[rusteberg]

don't know what it is, brett, but this latest one isn't as convincing to me as your others... seems almost "too" frosted, if that makes any sense?

[simmsimaging]

don't know what it is, brett, but this latest one isn't as convincing to me as your others... seems almost "too" frosted, if that makes any sense?

I don't know - this is very much like the last one that Tom thought was the better effort from me. There are things I have liked about all of them - I leave it to you guys to pick among all of the Vray offerings for the best one, whichever and whoever's that is

The one thing about this one is that the smoothness of the grain has lost a bit of the real feel, but it did bring it more in line with the other samples so I thought that was a good thing. (Personally I'd add grain to all of them if I was going to use them for a job, they look too illustrative when they are that smooth)

b

[rusteberg]

look too illustrative when they are that smooth

it does seem to verge close to "toon". maybe that's what seems a little off about it?

maybe you're looking too hard at tom's example instead of trusting your own intuitive instinct? what would the frosted dragon look like had you no example of it to compare to?

could it be that by investing too much time in the science behind it, you've lost sight of frosted glass 'sex appeal'?

edit: the best part about this thread is when frances peeks in, stirs up shit, then leaves

[simmsimaging]

look too illustrative when they are that smooth

it does seem to verge close to "toon". maybe that's what seems a little off about it?

maybe you're looking too hard at tom's example instead of trusting your own intuitive instinct? what would the frosted dragon look like had you no example of it to compare to?

could it be that by investing too much time in the science behind it, you've lost sight of frosted glass 'sex appeal'?

edit: the best part about this thread is when frances peeks in, stirs up shit, then leaves

Agreed - she's been quiet of late and it is getting dry.

Really, I haven't invested *any* time into the science of it - I'm just spinning dials and pressing render. Tom's got the science and covered

Regarding the look: I haven't lost sight of the sex appeal, that's why I thought I was okay with the first one: it looked fine to me (albeit the render was a bit rough). It's more that the sex appeal just got set aside as not being the relevant issue. It kinda grew into a thing about matching stuff after that, which is a different goal. Anyway, the last round was posted just to refer to a specific thing I mentioned to Tom, I hadn't intended it to be the new example for Vray, but whatever works.

In answer to your question though: it would look a lot different because I'd have spent the time on lighting it better instead of messing with matching materials to references



b

[tom]

Here's a cropped portion in one by one comparison.

There are 3 points leaving questions on my retina:
1- The shadows are very dark and not filled by caustics like it's something opaque.
2- The color bleed along internal reflections are not satisfactory leaving some parts shaded grey.
3- The frontal glossy reflections are dominating the inner reflections damaging the frosted look.

[rivoli]

I rendered this a bit ago, right after rusteberg posted his vray images. was just curios to see if I could match tom's output (unfortunately didn't quite managed it), and then forgot about it. now I see brett has covered all vray's possibilities, and you managed to dissect the differences between maxwell's and vray's output down to the least detail.
not sure if it's still worth it to post yet another vray render, but here it is:

[tom]

Thanks, Rivoli! I think your attempt has fixed points 1 and 2 good enough while only leaving the 3rd point so, I'm updating the table with this one.

[tom]

edit: the best part about this thread is when frances peeks in, stirs up shit, then leaves

Agreed - she's been quiet of late and it is getting dry.

I would consider it inappropriate to contribute a fryrender or an Arion image considering my position at RandomControl.

It's OK, Frances has a reason. But, is she the only fryrender user? It's been one and a half month by now and it doesn't sound like a contribution is on the way from their side. That's so interesting because, we're very used to their prompt replies on anything else.

[rivoli]

I think your attempt has fixed points 1 and 2 good enough while only leaving the 3rd point so, I'm updating the table with this one.

eh, that was a tough one, trying to match your maxwell output. I guess the shader models are very different, and react in different ways. this doesn't mean one is correct and the other is wrong. it's worth nothing that we don't really know how the actual dragon made of frosted glass would look if photographed under the same lighting, I just couldn't find a way to make the dragon look like yours. especially the head, and the neck/arm holding the ball, there are huge differences there. but I guess that was to be expected, and I guess it would prove true even if it was the other way around, with someone trying to match a vray output with maxwell.

[simmsimaging]

Nice job Rivoli. Regarding the opaque shadows on mine: might have forgotten to toggle that option. It's something that conflicts with full normal caustics in Vray and normally I don't use it as use caustics. However the beta I'm on does not have refreactivecaustics ATM so I should have activated that but probably didn't.

Regarding the third thing, it may be a case of having to refine the glossy reflect layer more, but I agree with Rivoli that pursuing an exact match is pointless.

[tom]

it's worth nothing that we don't really know how the actual dragon made of frosted glass would look if photographed under the same lighting

The second test covers this > http://www.maxwellrender.com/forum/view ... 48&t=34574 There's a photo comparison and you can also give it a try. I want to see something else than Maxwell coming closer to photo.

I agree with Rivoli that pursuing an exact match is pointless.

Exactly. It's just clear that V-Ray is producing a visually acceptable frosted glass.

[rivoli]

Nice job Rivoli. Regarding the opaque shadows on mine: might have forgotten to toggle that option. It's something that conflicts with full normal caustics in Vray and normally I don't use it as use caustics. However the beta I'm on does not have refreactivecaustics ATM so I should have activated that but probably didn't.

cheers brett. this one posted was rendered with bf+lc and direct caustics directly mapped into the image. of course it didn't compare consistently with maxwell, so I ran a full ppt test and found out that the result was almost exactly as the one I had already rendered. that's why I didn't even bothered to save it and posted this one instead.

The second test covers this > http://www.maxwellrender.com/forum/view ... 48&t=34574 There's a photo comparison and you can also give it a try. I want to see something else than Maxwell coming closer to photo.

I've seen it tom, but you posted it after I already did the dragon. and well, this one is getting more interesting even if less "scientific"..

[tom]

I've seen it tom, but you posted it after I already did the dragon. and well, this one is getting more interesting even if less "scientific"..

Less scientific than?

[rivoli]

less scientific than the one with the reference photo I meant. but it turns out to be more interesting to follow, now that a few people have started posting renderings with different engines.

[tom]

Oh, yes. Well, it may practically seem less scientific but, you shouldn't underestimate this Stanford guru and the physics he's using. Of course, his result is above a certain technical level especially considering it's done in 2003. Basics like Transmittance, Refraction and TIR are not a misery while roughness is the only x-factor in this test if you consider the rest of the engine is working perfectly without cutting off something for speed. For this reason, I find the dragon test as valid as the red cup photo test. I also think the personal perception is another valid tool. We can all clearly see which ones look frosted and which ones are looking like a silicon gel or having other oddities. You're using your eyes for tens of years and you're into visual arts, computer graphics and so on, and have a record of billion frames of how you've seen frosted glass here and there in your entire life. But, still suspicious? Go with matching the photography then, and when you match; it means your engine is physically correct enough to let you reverse engineer all the details. If not, there's something wrong in the engine before the artist's touch.