"Devide shading subidvs" leaves Vray working linearly? Is this correct?

[Lupaz]

More like what Maxwell is doing.

Oh ok. In that case I better shut up.

[joconnell]

Wouldn't that be the opposite of what we are talking here?
Some say they love the universal settings (which is contrary to what it was written for so many years by Vlado himself, that there are no universal settings).
Others say it's a completely wrong approach.

Yeah it is the opposite but if you're only using aa, then there's no other sampling settings that really have an impact for one thing and if the only really quality control is the colour threshold or the quality that vray has to achieve before it's satisfied then that's going to be as linear as you can get - the finer the quality you get, the slower it's going to be, guaranteed. If you're not using a high aa / universal type approach and you're using just enough aa and controlling quality through lights and material samples then changes in the noisiness of those samples will start to have a slightly more random effect on how much the aa sampler will work. That's the part that probably becomes more confusing for users - why when I turned down my settings did the render times get worse? I reckon it's mainly because they're turning down the wrong setting. The colour threshold is setting the standard that vray is aiming for and say for example you decide to turn down your light and material settings to try and speed up your render, all that does is hand worse results from those two over to the AA sampler which will see that they're not good enough to meet the colour threshold and then start using more AA to try and hit the quality you want. If you just wanted more speed then you'd be better off just using the colour threshold which will make vray accept lower standards and thuse speed up render time.

In terms of Vlado saying that there's no universal settings, he's right and what he means by this is that every scene is different with different sampling needs so there is no one set of settings that he can give people that'll work perfectly in every case. What the universal settings are though, is an approach that's guaranteed to give you very good results in a very easy to set up fashion. It sidesteps some of the sample balancing confusion in vray and while they won't be the absolute fastest settings you can get, they will work the same way for every scene which is a big benefit. I'm one of the people who uses a more custom approach but that said I'm more than happy to spend an hour tuning things - I tend to get a lot of shots in sequence that are very similar so an hour spent tuning one scene can give me speed increase that'll work across multiple shots. What's happening right now though is I'm rendering a lot more 3d motion blur which is generally high AA and you're nearly into a universal type approach anyway

I believe that other render engines are noisy because they let the secondary rays to do their job, as opposed to the AA. Isn't that right? Vray would be something closer to this with "divide subdivs" un checked. Am I way off here?

What's maybe truer of other renderers being noisier is that they have fewer smart optimisations than vray has internally and so they're less capable of dealing with the amount of rays that vray can fire for a similar amount of render time. In terms of turning off divide, vray would just break it's current relationship with the upping and downing of the other samplers. Ultimately though it'll always be the aa sampler that has the last word since it's all of the samplers for light, materials and gi combine to make a pixel effectively, and the aa sampler then has a look at whether all of those pixels are good enough to meet your colour threshold. In this regard it's almost impossible to totally uncouple dmc AA from the other samplers. The adaptive subdivision sampler maybe comes closer since it's judging things on a few different categories and not just the final pixel value but again it's more memory intensive than DMC.

[joconnell]

I'm increasingly working like this. For me it makes far more sense to spend time working on the scene itself and then have a noise control for render times. During testing I'll render low-res drafts with a reasonable noise level. For finals I'll just let it cook longer at full res. It's a nice workflow. More like what Maxwell is doing.

Yeah that's generally how I'll do my test renders too, colour thresh around 0.1 or 0.05 and then drop to 0.01 or 0.008 for final bits.

[Lupaz]

Ultimately though it'll always be the aa sampler that has the last word since it's all of the samplers for light, materials and gi combine to make a pixel effectively, and the aa sampler then has a look at whether all of those pixels are good enough to meet your colour threshold.

Yes. Now I get it. It was also very helpful to use the V-RAY DMC CALCULATOR by Akin to understand this.
For those who didn't yet, it's great! http://www.cggallery.com/tutorials/dmc_calculator/

The colour threshold is setting the standard that vray is aiming for

Do you know how the "min shading rate"" works? Cause now I'm not sure if it's redundant with color/noise threshold :s

The threshold I'm assuming it's just the contrast in color.

[Lupaz]

Ok. this is what I found from Akin on another thread:

Min Shade Rate basically raises the maximum number of secondary samples taken in your scene, regardless of what your AA is set to.

.

I guess that answers my question: Min shading rate behaves more like a global sample multiplier, and color threshold works as a limitation to the AA. correct?

[joconnell]

Righty. You can have a look at this in combination with Akin's calculator to see how this is working. I'm going to do some incorrect maths just cos it's handier for me to not think of the subdivs being squared so just look at the idea but not the numbers as being accurate.

