I've been giving this a lot of thought lately and I know that the opinions is probably mixed.... It's been the style of a lot of freestuff/marketplace providers lately to make their items with less geometric detail and putting most of their detail in textures. This lends to a greater ease of customization in the overall appearance, but it is limiting when it comes to out-of-the-box renders on some scenes, making postwork more necessary to adjust the painted-on light and shadow. In the minority, however, it seems, are the artists who choose to go with more of a high-resolution topographical style... namely using simpler textures but higher-resolution mesh topography and letting light and shadow create the details, using the textures to accent the modeled details. The downside to this of course, is that while there is less postwork involved to correct light and show, that it lends to less customization much of the time because the details are modeled in. Geom-swapping dials help greatly in this, but don't necessarily solve the issue. Neither is the right way or wrong way... they are simply different methods by different artists. I'm of the opinion that it all depends upon the item... what are your thoughts?

I agree, it does depend on the item, but because we do both, I still tend to go with detail on the model if it won't crash others computers. But that is a good statement. Sharen

I would personally like to have a hi rez mesh as i can always rework the textures to a higher resoulution. I dont really care for the tag line 'lower poly and or textures' to ease up on the system. I dont mind having a slower refresh rate when working on a scene, or a longer render time, because in the long run, its worth the quality. My last scene 'Point of entry' took forever for me to do as im only running a p3 866 with 512 ram, even though i did post-work that diminished any high quality textures or polys that you can notice.
Just my thoughts.
E

I am not a big fan of painted on details, they look fake in many lightning situations and are difficult to influence with material settings.
I think the ideal solution in many cases would be to add the detail with displacement maps, if necessary with the addition of specularity and reflectivity maps. That could give highly detailed models that look good in any lightning, are easy to customize and keep the polycount manageable.

Stefan

I'll take geometry over textures anyday.... :-)

YES!

A detailed model that allows for many material groupings would seem to combine the two. I'm with echo871, low res looks like tinker toys to me and is deadly when you move in close. Emily

I would only say this.....BOTH! yes both are important...but it's only from my point of view... Good modeling and good texturing is the key! Just remeber this...(in a way we are all the pioneer in the 3d digital world...One day, we will have a hard time to say what is real and what is digital) for this last said...I think really that both are important in a model...Good details...good UVs that leads to good texturing!

Detailed mesh... as another artist in the Vue forum discovered, raising the bump will never add detail that isn't there to begin with... As painful as it is, i've learned to use a detailed mesh instead of an expensive (in polys) texture. In Vue, there is danger zone above 1 million polys. I get to that point very easily, and have to start backing off. It's a fine line sometimes. Poser is an entirely different animal. The textures that you try out and don't use hang on forever if you don't save your character/prop back to the library. That adds up very fast.

I agree with Quinlor that painted on details tend to look fake. It's the reason I don't buy many texture packs. Whenever possible I model details into my mesh, and at one point I went WAY over what I consider high poly count but the darned thing still sold like hotcakes :D And I adore multiple material zones .. tis why I do it LOL

Well I must be the lone oddball here . I like lower res stuff , it's probably because I like doing post work . I think that alot of peoples models out there are higher res than they need . After all , do we really need 300 polygons for a flat wall segment when 1 would have done the same thing . I suppose what I'm trying to say is , put the detail where it's needed , not everywhere .

Yes. Meaning both...LOL. A good detailed texture on a good detailed model looks smashing, IMHO. Laurie

detailed model.. absolutely.. The textures can always be reworked best ofcourse is both.

I think Geralday is the master of getting amazing results with low poly figures. But... he always says up front that they're not intended for closeup renders.If you know that something's probably going to be more of a background figure then it's easier to get away with a smaller mesh and perhaps a better texture. Once you get closeup, it's hard to fake detail. Of course, in an ideal world, you'd provide a high and an alternative low resolution mesh. Give the low resolution one away for free and everyone's happy :-)

In an ideal world, all models come with three versions- a low-poly low-res texture version for long shots, a low-poly high-res texture version for medium shots and a high-poly high-res-without-painted-detail texture version for closeups.

Sounds like a job for a realtime dynamic polygon reduction and autotomated texture resampling engine tied into the distance of the model from the camera. I think its in the works for Poser 6, after Pixar buys out CL/EGISys :-)

LOL lmckenzie. :D

With what I do, I prefer to use the right amount of polys, no more and no less. Of course, I have no way of telling anyone what that amount is because it's something I just know when I make something. But one note: don't chamfer to make a sharp edge in Poser. Split the edge. Less polys AND a nicer edge. Of course, this is a problem when the split edge runs through two groups, because Poser will try to weld and make a smooshy spot. I can get around that, but it's a trade secret B^)

Actually, this bit of blurb from a laser scanning studio sounds very similar to what I proposed: ------------------------------------------------------------ "...an advanced algorithm for Dynamic Polygon Reduction. This algorithm, enables us to remove polygons (triangles) from a model in real time. Data is removed from the original model in an ordered fashion, removing the least important details first. All removed points remain in a "history" file, which allow the model to be "scaled" up and down indefinitely. ... We have seen models reduced from 80,000 polygons to 800 without any visual anomalies;..." ------------------------------------------------------------ I think the technology probably exists but putting it into a "consumer" app like Poser or having the horsepower to run it is something elae. Maybe a 4-way Opteron with a few gigs of RAM as Vicky X smoothly goes from a 250,000 polygon closeup to a Posette sized background figure. I like it!

Conceptually it wouldn't be all that hard, really, algorithm-wise. 1) always remove any polys not visible to an active camera (including shadow cams). No point in processing polys no one will ever see. 2) By interpolating the angle of vertices, remove any polys that would not come out at least 1px x 1px in the render and weld/smooth around them.

A good bump map can do a lot in combination with a good texture. I think it would be overkill in most cases, for instance, to model every rivet and line on an aircraft. Now a good nipple, say, that's worth blowing some extra polygons on. Gotta keep your priorities straight :-)