-32.html)
+--- Thread: TEXMAP Fallback (/thread-26660.html)
TEXMAP Fallback - Orion Pobursky - 2022-08-29
Currently a fallback pattern is not required for TEXMAP parts. I've been musing making it a requirement that a fallback be used even if it is lower quality than the TEXMAP image.
Thoughts?
RE: TEXMAP Fallback - Philippe Hurbain - 2022-08-29
(2022-08-29, 17:14)Orion Pobursky Wrote: Currently a fallback pattern is not required for TEXMAP parts. I've been musing making it a requirement that a fallback be used even if it is lower quality than the TEXMAP image.
Thoughts?
Not so keen about that. Most (actually I think all) editors currently in use are able to display Texmap, so what would be the point?
RE: TEXMAP Fallback - Orion Pobursky - 2022-08-29
Backward compatibility mostly. Also, support for the full spec (including GLOSSMAP) isn't consistent.
RE: TEXMAP Fallback - Roland Melkert - 2022-09-01
(2022-08-29, 18:29)Orion Pobursky Wrote: Backward compatibility mostly. Also, support for the full spec (including GLOSSMAP) isn't consistent.
Having a fallback is kind of automatic, just use a TEXMAP block without the FALLBACK parts and don't use "0 !:" in front of lines.
It will be just a solid color mesh, but it prevents gaps on non texture supporting software.
If such a fallback is what you are going for I'm fine with making it mandatory.
But I don't think you can demand a "vector art" fallback to be mandatory.
RE: TEXMAP Fallback - Orion Pobursky - 2022-09-01
(2022-09-01, 17:27)Roland Melkert Wrote: It will be just a solid color mesh, but it prevents gaps on non texture supporting software.
A fallback is currently required but only as you state above. I was thinking more along the lines of traditional "vector art" as you put it. If we don't want to do that (which appears to be the case, even in my opinion), then maybe we should define some minimum picture resolution. Basically I'm wondering if we should to set some sort of formal standard for pattern resolution while TEXMAP parts are still few in number.
Also: does LDCad support GLOSSMAP?
RE: TEXMAP Fallback - Roland Melkert - 2022-09-01
(2022-09-01, 19:48)Orion Pobursky Wrote: Also: does LDCad support GLOSSMAP?
Not at the moment, as it currently uses fixed pipeline OpenGL.
That said it is very doable to change some things around to (also) support native shader oriented rendering (as an HQ mode or something).
So I might add it in one of the upcoming 1.7 beta versions.
RE: TEXMAP Fallback - Orion Pobursky - 2022-09-01
(2022-09-01, 20:04)Roland Melkert Wrote: Not at the moment
Well that answers the question of what LDraw programs support it: none.
RE: TEXMAP Fallback - Magnus Forsberg - 2022-09-01
(2022-09-01, 19:48)Orion Pobursky Wrote: ...., then maybe we should define some minimum picture resolution. Basically I'm wondering if we should to set some sort of formal standard for pattern resolution while TEXMAP parts are still few in number.
When I made the globe I also played around with a rudimentary fallback pattern.
In order to keep a good surface for the texmap, there needs to be a good match between triangles and condlines.
A pattern line across a condline will result in a non-smooth surface and a bad surface for the texmap.
RE: TEXMAP Fallback - Orion Pobursky - 2022-09-01
(2022-09-01, 20:19)Magnus Forsberg Wrote: In order to keep a good surface for the texmap, there needs to be a good match between triangles and condlines.
A pattern line across a condline will result in a non-smooth surface and a bad surface for the texmap.
Upon reflection after writing the above, I'm not sure any quality standards can be codified and this is probably best left to the review process.
RE: TEXMAP Fallback - Travis Cobbs - 2022-09-01
(2022-09-01, 20:19)Magnus Forsberg Wrote: In order to keep a good surface for the texmap, there needs to be a good match between triangles and condlines.
A pattern line across a condline will result in a non-smooth surface and a bad surface for the texmap.
That's only true if you don't use texture-specific geometry, and don't have an actual FALLBACK section in the TEXMAP section. When the vector geometry is complicated, it belongs in a FALLBACK section that doesn't get rendered at all when TEXMAP is enabled. The geometry that gets textured would then be done via 0 !: lines.
So, if the vector geometry is really simple (or nonexistent), then the part can have something like this:
Code:
0 !TEXMAP START <params>
<simple shared geometry>
0 !TEXMAP ENDThe <simple shared geometry> section will render in both TEXMAP-supporting renderers and renderers that don't support TEXMAP. On renderers that support TEXMAP, the texture will render over the geometry.
In my opinion, as soon as there is any vector geometry, the following setup should be used instead:
Code:
0 !TEXMAP START <params>
0 !: <simple textured geometry>
0 !TEXMAP FALLBACK
<complex vector-only geometry>
0 !TEXMAP ENDThis means that only the simple geometry will be rendered when TEXMAP is enabled. Even with your example with the simple globe, I would argue that it should be done this way, and not draw the TEXMAP over the vector geometry, even though that vector geometry is relatively simple.