Not there yet? Post your work in progress here to receive feedback from the users.
User avatar
By Nasok
#400215
Hey guys, I'm recently been tasked with some look development and I stumbled upon this pretty common situation (in FMCG) when we need to visualize an opaque product inside a glass object.

In my case, it is a cream (let's say it is fully opaque) in a clear glass jar.

Here's my jar set up:

Image

As you can see it's pretty simple - as it is just not a final object but rather a template to test material and geometry setup in general.

- Cream material is intersecting with Glass object's walls - and Cream material has a lower nester priority number (cream has higher priority than glass)
- Jar is a pure cylinder in shape - pretty small only 63 mm in diameter and 38 mm in height (including the cap)

- Glass material - is a simple BSDF with almost white transmittance colour (we don't like tints in cosmetics 😜), Attenuation distance about 30 cm, Nd 1.420, with forced fresnel, and very low roughness (~1.5).
- Cream material - is a 2 layer-based setup - 1 layer with base colour - single BSDF - with almost white colours for both 0 and 90, Nd 3, no fresnel and roughness 97, 2nd layer is additive (95%) and has a single BSDF with blacks for colour, Nd 3, forced fresnel, with the roughness of about 10.


The main concern was that whenever white (or very light-toned) cream is placed inside a glass object it starts to look quite dirty (looks much darker)


So here are a few test-renders that I did in "studio setting" (light setup is pretty simple - as you can see in the mirror ball)
Image

and another one with reduced intensity of the top light panel and without chrome caps + slightly adjusted angle of the jars (to compare refractions)
Image

And here are some testing that I did in other environments to see the performance in different light scenarios
Image
Image

Altho the primary request was specifically about white cream, I did another one to see how the colour of the cream will influence the glass tint.
Image
... and, well, couldn't resist in trying the Random Colour procedural :lol:
Image

So that's that.
Please let me know what you guys think.
Would really appreciate your suggestions on how to improve - feels like I'm missing something. Probably something simple (as usual).

To increase the brightness and purity of the cream, I even tried to render in different setups, like when the cream geometry is smaller than inner wall of the glass jar, or with the "infinitesimal gap" approach - so far the nested priority gives best results.
I've also tried back wall culling, making AGS glass or single-sided, excluding from global Illumination rays and etc.

All your suggestions are super welcome!
Cheers, 🙌🏼
Artem.
By PA3K
#400220
Hi Artem,

I think Your cream material is too glossy = reflect more light than it should. Your cream has an almost pearl effect. Maybe it is for purpose but if not, try to use one layer 2 BSDF material with one 99 roughness white (225) combined with 10% additive BSDF with 0 roughness 1,4 fresnel. I tried to use it in the attached image back two doses. You can see, it is not enough and it shows the same darkening effect as on your images. So I added white (225) sss to white 99 rough BSDF (500 and 0,5 asymmetry). You can see the front two doses without darkening. I think all creams, in reality, have sss. Hope it helps :-).

Patrik
Image
#400223
Hey Patrik, thanks for the suggestion. 🙌🏼

Creams in cosmetics usually pretty glossy 🤪 .. even if they are not so in real life - we like to make them that glossy :))) for some reason.
And yeah, of course, I agree that creams are SSS - I guess I was just trying to get away from it to save a bit on render time 😔

That is an interesting suggestion to make it out of a single layer with 2 BSDFs and with a scattering - will definitely give it a go. Also would be more convenient to adjust glossiness - shall designers decide to go with less glossy look :)

Thanks, man - appreciate your support! 🙌🏼
Artem.
#400231
Here are some more renders.
I've reworked the cream material a bit

So without SSS is slightly reworked 2 layer material (additive approach) and with SSS is a single layer ( multiple BSDFs )

No SSS
Image

With SSS
Image

No SSS
Image

With SSS
Image

What do you guys think?
Without SSS is for sure more render efficient (benchmark is almost twice as high)

Cheers,
Artem.
#400232
I like them both. Great work! :D

Seeing both, I think the SSS version is more "attractive" visually. It suggests moisture more successfully --but see note below.

