In rawkstar’s case I guess you would. However remember that he only zbrushed two elements from the max scene.
In 3d Studio Max you can choose element mode. Something Maya severely lacks and something every 3Ds Max users can not do without.
I choose a very broad selection of elements, like “Boots & Straps” or “Torso & Head”, and I seperate them into their own mesh. If I choose to zbrush any of those elements I now only have to bring in very little from Max.
As I mentioned earlier, seperating your high resolution geometry into catagories, helps when projecting the normals onto unwrapped low polygon geometry.
This is because when projecting such a complex model onto it’s crude low polygon counter-part can result in 2 or more elements sharing the same low polygon face. To prevent such an overlap from happening, ideally you want all your high polygon elements to have a low polygon structure built around them in its form.
Unfortunately we are building for games that require ruthless optimisation. The obvious and visually seperable parts of the high resolution geometry get projected onto their very own low poly element as other less obvious geometry may share the same low polygon face. It’s about balance and clever selectivness when choosing what it visually more important to the viewer.
So to wrap up what I said, once you have all of your low polygon geometry built around your high resolution mesh, its time to break the models up into peices. The high polygon model’s peices match the low polygon’s peices respectively. And when it comes time to project a high polygon peice onto it’s low polygon peice theres no confusion in which low polygon peice it’s going to project onto. Therefore resulting in minimal or no artifcacts in your final normal map. Of course doing things this way means you havfe to composite or stitch your various normal maps together into one or two large normal map which is why its better to model - unwrap and then sererate your model into groups and catagories. You see your UV’s must be created, given their own space in 1:1, and must not overlap, before you project normals from your high poly peices onto your low poly peices.
Its very simple. I may have said the same thing four times I apologise for this, but one must really take four or five simple steps to assure themselves a perfect procedure. You save memory, not to mention heart break, and you also acheive a greater graphical finesse with your final normnal map as it’s detail doesnt overlap or blend over two course low polygon faces.
I hope you follow. I am in a bit of a daze myself.
Must be the coffee.
Let us know what your method you adopted, rawkstar.
I think we’re all watering to know how that normal map of yours ended up being created. Did you stitch elements of your model up into one large normal map as I have done?
Cheers folks.
