[taumel]

But just because something is theoretically possible, doesn't make it automatically likely nor reasonable. As i already suggested, choosing the right way for your GameEngine involes a couple of questions you first have to answer to yourself according to your specific project/situation/strategy.

You can write you own renderengine, you can licence one, you can use the one which comes with the gameengine, ...

What's suprising about that? It's the same like for instance with sound. If you can afford it, you will most probably go with a FMOD licence, if you can't you'll use BASS, openAL, ... write you own audio lib whatever. It's the same thing with a lot of components you need for a game (graphics, sound, ai, physics, ...) as long as your game engine is capable of integrating such components.

Otherwise you might just use such middleware and write the GameEngine on your own or you can't integrate such things because your engine wasn't designed for such purposes or it would mess up your engine's workflow. There are many possibilities why you can and why you can't do such things. But if you don't want to write a certain component on your own due to some of the reasons we have already talked about and can afford the licence, then why not?

a) There does exist a mature crossplatform soundengine? Let's see what it costs.
b) We need some more performance and are in need of some advanced occlusion culling. Let's take a look at Umbra.
And so on...

Theoretically you could enhance a Quake I engine to a cryEngine 3 but it would be more a complete start from scratch. As this is a money driven business you'll want to go the way which serves you the best.

Of course you can licence render engines. With shaders the days of fixed function piepelines are over since years, at least on the desktop. Normal mapping is a concept. You're free to implement it on your own (it's all just shadercode) and it has to wok together with the rest of the lightning system. Beside of this, a renderengine is a bit more than just a pack of shading possibilities.