RE: Lies' Project

Hi Lies, 

[Sorry this post is a bit long, but most of it should be relevant for anyone else who is starting lighting for the first time.]

I love your project and its great that you have a clear idea of what you want the end product to look like.  Your reference is beautiful, your detailed model should pick up the light nicely and make a really interesting image.

When shading the hardest part will be the skin and hair, if you have access to subsurface scattering, use it, but keep it subtle, you have a strong key so it should work well.   You might want to work on lighting and shading at the same time, the subsurface settings will probably need tweaking for your shot, once you start lighting it.

As your new to lighting, the first thing I would say is keep it simple and do one thing at a time. Try to get the look you want in 3 lights, persist with (rotating,and moving around) just those lights until you are completely happy, this might take some time if your new to it.  

I know this might sound a bit odd, but I think of it a bit like playing snooker or pool :)

at first getting the angles can be a struggle, but you develop an instinct for it after a while. 

Remember to get your camera locked off before you start lighting so that you know what direction you are going to be viewing your model from and can light it accordingly.

Think KEY -RIM - FILL.  Render a frame through the camera for each light, turning off the others so you have a clear picture in your head of the area that light is illuminating.  Think about what type of light you want to use, it might seem the obvious choice to use a point light for the lamp, but perhaps it would be easier to light her separately (not with the lamp model or is this a matte painting?) using a spot light?  You get a clear light direction compared with a point light, the results might look better.

Ignore colour, intensity and shadows at first, just stay with white lights set at intensity 1.  Get the light directions right first.  You have the perfect image to match in your reference.

Then work on the intensity, obviously the key light from the lamp with be the main light and should be the only shadow casting light.  You will need to tweak shadow settings to get what you want, if your using raytracing you will need to increase the number of rays, or using shadow maps its the resolution, filter size, shadow blur, and bias.  In your case (as you are rendering a still image I presume) it would make more sense to me to use a depth map shadow as you can save out the map and re-use it.  This should save you time when doing test renders.  In general in lighting (particularly film industry), not using raytracing is a good thing unless you really have to :)  (this is my renderman background coming out)

If you feel you need more lights, add them one at a time, carefully considering if they are really necessary.  For this project I would steer clear of using global illumination and other light simulation programs  (like final gather).  We can talk about occlusion later, but that will help in the final stages of lighting to bring it all together (this is the one occasion you will have to use raytracing).  Personally I would render it on a separate pass and use it as a multiplier in the comp rather than having it in the main render.

Colour is the last thing you will be looking at, and its not as simple as you might think.  Quite often you hear people say, can you remove that green or that needs to be less magenta, when you have nothing in your lighting to suggest either of those colours. It is important to be aware of perception.  A lot can be learnt from experimentation so I would not worry too much about it now unless you come across a problem.

It seems to me like you can do a lot in comp to match your filmic reference with nice glowing yellow/oranges in the key light, and a strong contrasting vignette.  Don't worry about getting a perfect match in your main render, and try to start compositing early.  You might spend ages trying to fix something in lighting which takes 10 seconds to fix in comp, its worth experimenting with your early renders to see what you can do with them, and evaluate/designate what you want to do in 3d and 2d.

That should be enough to get you started.  Because I have been in industry a few years it is hard to remember what is obvious and what isn't, so feel free to ask questions if you want anything explaining in more depth (and sorry if any of it is patronising).  Most of my personal experience is based around trying to find an optimised workflow, which is less necessary now that computers are more powerful, and companies have huge render farms, but I believe it is still good working practice, and will make you a more valuable employee when you graduate.  

Looking forward to the next post.

Tessa