The non-SSS version was just fine too by my eye. And, had I not seen both . . . would not have thought anything about it. The non-SSS version looks just like any other face cream image on Google. So, I think you're good to go with that one.

Just a note (from a layman's perspective): the visual appearance of the viscosity on your SSS version seems slightly off. To me, it looks as if the water has separated from the emulsion. I wouldn't want to buy that particular jar of cream. :wink:
#400233
Thanks for your feedback 🙌🏼

Both versions could have either smooth or sharp reflections. It's just I thought that for SSS making sharp reflections could look pretty interesting - but yeah - I agree at some angles (with some light scenarios) it could look a bit weird - like what you said when the emulsion is separated .. I guess I need to soften it a bit.

In cosmetics, we have different types of cream - some of it are more shiny - like oil with sharper reflections and not so much color weight, and some of it are super soft like .. shaving cream. I guess we'll use both.


I was wondering if you guys have any cool tips on rendering / optimization / geometry setup / when in such situations (opaque objects inside glass)

Cheers,
Artem.
#400236
Hi Artem.

Your work looks great and I can't tell that cream looks too darkened in any given render. Every time I try to show white objects with decent details, it's a struggle between whiteness and details. As for me, I'd use just single glossy SSS cream with Nd about 1.4-1.6. Nd-3.0 is way too high IMO and it can be tricky especially when having basic BDSF with nearly 255 white and additive layer of so high Nd - the resulting color can go over 255 in additive mode producing unwanted effects and render time.

Regarding darkening of opaque or SSS objects inside dielectrics... the first thing that comes as an issue source can be light setup and environment. And these another players join the struggle. Now it is going between light setup, environment objects brightness, wanted reflections on glass (details) and decently white contents of the glass. So my humble suggestion would be - want more white, give it more light. And prepare for the struggle. The hardest thing is to show bright white thing on white background with cool details in reflections.

And BTW - another thing about darkening of really big number of SSS objects or just clear transparent objects inside a glass. There is really huge darkening effect. For example you need to render a transparent bottle of sea salt or plastic container with sugar or any other grainy product - it is darkened VERY VERY much because of multiple refractions which are highly time consuming. You will have to wait for ages for that darkening effect to decrease as the SL increases. And noise will go away in the same slowly manner as the darkening effect decreases. It can take literally few days for a really powerful workstation to get some render of such a product. It will not sound good, but with Maxwell it can be helpful to avoid such tasks or projects... or use other render engine for this part. The real issue here is that you can't get fast preview for needed resolution to see if it is going to look great or not. The only friend here is your experience with lights. And Multilight. And having a lot of time. And a render farm. But wait, what? Render farm for that glass of salt? Anyway, I warned you! )
#400237
Thanks Artistus, appreciate your feedback and kind words :)
Yeah so for non-sss version I went with additive approach - which is just a classic 2 layer setup - to make it brighter .. probably more than it should, as in cosmetics we like to have things to be "perfect" which is more artistic VS realistic.
with SSS - it is a single layer setup with 3 BSDFs: First BSDF is just a base diffuse colour, opaque (main colour control), second BSDF is with like you said nd of 1.4. and with SSS, and third BSDF is just a glossness control. Nothing fancy.

Agree - I tried quite a number of light setups - in some cases it was actually pretty helpful to exclude Glass container from Global Illumination - I know I know - not cool - but honestly it always makes contents of that container much lighter (obviously) and in some cases it had no visible benefits (when it is included in Global Illumination).

So that was one of the "possible cheats" which is kinda applicable in some cases (like when no contact shadow is required or so).

But yeah - well planned light setup and environment is kinda a big deal.

Speaking of white - the brightest I went with those materials was 220-225 - so not so "white" :)
Hmm sea salt in a glass jar .. interesting .. need to try :))) will post here once I'm done with my little R&D and current projects - Now I'm curios to try :)

Also, de-noiser does a great job in there situations. and extra sampling, obviously :) I can try render same scenes with other render engine - it's just that I had a request specifically for Maxwell Render - since Maxwell is kinda big for product viz or "digital prototyping".
Will there be a Maxwell Render 6 ?

Let's be realistic. What's left of NL is only milk[…]