By Half Life - Mon Jan 10, 2011 12:49 am
- Mon Jan 10, 2011 12:49 am
#335966
I'm glad somebody else is spending lots of time considering this besides me
All good points -- however the overwhelming majority of the data that is needed is being gathered (and has been gathered) in the name of science via satellites for a while now and could be re-purposed for rendering with a bit of work. Alot of the work being done for Google Earth proves this (satellite imagery/terrain data getting more detailed and near real-time weather to name a few)
As to the application of said data -- I can think of many, just not necessarily in our current workflows... and with our current hardware processing power it may not be feasible yet to be crunching complete/unbiased real-time atmospheric data. But I do think it's coming sooner rather than later -- the obvious question would be is it Next Limit that will do it first? It makes sense they would considering their heavy emphasis on physics/math and that fact that all their products are based on real world simulation with as few "cheats" as possible.
Once it's in place I imagine it will be alot like SSS or dispersion -- not something used on every project but certainly a difference maker when you want to push boundaries... and when the processing power arrives then it will be something used often.
But truly, to take the physcial sky settings we have now and make them date/location automatic wouldn't really require much -- the addition of more parameters, clouds, stars, moon, etc would be the things that would be more long-term additions.
I like Maxwell, not because it is the fastest renderer, or it makes the prettiest pictures or even has the most features -- I like Maxwell because I can rely on the correctness on the output if I input good data/geometry/settings. I know that isn't why many other people like Maxwell but I still think it may be one of the most important differences from most of the rest of the huge array of alternate rendering choices out there.
Best,
Jason.

All good points -- however the overwhelming majority of the data that is needed is being gathered (and has been gathered) in the name of science via satellites for a while now and could be re-purposed for rendering with a bit of work. Alot of the work being done for Google Earth proves this (satellite imagery/terrain data getting more detailed and near real-time weather to name a few)
As to the application of said data -- I can think of many, just not necessarily in our current workflows... and with our current hardware processing power it may not be feasible yet to be crunching complete/unbiased real-time atmospheric data. But I do think it's coming sooner rather than later -- the obvious question would be is it Next Limit that will do it first? It makes sense they would considering their heavy emphasis on physics/math and that fact that all their products are based on real world simulation with as few "cheats" as possible.
Once it's in place I imagine it will be alot like SSS or dispersion -- not something used on every project but certainly a difference maker when you want to push boundaries... and when the processing power arrives then it will be something used often.
But truly, to take the physcial sky settings we have now and make them date/location automatic wouldn't really require much -- the addition of more parameters, clouds, stars, moon, etc would be the things that would be more long-term additions.
I like Maxwell, not because it is the fastest renderer, or it makes the prettiest pictures or even has the most features -- I like Maxwell because I can rely on the correctness on the output if I input good data/geometry/settings. I know that isn't why many other people like Maxwell but I still think it may be one of the most important differences from most of the rest of the huge array of alternate rendering choices out there.
Best,
Jason.
My Video Tutorials: Maxwell Render 3 Materials