If I want to model a complex product like a highly detailed digital camera, which one would give you smooth and the best result without any artifact in close up views, Nurbs modeling or Polygonal modeling?
The digital camera is just an example. There are other industrial objects such as car engine.
The matter is that I want to render photorealustic image, not manufacturing.
In many cases ceeating a good topology is very hard when there is lots of detail on the model and you need to use lots of supporting loops and these supporting loops damages other areas of the model when you add smoothing. (In polygonal method(
Either way can get you the results you want, but with Rhino being a surface modeller I’d stick to that.
Getting good visualisation is more about what you do with the models, how you texture and light them, how you compose your scene.
If you are going to do both close-ups and wide shots you’ll probably have different versions of your objects. For wide shots you probably don’t need high density geometry on the smallest of the smallest objects (or objects that aren’t visible anyway). If you are going to model a car engine and have a shot from outside that can’t see in you’d hide anything you can’t see.
Things to keep in mind:
- Use as little geometry as possible, but no less than the required details
- Use as small textures as possible, but no smaller than required
- Use as few lights as possible, but no fewer than required
- Expect to do post-processing of your renderers however final they are - you’ll be hard pressed to find any untouched professional photo or rendering in publications, so don’t try to do it all in one go
- Cheat where you can
- Rendering is cheating (non of it is real, it is all just numbers)