Lighting and Rendering setup advice?
polycounter lvl 7
Lately I've been getting back into 3D and find I am really rusty in a lot of areas that I just never got enough practice at to begin with years ago, lighting and rendering being a big one. I've really started to notice that if any one part of the process is sloppy or forced, it just makes the overall piece and myself as a reflection look amateur.
Anyways, been reading a lot of stuff like Polycount's wiki on the topic and doing lots of google searches on lighting and rendering techniques while I get ready to do up my models fresh and figured it would be a good idea to ask how yous guys usually set up your lightboxes and do renders. Do most people still do 3point or something else? Skyboxes? Mental Ray or Marmoset or Xoliul or something else?
Any advice or pointers is welcome!
Anyways, been reading a lot of stuff like Polycount's wiki on the topic and doing lots of google searches on lighting and rendering techniques while I get ready to do up my models fresh and figured it would be a good idea to ask how yous guys usually set up your lightboxes and do renders. Do most people still do 3point or something else? Skyboxes? Mental Ray or Marmoset or Xoliul or something else?
Any advice or pointers is welcome!
Replies
Theres the 2 sides,
1. Offline
2. Realtime
Some people only present their models in realtime environments because they think its the true way to go, dont know about offline rendering or think the rest is cheating or all of those 3.
If you are one of those, you may want to use presentation techniques like xoliuls shader for 3DS which is proabably the easiest way. Just render in viewport, or you may want to sit down and get on marmoset. Its pretty easy to use and gives good results. Then there are DirectX realtime modes of 3D programs and stuff.
The other side is offline rendering, there are classical scanline renderers every 3D package has, or popular standard mental ray in 3DS and Maya. Or even vray, maxwell and whatnot. It takes some time to surpass the results you would get in marmoset initially Id say, but since theyre a 10x more complex, you get way better results in the end. But they have a huge learning curve. Those are for the people with the mindset that what you do is art and should be presented in the best possible way, maybe even with post production.
Personally I use Cinema 4Ds physical renderer which is offline ofc. I like to make
the best possible presentation, but its a lot of work you may not be willing to do in offline. But for me, I love rendering, its the most fun part of the pipeline for me, and cinema 4D is just a lot of fun to use. (Im a presentation guy tho, not a real modeler, more of a graphic designer)
The upside of offline knowledge is that you can do amazing shots of highpolys with procedural texture even, without laying out an uv and all that stuff.
About marmoset, in my opinion marmoset is also not as true as people claim, it uses image based lighting which is only used in a hand full AAA car games and simple offline lightsetups would work in realtime aswell if they had the DX api with the right settings or look very similar atleast.
The classical thing to do is a 2-3+ point setup, rim light maybe, fill light and spot. Or one on top one below. Just experiment.
Remember to color them complementary. If you do offline, get a HDRI lighting.
Its probably the best and easiest to just use marmoset, it dosnt bring a big new factor into your workflow.
Also, care about your backgrounds. Grey backgrounds are nicer the darker they are.
Flat color is always worse than simple gradients. (preferably a radial one with the lightest point in the middle)
Avoid harsh gradients at any cost. Grey to white, black to white, stuff like that is awful to look at. Heavily desaturated color in a gradient is the best and easiest way to go imo. Get a complementary to your texture, and gradients are best when you barely notice the color difference.