[otacon]

1.5

[Jun In Gi]

i can't find any different thing from fprime.

Though it's a relatively *good* scene for fprime.

i agree Sinik

[Micha]

Upps ... 62 and 5 hours ... and no big differences in the look ...

[psanitra]

i`m impressed by the Fprime render! nearly the same result in 1/10 of time!!

[hdesbois]

very interesing comparison, and excellent images, as usual. Fprime light is harder and lacks the warmth of Maxwell's, but it's still a good render, specially considering the render time. Fprime is a great companion to Maxwell. You can really fine tune the DOF with it.
HD

[sidenimjay]

i am not noticiing DOF in the Fprime render am i wrong?

if there is no dof, then the test is nowhere near fair as dof calculations are intensive and i would be willing to bet that it would at least add a couple more hours to the fprime render . . . .

what did the maxwell render look like after 5 hours? after 10?

was there a point along the 62 hours that it was pointless to be rendering any further? i mean did an extra 32 hours make a noticable difference ?

plus the aliasing on the railing is much more harsh than in the maxwell render . . .

[otacon]

The colors are better in the maxwell one, its hard to try and simulate the nice physical sky in lightwave. There is dof in the fprime render. The aliasing looks worse because i sharpened it a little too much in post, originally its softer. I also had to adjust the color because it was too yellow. The maxwell one had to go the 62 hours to look decent, i was testing it every so often to see when i could stop it but there was too much noise. There are certain scenes that could look good in fprime and maxwell, but there are other scenes that would not look so good with fprime. Mainly because the way it handles light bouncing, corners tend to get too dark.

[jep]

interesting stuff otacon - thanks!

[lwan]

i am not noticiing DOF in the Fprime render am i wrong?

if there is no dof, then the test is nowhere near fair as dof calculations are intensive and i would be willing to bet that it would at least add a couple more hours to the fprime render . . . .

NL said it don't matter if there DoF or not for rendertime (same for MB if i remember)

[otacon]

If im not mistaken with maxwell you will always have dof, its just whether or not it will be visible, and thats based on the fstop and focal length of the camera.

[rivoli]

If that's the case, try rendering a scene with DOF and without it

as jason says, maxwell cameras act as real world cameras: you'll always have a depth of field and a focal point (even if in most cases dof is so deep that you can't actually find anything blurred at all).

[sidenimjay]

stronger dof may add time during the voxelization as does motion blur


but if you add dof to mantra(houdinis renderer) or renderman, you must up the sample counts to resolve the dof. this takes forever which is why we use depth mattes to do dof post in most cases


that was my point earlier , to be fair , fprime render or any other render test compared against maxwell should use dof else its not apples to apples . . .