PDA

View Full Version : mia_material AO


royter
09-06-2007, 12:57 AM
i am trying to exagerate the effect of the AO of my mia_material.
but even if i raise my spread and set a negative value to the the dark swatch,
i still get very weak AO effect.

now when i do this in the traditional mib_amb_occlusion i get the exagerated effect but in the mia_material, why is that?

modernPrimitive
09-06-2007, 01:21 PM
Got the same problem. It's like the AO does just about nothing.

DimensionalPunk
09-07-2007, 02:28 AM
Also curious.

djx
09-07-2007, 04:20 AM
The mia_material's built in amb occlusion is not the same as mib_ambient_occlusion.

The documantation explains it this way
The built in Ambient Occlusion can be used in two ways. Either it is used to enhance details and contact shadows in indirect illumination (in which case there must first exist some form of indirect illumination in the first place), or it is used together with a "specified ambient light" in a more traditional manner. Hence, if neither indirect light exists, nor any "ambient light" is specified, the AO will have no effect.



When "Details" is not flagged, then that "specified ambient light", they are referring to is the "Ambient" swatch color which acts like an ever-present ambient light. This is what gets occluded. So it doesnt actually darken the shader, it adds some ambient light which is occluded first.

However I think that generally you should have "Details" flagged, and the "Ambient color" set to black. Then the AO acts as a multiplier of the indirect illumination from the FG and GI.

One more quote from the docs
Sometimes people use AO as a general multiplier to all diffuse light. This has the distinct drawback of affecting even brightly directly lit areas with AO shadows, which can look wrong. This use is not covered by the built in AO shader because it is trivially acheived by simply applying the mib amb occlusion shader to the diffuse color of the material and putting the materials original color into itís Bright parameter.

So I guess they are saying you should be using mib_ambient_occlusion to get that exagerated effect.

-- David

royter
09-07-2007, 04:55 AM
So I guess they are saying you should be using mib_ambient_occlusion to get that exagerated effect.



i really like the built-in AO of the mia_material but iguess it's always going to be weak.
is there a way to plug the mib_ambient_occlusion node to the AO swatches of the mia_material?

djx
09-07-2007, 05:06 AM
Even though they say "it can look wrong", I have often used mib_amb_occlusion as a multiplier for my diffuse color - and it looks good most of the time. You can probably find examples where the darkening is not truely accurate, but to me it looks fine.

I dont think it would work trying to connect it to the built in AO though, especially since you already tried using a negative "dark" color.

Have you looked at the mib_fg_occlusion? Maybe it could give you the best of both worlds.

-- David

DimensionalPunk
09-07-2007, 05:10 AM
Only good for simulating shadows, not dirt build up. We should again start referring to the mib_amb_occlusion technique as dirt mapping, since that will be its main purpose now.:)

modernPrimitive
09-07-2007, 07:11 AM
Does Mia AO only work with FG/GI - because I think that might be the problem?

djx
09-07-2007, 11:37 AM
Well its not really a problem, except that I think royterr is trying to use it to exagerate the occlusion effect, which I think might be better done using mib_amb_occlusion instead.

I'm pretty sure the intended use of the mia built in AO is to use the "details" flag to get the "appearance" of more accurate FG/GI using lower accuracy settings on these, and thus improving render speeds. With no fg/gi and therefore no indirect illumination, the builtin AO would do nothing, so they provided a built in ambient lighting effect (the Ambient color swatch) - even with fg/gi you could also use this to add some extra light, for example if the fg/gi bounces are not strong enough.

-- David

modernPrimitive
09-07-2007, 12:13 PM
i really like the built-in AO of the mia_material but iguess it's always going to be weak.
is there a way to plug the mib_ambient_occlusion node to the AO swatches of the mia_material?

I've been plugging my colour or texture into the bright colour of an mib_amb_occlusion and then plugging that into the diffuse of a mia.

royter
09-07-2007, 01:40 PM
I've been plugging my colour or texture into the bright colour of an mib_amb_occlusion and then plugging that into the diffuse of a mia.

ok, but if i a have a bumped mia material, will your AO method be accurate? will it be bumped ?

drossxyu
09-07-2007, 06:15 PM
Does anyone know if the built-in MIA AO takes light direction into account like the Ctrl.studio AO shader?

modernPrimitive
09-07-2007, 09:38 PM
ok, but if i a have a bumped mia material, will your AO method be accurate? will it be bumped ?

Yes - from what I've seen - An "external" AO plugged into diffuse will generate occlusion in the Mia bump detail if the bump is strong enough and the distance and spread values in the mib_amb_occlusion are right.

modernPrimitive
09-07-2007, 10:14 PM
Yes - from what I've seen - An "external" AO plugged into diffuse will generate occlusion in the Mia bump detail if the bump is strong enough and the distance and spread values in the mib_amb_occlusion are right.


hmm....I just tested that it doesn't generate occlusion with bump. I could have sworn it worked before????

royter
09-08-2007, 03:45 AM
one can have more flexibility by applying a traditional mib_amb_occlusion shader to the mia_material just like when you are rendering an ao pass and compositing it in post.
( i didn't know thow that you can apply it directly to a shader via the diifuse swatch).

but, it seems that combining the mib_amb_occlusion and the mia_ao, gives more convincing results since you have the flexibility of the mib_amb_occlusion and the fine details of the mia_ao just like in the 3rd picture.



http://www.aedii.qc.ca/v2/fichiers/uploads/section_4/4.jpg

CGTalk Moderation
09-08-2007, 03:45 AM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.