[Leonardo]

˙˙˙ooʇ ʇı ƃuıddılɟ unɟ ǝɯos ǝʌɐɥ sʎnƃ noʎ ʇǝl ı 'ʎɐʞo




http://www.revfad.com/flip.html

[devista]

sınl
sopnlɐs
(-؛ oǝl sxɥʇ

[Leonardo]

(: ɯǝlqoɹd ou

[KurtS]

˙˙˙ǝƃuɐɹʇs ʎɹǝʌ sʇɐɥʇ

[morbid angel]

i think maxwells approach to rendering will become the new norm in digital production pipelines. for a long time renderman has been the render of choice. (i know i love it). it been at the forefront of being able to breakdown complex shot in a highly optimized way. normal maps, per object render attributes. incredible shadow and other map based optimization techniques. the list goes on and on. but also being able to hit render and know you don't have to worry to much about it crashing.

but now comes transformers. a movie where the only way to complet the shot was to build and render all the highy detailed geometry. not baking out high res on low res, but full blown 10,000 objects per autobot. this is not what renderman is really good at. it took 38 hours per render on one of the most advanced farms on the planet.(don't you wish you had 4,300 processors just for rendering)

i think as computers get faster and movie push harder we will see shift from 'faked' geometry to actual geometry. and maxwell see's little decrease in speed compared to other render engines as scene complexity rises (i find maxwell takes 38 hours weather you have 100 or 10000 objects). i think this is the right way to go in terms of render technology.

my thoughs anyway.

reyes + raytracing = bad combination