Rendering Workflow / Quality

In response to the default settings for metals and not tangling up in the other areas of this thread…

Most metals have different reflection properties or roughness values based on their make up. Personally I don’t think you can just say that all metals have a default roughness value of (x). In terms of precious metals like Gold, Platinum and Silver all 3 have different reflection characteristics from one another. Silver, no matter how much it’s polished has the blurriest reflections of them all. So which one wins as the default? That’s up to the user’s own personal bias and I think it’s probably better to have a sub set of metal presets that more directly correlate to the desired metal. Microfacet GGX has certainly been a huge improvement for metal reflections. I do agree that, out of the gate, the metal reflections are very mirror like vs metal like for most applications.

It’s hard not to agree with Gustavos last post.

One other thing I find frustrating is what happens when you show prototypes of new tools.
If it is Snapshots or also SubD – I do not feel that there’s a disposition to actually rethink the already chosen fundamental approach. Of course Blender’s Scene management is more powerful and less fuzz: McNeel still keeps trucking with a system that needs a dialog coming up and decision-taking, as soon one adds new items to the file.

What’s always welcome is the sort of feedback on minor details which can easily taken of.
That’s all understandable.

But it’s also understandable that users who were generally willing to contribute not just on cosmetic level decide to rather spend their time on something else.

Is there a way to achieve similar reflection and scratches like this Fstorm material example:
https://cgmood.com/3d-model/34/set-of-scratched-metals#carouselExampleIndicators

Here some similar pieces of metal in reality:
image




1 Like

I’ll continue on the metal as well.
The bump texture is new to me and it has great potential, but doesn’t seem well enough thought through yet:

The first thing is Scale:
It has no reference in reality as it doesn’t care if I change from Meters to Millimeters. It just abstractly refers to Rhino’s units (through labels like XXXXL) and not to a real life scale, so I think it is much better to use % so we can type in the amount we like (and also go beyond 100%, to 10 000% or beyond if we like, or down to 0.0001%)

Take a look at this:
I had to scale the ball down to 10 mm and set “Dots” to XXXL for them to show:


Is that correct?
Because to me it makes no sense. (I understand that this is a microstructure that is not intended to be seen, but still, to be able for new users to use it it has to be intuitive. And how much bigger is XXXL than XXL? Now each of us has to explore that and make up our own minds.

And it feels very limiting that the “scale” only has a fixed depth. I expect the ability to adjust both scale and strength.

And to be able to add a custom bump as well.

Remember the old design method saying:
“As simple as possible, but NO SIMPLER!”
And I think you have crossed that line by not providing some of the necessary controls.
And it would be great with icons of the examples of the structures too ofcourse. (I understand that it would require a huger rewrite)

Freudian slip… :smiley: Yes…
And thanks, flipping fixed it.

Ok, good. I’ll look at some point at getting those normals properly handled, but not highest priority. Btw, on your mini you might want to ‘solidify’ your glass panes as well and use Rhino Glass if you haven’t already. I have some older model of your mini (dutifully lifted from the holomark of course :wink: ), and that has single-surface glass panes. We don’t have at this moment any glass material that simulates thin glass sheets by means of a single surface unfortunately.

Custom bump is something for the future, in the very least V7 when we go for principle based materials anyway.

I find myself grappling with this often, too. I don’t think we’ll be making big changes to this for 6.0, but we definitely will be going over this more, as we’ll be doing with the reflection values for different metals as well. Again, how much will make it into 6.0 will remain to be seen, but there is always life after 6.0… :wink:

Thanks for the feedback on them bumps.

Yeah, I know you have to lock down before V6 ships, and it needs to ship soon! And I know you know that we push as hard as we think you can handle, but waiting another 5 years for V7 is another 5 years with pulling hairs when using the production certified version, so do your best in pushing upgrades while V6 is in the loop. (Please! …or I’ll show up on your doorstep with a bottle of akkevit and we’ll drink until we agree) :wink:

My GRANDEST wish though is an exposure control for the viewport. Having to fiddle with the damn HDRI’s each time I add a light to the scene is a show stopper in the long run. Keep in mind that I have both V-ray and Maxwell installed on my machine, so ease of use is HIGH on my list. But the number one reason to use Cycles instead is that it has the potential of being more future and coworker proof as plugins has the risk of going out of market.

