The main thing that I want to see improved is the Blender scanlines process in which the renderer interprets photoreal lights, surfaces and shadows. We want photoreal features like Yafray but without the waiting. What steps are taking place now in this, area?
If you want a photo real render you will have to wait.
This is what I have read- Photo real renderers use complex maths therefore you have a slower render.
I think that blender's scanline render could...
- Use less memory, shadowmaps seem to chew up ram.
- Output high colour depth images - 48bit, since the renderer is floatring point anyway.
Render at a higher res then 10000x10000
But thats all. Im pretty happy with it.
Scanline renderer improvements?
the reason for "the waiting" as you refer to it is that when using high definition raytracing or/and radiosity light calculations (ie photorealistic renders) you are to simplify things "computing a lot". the only way to reduce "the waiting" and maintain the same quality is to not recalculate the lights, shadows, reflections for every frame. this makes sense in say architectural scenes where no objects are moving. actually blender can do that today (and has done so for quite a while now), for radiosity simulations anyway. sadly enough the workarounds for making the radiosity information stick to the scene is that all the materials used on all of the involved objects get baked in to a solid object, that means that even though you can get quite pleasing results portraying buildings or cad objects on display you cannot use reflections in conjunction with radiosity. at least not using the same light simulation.
