Difference between Cinematic workflow and low poly workflow??

Farina

Hello,

because of this community, i know now more about different workflows. Thanks!

I want to know now, what the difference is, between the high poly to low poly workflow and the cinematic workflow?

I mean..


For cinematic.

1. Base Mesh Zbrush
2. Sculpt
3. High poly
4. Duplicate High poly, zremesh and then project the High poly to the zremeshed, so we have subdivisions.
5. Uv Master and unwrap
6. Paint Maps (maybe normal and discplacement map)
7. Export to your program (Maya, 3dsmax)
8. Render with Mental Ray or Vray

For games, it should be from 1 to 3 and then

4. Retopology
5. Export Retopo to 3dsmax or Maya and unwrap
6. bake normal maps and other maps
7. Marmorset ;-)


Correct me please, if i am wrong. 
------------------------------------
The big thing is for me, i want to know the workflow for accessories too, but i am not sure. Is the workflow maybe for a jacket or a shoe the same like step 1 to 4, when we want to make a cinematic or game character?  
+


greetings

Farina

ActionDawg

Your steps are spot on overall, but they're very specific towards one type of work:  organic/environmental. Even for this type of stuff though classic polygon modeling like in 3ds or Maya can do just as well. Just keep in mind that a studio might not be using zbrush, mental ray, etc at all. Another example of a different way of working is Ptex, which doesn't use UVs to texture objects. One workflow is not always the right answer. You could sculpt a shoe in zbrush, but if you could model it in 3ds max faster then that's clearly the better answer.

I would say that more generally, the only differences between realtime and prerendered are:
1) In realtime everything must work from every angle and place the player can go to, whereas with prerendered it's just whatever gets the job done.

For example, I couldn't immediately find anything online about this but remember it pretty well from the DVD extras for Ratatouille. The head chef, pictured above, has a cloth that's tied around his neck. They had trouble simulating it correctly 100% of the time (their cloth simulation could glitch out, especially when under stressful forces like self-intersection which the guy's short stature could quite easily cause), so for some shots they would stretch his neck out extremely far and use forced perspective to get it looking and simulating right without the cloth freaking out. You would never notice this though because you only ever see the movie from one point of view.

2) Realtime has to fit into the constraint of, well, running in realtime. Prerendered is vastly more flexible. Hence why we bake out tons of texture information. The rendering engine can't compute that stuff in realtime, at least not at any quality that makes it worth doing (yet!)

I find it extremely helpful to just think of realtime art as being an approximation of what you actually want, and your goal is to get that approximation as accurate as you can. This is of course ignoring very stylized art. With this generalization you can restate your steps as...

For prerendered (cinematic) work:
1) Make a high poly
2) Make textures
3) Light and Render

And for realtime:
1) Make a high poly
2) Make a low poly
3) Make bakes and textures
4) Light and Render

The methods by which you accomplish those general steps is entirely up to you and the software you use.

Farina

thank you so much for your answer. I understand now, what you are talking about, because i have read it 5-6 times. It helps me a lot