[itsallgoode9]

in the past I have done many many glass products in the past and decided i need to start getting into plastics a bit more. It turns out don't really nice looking plastics for products, is quite tough so far!! These beauty and healthy products have tons and tons of semi translucent plastics and it's turning out to be a bit of a rendering nightmare...even with using a xeon 980x :-\ anyways, I'll post some updates as I go.

Here's where I left off with this about two months ago. I felt good about how this looked, ut the plastics just really all looked a bit too solid and not like the semi translucent plastics of the real thing.




Here is the latest iteration I'm working on. It's still rendering, so it's pretty noisy, but far enough along to show. I think these plastics are getting pretty close to the real thing, although i'm thinking they don't look as nice...like in a "magazine advertisment retouched" way. The first, non sss render, looks really smooth and sexy and clean, this one just kinda looks...real but not sexy. I dunno, what are your thoughts? more updates to come as I figure this out

[m-Que]

I personally like the second one - the gray and orange parts look way better. And most important - they look realistic. The top white part needs the same changes as the last two.

As for that look - it depends what you're after - a realistic look, or that 'magazine look'; as you said - it looks sexy, but sort of unreal.
Actually it's not only the case of products like cosmetics; I've noticed the same problem with magazine babes

[itsallgoode9]

43 hours rendering on an intel 980-X (12 threads at 3.8ghz) using IBL.... This is SL 23.58...it will literally never clear up. Even using the ranch render farm, it would take 4hrs to get to this level and cost $685...all of this for a single noisy, unfinished image.

WTF, SSS is absolutely fucking useless.

[rusteberg]

43 hours rendering on an intel 980-X (12 threads at 3.8ghz) using IBL.... This is SL 23.58...it will literally never clear up. Even using the ranch render farm, it would take 4hrs to get to this level and cost $685...all of this for a single noisy, unfinished image.

WTF, SSS is absolutely fucking useless.

OMG - that's just unacceptable!

sorry dude, just trying to lighten up your mood... your example could quite easily be solved with an object/material id selection with just a "hair" of noise reduction applied to the cap.... let's just take a deep breath amigo.

[itsallgoode9]

haha thanks rusteberg, that's hilarious

[zdeno]

just try biased approach and do supersampling behind back of maxwell.

Just render 2800X2800 picture in 24 hours with this awesome machine. SL will be much smaller, but when You shirnk picture to half 1400x1400 it should be much better.

try this first on small region, render 2x bigger or 3x bigger with compared amount of time and just resize output in photoshop to see results

and please don't use F words . if you are using maxwell You should have nerves of stell already.

[itsallgoode9]

that's the approach i am going to try next, even though it is still fairly unrealistic time line for real world projects. Probably be looking into other option ans solutions along with that, i just think Maxwell SSS won't be compatible with me until some sort of speed increase has happened in future versions. :-\

[Half Life]

maybe give it to a render farm?

Best,
Jason.

[itsallgoode9]

My anger/tantrum is back in check now. Thanks for feedback guys, I'm ready to move forward and figure out how to work with SSS. I came from a video game design background, so I've been working with limitations for the past 5 years in that field...I'll figure out how to work with the SSS in maxwell, I just needed to step away for a bit. Something big will eventually be happening with this SSS render. A personal vendetta of mine to let SSS know that I defeated it.