|
|
 | | Re: duplicate primitives?
|
| (...) Sorry, it was a false alarm. There are no duplicated primitives. AFAIK. Steve (22 years ago, 4-May-02, to lugnet.cad.dat.parts.primitives)
| | | | | duplicate primitives?
|
| (...) ????? Just out of curiousity, what are these two? Thanks, Franklin (22 years ago, 3-May-02, to lugnet.cad.dat.parts.primitives)
| | | | | Re: ring 3 to 5
|
| (...) I understand that. But the question specified that the number of triangles (actually polygons, but feel free to assume triangles) is fixed. (...) So size of the polygons doesn't matter. Hmm. I'll have to remember that. Steve (22 years ago, 3-May-02, to lugnet.cad.dat.parts.primitives)
| | | | | Re: ring 3 to 5
|
| (...) I don't know about all situations (POV, etc), but for real-time rendering in OpenGL, the fastest rendering generally occurs with the fewest number of triangles (assuming that the triangles specify the same final geometry). The fact is that (...) (22 years ago, 3-May-02, to lugnet.cad.dat.parts.primitives)
| | | | | Re: ring 3 to 5
|
| (...) Not necessarily. We usually stick with units of 1 because they're the most easily scaled, and most easily implied (ie, if every ring has a n:(n+1) ratio of radii, we don't need to state both radii in the filename). OTOH, boxes have side length (...) (22 years ago, 3-May-02, to lugnet.cad.dat.parts.primitives)
| | | | | Re: ring 3 to 5 [DAT]
|
| (...) Very true. In some cases (like the minifig arms), is it better to go with an all-polygon approach, or to use primitives as much as possible and fill in the rest with polygons? (...) When filling in around a hole in a flat surface, one should (...) (22 years ago, 3-May-02, to lugnet.cad.dat.parts.primitives)
| | | | | Re: ring 3 to 5
|
| (...) There are some complex curved parts (e.g. most of the minifig headwear, the minifig arm I'm working on right now, some wheels) where some regions can be represented by the regular cyli, disc, cyls, cyls2 primitives, but there will be regions (...) (22 years ago, 3-May-02, to lugnet.cad.dat.parts.primitives)
| | | | | Re: ring 3 to 5
|
| (...) Yes, otherwise you get anomalies like this one: (URL) the outer ring of the tile got high-quality-ified but the fill in (non primitive?) circle didn't, leaving gaps. Also the plate below it, since it has stud cutouts, isn't as "round" as the (...) (22 years ago, 3-May-02, to lugnet.cad.dat.parts.primitives)
| | | | | Re: ring 3 to 5
|
| (...) True. But it's more than nothing. (...) LDraw (and LEdit, I assume) don't cache any files in memory. Read it (line by line), process it, and throw it away. (...) Reducing the number of files may not be important to rendering speed, but it is (...) (22 years ago, 3-May-02, to lugnet.cad.dat.parts.primitives)
| | | | | Re: ring 3 to 5
|
| (...) Actually, for rounded elements, it's almost always[1] preferable to use primitives. That way, programs that do primitives substitution can replace the polygonal primitive with a true round object. -- Steve 1) 'almost always' meaning, 'I can't (...) (22 years ago, 3-May-02, to lugnet.cad.dat.parts.primitives)
| | | | | Re: ring 3 to 5
|
| (...) However, when deciding whether or not to in-line sub-parts, reducing polygon count isn't a consideration. Any other advantage, small or not, should be maximised. (...) Whether they do or not, there's still extra time spent identifying the (...) (22 years ago, 3-May-02, to lugnet.cad.dat.parts.primitives)
| | | | | Re: ring 3 to 5
|
| (...) I think it is an extremely small advantage. What programs actually read any file more than once? I mean, Are there programs that actaully open the file and read the part in every time it's referenced? Or are there programs that even if they (...) (22 years ago, 3-May-02, to lugnet.cad.dat.parts.primitives)
| | | | | Re: ring 3 to 5
|
| (...) Sub-parts are handled differently to primitives (from an authoring view point). While primitives are generally not in-lined, sub-parts used during authoring are often in-lined, if it doesn't increase the total file size of the part too much. (...) (22 years ago, 3-May-02, to lugnet.cad.dat.parts.primitives)
| | | | | Re: ring 3 to 5
|
| (...) I can see how that wouldn't be desirable. (...) Correct me if I'm wrong (please!) But inlining only replaces the type 1 line with the (transformed) lines from the subfile it referenced. Right? It doesn't mean figuring out if 2 polygons could (...) (22 years ago, 2-May-02, to lugnet.cad.dat.parts.primitives)
| | | | | Re: ring 3 to 5
|
| (...) I'm not a mech. eng. major, myself, so my knowledge of this CAD stuff is just from my math. skills. That said... As I understand it, a "primitive" is supposed to be an *atomic* unit, something boiled down to its essentials. A disc/circle (or (...) (22 years ago, 2-May-02, to lugnet.cad.dat.parts.primitives)
| | | | | Re: ring 3 to 5
|
| (...) One thing you should be careful of, though. If you do inline it so that your new sub-part is used and doesn't reference any actual primitives, programs like L3P and LDView can end up displaying things with gaps. Both LDView and L3P perform (...) (22 years ago, 2-May-02, to lugnet.cad.dat.parts.primitives)
| | | | | Re: ring 3 to 5
|
| Hello, in this case i would agree tp Franklin's opinion regarding using two ring instead of creating a new primitive because if we create a new primitive section for this you will have very big number of possible combinations (ring 1 +ring 2, ring 1 (...) (22 years ago, 2-May-02, to lugnet.cad.dat.parts.primitives)
| | | | | Re: ring 3 to 5
|
| (...) It's all a compromise (just like life!). If you go creating primitives for every little sub-part that's used a few times, you end up with a primitive directory that's unwieldy & lots of parts which inline because they don't know the (...) (22 years ago, 2-May-02, to lugnet.cad.dat.parts.primitives)
| | | | | Re: ring 3 to 5
|
| (...) As a program author, I believe that yes this would result in twice the polygon count. While I aggree that programs should generally do everything they can to make authoring parts easier, I also think that Part Authors ought to keep in mind (...) (22 years ago, 2-May-02, to lugnet.cad.dat.parts.primitives)
| | | | | Re: ring 3 to 5
|
| (...) Wouldn't that result in twice as many polys that have to be rendered? Forgive if that's a dumb question, I ain't much of an author, you know. :-) (22 years ago, 2-May-02, to lugnet.cad.dat.parts.primitives)
|
|
