To feather or not to feather, ...


#21

I’m not assuming any “only”, I’m suggesting option that seems to be better in terms of performance/quality balance under my best knowledge and without intention to mislead anyone. As you seem to agree on 3delight’s usefulness here, there is no reason to disagree.

The fact that you even mentioned 32bit is a clear indication of the problem, which force people to try rayes engines while developing feathers-like shots. You can of course do feathers in old LW 6.0, but that’s not a topic of the discussion (I supposed), possibility it’s not an usability.

Yes, I’m aware of rays in reyes ;), as I’m familiar with rasteriser. The point is that while mray looses its temper without casting bunch of rays, rayes by nature, are still fully capable. How would you use any of mray advanced shaders without casting rays? Internal scattering? Without mental images guys behind you, that would be difficult.

You don’t see too many Mantra renders for a known reason of small availability of it in studios ;). Its strength in respect to our topic is quite similar to the rest of the reyes crew. It can do, similar to 3delight, nice fur without a single ray, what makes a difference.

After all, I agree, talent is most important… and pertinacity for that matter :).


#22

Sorry, but I’m still, sincerely, really not understanding what you’re saying.

If we want to discuss REYES as an architecture we can, but the main point here for what we’re discussing is first and foremost rasterization, which is one of the principles at the base of REYES renderers (not RAYES btw). and which is available in MRay too and has been for a while now.
The rasterizer in MRay is pretty much an engine within the engine in fact, and has little to do with what people mostly think MRay is like normally (a pure raytracer, or a hybrid with a strong bias, which it isn’t anymore, if ever it’s been). It also comes with its own issues (many of which similar to rman standard engines), but with some of their strenghts too, like a better scalability.

As for feasibility, the comparison with LW6.0 doesn’t stand, not even hyperbolically.
MRay is perfectly capable of rendering millions of MI_HAIR primitives with the rasterizer. It’s been done in production without needing MI’s direct support, and it’s not even that hard.
And if you think that PRman can render millions of l-sys like ricurves for feathers, for production shots, inside a 1GB of ram footprint, then sorry, but you’ve never tried, because it really won’t :slight_smile:
If you know otherwise please share, because I’m sure people here and elsewhere would love to save the money it costs to upgrades several hundreds of blades to 8 or 16GB of ram per blade to render the bloody things :stuck_out_tongue:

The rest about mental images guys having to be behind someone, mray casting bunch of rays and losing its temper, and not being able to use advanced shaders without doing so is beyond trying to make sense of, so I won’t try :slight_smile:
Far from me wanting to defend MRay, but there’s a limit to the fluff I can see posted about :slight_smile:


#23

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.