Btw: Regarding the mini’s: I trust you know you can use it as freely as you like! (And yes, I gave the glass thickness too, it was made for OpenGL only, so back then it wasn’t needed )

2 Likes

I wouldn’t mind…

Yup, and it is still on my list :slight_smile: https://mcneel.myjetbrains.com/youtrack/issue/RH-38985

1 Like

@nathanletwory
Is implemnting UI of UV scale difficult for V6? There will be much that can be done if patterns can be stretched as well.

Yeah I wanted the toy feel due to the plastic paint look. And as you say DOF + camera angle (above eye height) + Lens length are the three factors that gives the impression of scale. (well flat lighting adds to the small scale illusion too)

Here is a test of the car pain shader (Or “Cycles Flaked Car Paint Material (DEV)” as it is called)
Less DOF, low camera angle and 25mm lens lengt gives more of the impression of being up close to the object.

This is still the same HDRI, as this is a study in just tweaking materials.
We do need a “rubber” material though as it needs both glossyness and strength.
(I think all materials need that thought)

1 Like

Contrary to better options (I really should have been on my bike), curiosity got the better of me and took another dive into the morass. Fired up Windows Rhino, ick (at least that’s a personal problem), and had a go again. Wrestled the beast of an interface in an attempt, as best possible in short order, with limited experience, at an apples-to-apples shootout with one perceived single purpose leader - Brand X. Fair fight, no! Interesting, yes!

Used @BrianJ rough approximation of the Eames Fiberglass chair as the stand-in model, the same vanilla startup default HDRi as in Brand-X, ubiquitous ‘white’ studio parameters, and did my best to equalize all else with very limited Rhino 6 experience.

I think, with effort, I can beat the thing into submission and get something out of it, as @holo has demonstrated above. Not likely presentable world class renderings, but so what, there certainly will be an appreciative market for what it can do.

I’m still most interested in, and think the greater value to mine, is the digital rapid visualization photorealistic aspect of ‘Rendered’ inside of native Rhino; for quick peeks at in progress work. It is good to see that such has progressed from - “WTF” - to merely - “MEH”…Progress! If progress is linear, we might yet have something down-the-line. That’s a challenge @nathanletwory

And here’s an idea about the render setup interface several experienced users have squawked about, that perhaps is already underway. Have @Marlin and crew use the Mac version as the test bed to un-!@#$ the Windows interface…:wink: Just a thought.

So, with rapid as possible, yet photorealistic, interactivity as the goal, here is what I got:

after 10 seconds of not moving the mouse (hint - needs to be faster!)

after 1:30

after ‘completed’ at 2 min 21 second

To compare as a benchmark future goal, here is ‘Brand-X’ after 5 seconds (actually looks like this after 2-3, but takes me 5 to grab the screenshot)

Here’s a vid of interactivity in Brand-X from the Mac side of the same machine. A side-by-side vid of both would be cruel and unusual punishment, plus I’m too lazy to procure Windows screen recording.

And for doo-doos and giggles, here is one actual marketing shot of the Eames from Herman Miller

Keep beating this bad-boy. You’ll get there. Just not too much in v7. My vote, as always, is a focus on modeling.

Rhino and Brand-X models attached if anyone wants to have a go with same, or show us how to improve interactivity speed w/texture.

Eames.zip (1.8 MB)

3 Likes

All! The more studio HDRis the better IMO, especially for beginners. I already have a turd-load of em, but hey, goodie for me.

B,C,E and G seemed to produce decent vanilla results possibly suitable for default status. Still, perhaps something a tad more vanilla for default, such as this???

Center row middle image, or bottom row middle image, looks like best best default candidate to me. Still, someone has to test in variety of conditions, if goal is to present best ‘idiot proof’ starting point for beginners. Sorry, I don’t volunteer…

You don’t want the kiddies all over the internet saying how your renderer sucks, just because they are clueless, and Brand-X made then look smart, do ya?

Edit: Correction, this CPU load is when the GPU is used as source.

Regarding speed: I think this might be part of the riddle:

And this is with the CPU sat as render source.
EDIT: My mistake, this cpu load is when GPU is used.

Hum, never looked at that, interesting. (And not going to bother in WinBlows, though) Brand-X lets me hit all the cores I want, till I run out of em. Still I think the fan was going in Windows. I’m too damaged from the experience. Not going back over there. Need recovery time…:imp:

Dinner and some booze should fix!

So what did you do to “beat it into submission”?

This hdr won’t work as a default because it is impossible to get sharp reflections. A better solution is to lower the default polish and allow the user to crank it up if necessary.

Using what? Cycles? Running on a GPU? You’d expect the CPU usages to be low in that case, right? In fact, you’d hope it would be.

What GPU are you using for Cycles? My guess: not nVidia, right?