- Say you have aa set to 1 / 5 and a material set to 50 subdivs, the aa's max value of 5 divides our 50 value so we get 10 samples for each time an AA sample is used. Lets pretend in this render that our materials are doing something like driving glossiness by a detailed bitmap so they're still noisy, so we need to up our material samples to clean this up.
- Next we'll keep our aa 1 / 5 but increase our material subdivs to 500. Again our max aa of 5 comes in and divides our material samples but we're now getting 100 samples for every aa ray shot. Lets say with this render our materials are now really clean. Lets say to throw a spanner in the works we now decide we need to render with 3d motion blur or depth of field. This is a geometry issue so that means we've to increase our max aa samples to clean this up.
- Say if we now go up to aa of 1 / 50, again the aa's max value divides our 500 value again but now we're only getting 10 material samples for every AA sample used. This is kind of back down to a similar amount as what we started off with and lets pretend that our motion blur or depth of field is nice and smooth due to our 50 aa samples but our material has become a bit dirty again, here's where the aa division of material samples becomes a bit annoying. If we need to get back up to our 100 samples for every aa ray shot, then we'd have to set our materials to be 50 x 100 = 5000 (Again a warning, not actual proper numbers, just ideas here!). It's actually hard to do that since the spinner for all of the subdiv settings are locked to a max of 1000, and it's a bit of a pain in the ass to do anyway.
- Now on to our min shading rate! What this rather handy control will do is not give a shit about the division stuff going on, and it'll set a minimum amount of secondary samples that vray must take for every aa ray. If we continue with the above example where we've got our aa at 1 / 50 and we'd have to set our material at 5000 to get our previous number of 100 secondary samples per aa ray, we can just set our min shading rate to 100 instead. What it'll do is tell vray for every time you use an AA ray, make sure that you take 100 secondary samples every time if the amount of secondary samples you were going to use was less than 100. With this what you can do is freely set your AA to whatever you want, and then instead of having to use really high numbers in each of the secondary samplers controls, you've got one control to gradually up all of these things without having to give a crap about any of the division stuff.

The previous issue with the universal stuff was that the 1 / 100 settings would sometimes make it difficult to get some of the secondary sampling clean - the AA was fine but materials and light could be a bit grainy. The min shading rate gives you a handy control to force up the standards of that.

You're pretty much correct on the colour threshold - the lower down that goes, the closer all of the results have to be in vray before it'll consider the rendering good enough so ultimately it's the boss of absolutely everything. Vray doesn't really give a shit where the clean quality is coming from whether AA or material sampling, it's just constantly aiming to hit that colour threshold.

[Lupaz]

J, that was a great explanation. Really, REALLY thank you for your time.

Now there's going to be people here that are gonna roll eyes. But I have yet another question. If min shading rate

set a minimum amount of secondary samples that vray must take for every aa ray

, what's the difference with "min samples" in the DMC rollout? Are these min samples BEFORE the division as opposed to min shading rate that comes AFTER the division?

Min samples - determines the minimum number of samples that must be made before the early termination algorithm is used. Higher values will slow things down but will make the early termination algorithm more reliable.

I assume the early termination algorithm is the noise threshold.

[RockinAkin]

Min Samples can be thought of as Min Adaptive Samples - and is limited by your 'MAX Secondary Samples per Primary Sample' (refer to my calculator).
Min Shading Rate is not limited by MAX Secondary Samples per Primary Sample and will infact raise it.

So Min Samples is much much more limited than Min Shading Rate.

Hope that makes sense.

[Lupaz]

Ok, so it's basically the same but with the limitation that in the case that "Max secondary samples" is lower than "min samples", we are in trouble. This limitation doesn't exist in the newer "min shading samples".

Edit: if it's not like this, I'm fine with it cause something is telling me that I don't really need to know that.

Searching for "adaptive samples" I found your complete description in a thread of 2013:

Simple explanation:
The Min Samples value is limited (clipped) by the the Max Secondary Samples (Per Primary Sample) value.
The Min Shading rate value is not limited and will in fact raise the Max Secondary Samples (Per Primary Sample) value.

More complicated explanation:
Min Samples is better thought of as 'Min Adaptive Samples' (as it is named in V-Ray for Maya, and which I'm hoping vlado also renames in V-Ray 3.0 for max).
This value basically says: 'of X number of Adaptive Samples possible, this value will set the Minimum number of Adaptive Samples'.

When your AA settings become higher (like 10 Max / 1 Min), the max number of secondary samples gets lower and lower until it eventually becomes 1 Max / 1 Min Secondary Samples (Per Primary Sample).
In this scenario, there's no more room for adaptive secondary samples anymore, and the Min Samples value has no effect at this point.


The Min Shading Rate value however isn't bound by any other values... it sets the floor for the maximum number of secondary samples to be taken per primary sample... regardless of any other settings.

By the way, I can foresee your calculator being the future V-ray's UI.

[RockinAkin]

Ok, so it's basically the same but with the limitation that in the case that "Max secondary samples" is lower than "min samples", we are in trouble. This limitation doesn't exist in the newer "min shading samples".

Bingo.

By the way, I can foresee your calculator being the future V-ray's UI.

This has been proposed a few times as well.

[LudvikKoutny]

Need of a calculator to set up something as trivial as render settings for a particular scene.

Come one guys... do you really want THAT to be a future of rendering?

[Vizioen]

Need of a calculator to set up something as trivial as render settings for a particular scene.

Come one guys... do you really want THAT to be a future of rendering?

So you want a one button push render solution that everybody can do and will put everybody out of business ?

[LudvikKoutny]

So you want a one button push render solution that everybody can do and will put everybody out of business ?

There are many one button push solutions, yet most people are still in business. Knowing how many glossy subdivs to set when you use adaptive sampler, your max AA subdiv is X, your min shading rate is Y, and your noise threshold is Z won't make you better artist.

"It needs to be difficult, only because we need to keep our jobs safe from people, who could do as pretty images as we do, if render settings jungle was not an obstacle for them." is really sad argument. Kind of argument those last few remaining Mental Ray fanboys use. I believe you know better than that

[Dariusz Makowski (Dadal)]

So you want a one button push render solution that everybody can do and will put everybody out of business ?

Yes.

Becaue then artistic people will create art. Not technical people attempt to create art.

[Lupaz]

I just read this from Vlado on another thread (since Brazil vs. Chile got too boring!).

The "Adaptive subdivision" sampler is the odd one out, as it does not divide the shading subdivs by the AA samples.

Shouldn't "divide shading subdivs" be grayed out?

Edit: I meant grayed out when "adaptive subdivision" is selected.