ahaaaaa
i found new problems.  the difuse node is a matte to the blender node. so when we use an IBL then the matte is not white-black but white-grey. which means that the AO will not be dark or black when its very close together. this is a huge problem since we use AO to have dark shadows when  something is close together.

That's what I meant by the need to tune it. If the ambience is high enough, you'll never detect that you are in shadow with that setup.

This is the same story as with my shadow catcher. Given enough IBL, there is no specific pre-defined level that means we're in a shadow. It has to be calibrated to detect a shadow in each situation. There will be times where there is so much IBL that the lowest value coming out of the Diffuse node is still more than 1/2.

just thinking out loud, as Im at work and cant try these things, but couldnt you use a clamped inverted diffuse node to control the level of AO based on light on the object?

I think I found an easier way. Poser has a strange quirk - IBL cannot create a specular effect.

So - don't use a Diffuse node, use a Specular node. It can only detect directional light. It cannot detect IBL. That's half the solution. If only the Specular node would behave like the Diffuse node and produce a broad reflection.

But wait - if you set the roughness very large, it does!

In this demo setup, I'm not really lighting the ball. I'm coloring it red or blue - red is in shadow, blue is lit by directional light. (An infinite in this case.) The red part is connected to an AO node. The blue part is not and will have no AO applied to it.

Of course, you should not use red and blue - just use white, and connect that to your Diffuse_Value.

so it works? I knew thered be a way to control it automatically with a node.

but the question is now, how does it hold up in differing angles of light. say you had an animated scene with the light rotating around your object on an oblique angle. would the AO be applied correctly at all angles.

I must test this tonight!

Here are two infinites. The bottom of the upper box is completely in directional shadow and shows a contact AO shadow across its whole face. Whereas the top of the lower box is partially lit by the two infinites and only shows a contact AO shadow where the infinites do not reach. Geometrically, the top and bottom are the same and would normally show AO in the same pattern.

Which means we've solved the problem of material AO darkening the directional lights!!

Wow.

Of course, I have to observe that the light-based AO is working perfectly here in Poser Pro, so I'm not sure why we should bother with material AO at all anymore. It does the right thing.

but its still nice to work something out, and now all those non Poser Pro people can have proper AO!

I like you baggins, you do things very similar to myself

That's fantastic. Great find, ice-boy and bagginsbill!

Baggins, didn't you say at some point that light based AO had some problems even in Poser Pro? I don't follow these threads as closely as I should, but I seem to remember reading something along those lines.

Quote - Baggins, didn't you say at some point that light based AO had some problems even in Poser Pro? I don't follow these threads as closely as I should, but I seem to remember reading something along those lines.

Every time I try the light-based AO, it works right. So I can't find too many problems with it lately. I don't use Poser 7 much anymore, but it seems to be working well there too.

Probably the only reason to avoid it is that if there are large areas where AO is just not important, you can render faster if its on the materials only where you need it.

Quote - That's what I meant by the need to tune it. If the ambience is high enough, you'll never detect that you are in shadow with that setup.

This is the same story as with my shadow catcher. Given enough IBL, there is no specific pre-defined level that means we're in a shadow. It has to be calibrated to detect a shadow in each situation. There will be times where there is so much IBL that the lowest value coming out of the Diffuse node is still more than 1/2.

i am happy that you didnt tell me. it was nice that i found it myself. 

Quote - Of course, I have to observe that the light-based AO is working perfectly here in Poser Pro, so I'm not sure why we should bother with material AO at all anymore. It does the right thing.

we still need to use  material AO sometimes.

i use it for example when its a cloudy day. the light bounces from the ground on the body a little. so teh AO on the body should be lighter then the AO on the ground. at least in the light situation that i do. not always.

so this works when the specular is 100% on the directianl light. right?
but it doesnt work when the light doesnt have any specular right?

Quote - so this works when the specular is 100% on the directianl light. right?
but it doesnt work when the light doesnt have any specular right?

It doesn't have to be 100% , but it can't be completely turned off. Then it will not work.

yeah i tested it. if we turn it off completely then it wont work. we can maybe the specular value 3.
great work.

bump. start using AO. makes the render more realistic.

i noticed something. when i use bagginsbill's  ENV_sphere and if i turn off raytracing then it renders very long.
so basicly if i make the ENV_sphere invisible for raytracing then it renders longer.
of course when AO is turned on. so i made some more tests. if i make a ball and make it invisible for raytracing and turn on AO it will also render longer.  if i turn raytracing on the ball on it renders faster

is this normal?

I've never noticed such a thing, because, of course, I wouldn't normally make the ESphere invisible to RT. Hmmm. Don't have time to test this myself. Very busy today with BITE MY TONGUE.

Quote - I've never noticed such a thing, because, of course, I wouldn't normally make the ESphere invisible to RT. Hmmm. Don't have time to test this myself. Very busy today with BITE MY TONGUE.

i hope it is what i think it is he he he :)

I've noticed this as well... no sphere... longer render times...
My personal explanation was Poser must be calculating reflections, AO, etc. to the end of the universe and back :-))

I see the opposite with Poser Pro on my laptop.

Scene: ESphere, Ground Plane, a 10" sphere on the ground. RT Shadows on one infinite light + IBL no AO. No reflection.

E Sphere Visible and Visible to RT = 17 seconds
E Sphere Visible and not Visibie to RT = 7 seconds
E Sphere Invisible = 4 seconds

then something is wrong with my computer.

This is an excellent thread, thanks.

Fantastic! I didn't know about this...

Thank you.

Laurie

bagginsbill do you maybe know why this is happening? 
on the first pic i have a simple ball on  a plane. on the plane i have ambient set to 1 and AO connected to ambient value.

on the second render i made the ball 100% reflection. so why did the AO on the plane change? 

if we use light based AO then AO should be ignored in the parts where the light is.so AO should only be in the shadows.
this is how it looked in poser 7 and pro from what i remember.i think in poser 8 it is not anymore like that.

The AO is the shadow. I'm not sure what you mean.

When there is more than one light source, the illumination at any given spot is the sum of the illumination caused by each light source. A shadow is nothing more than the attenuation of that light source due to surrounding geometry.

The AO shadow exists even if another light shines direclty into the same spot. That area below the ball is unable to see all the omni-directional ambient light, so it should be darker all round the ball. Very close to where the ball touches the ground, the contribution of IBL should be zero, even if a spotlight is shining in there. The spotlight does not cancel the fact that light from the sky is not going under the ball.

In Poser parlance AO is not a shadow. But in real life, the effect that AO is simulating is indeed a shadow.

but wasnt AO ignored where the light was coming from? 

No. AO doesn't pay attention to lighting at all. It calculates occlusion by sampling a bounded hemisphere, looking for stuff nearby. When nothing is nearby, it generates a 1. When stuff is nearby, it generates numbers closer to 0.

Why should AO be ignored if a particular light is shining on it?

Of course if a light is shining on an area producing an AO shadow, that light will make it less dark. But it will not make the AO shadow go away completely, because it shouldn't.

These are nitpicky issues with AO. It really doesn't matter. For realism, AO is wrong. If you're interested in realism, AO is too much cheating.

The shadows produced by P8 IDL are much more accurate. AO shadows are the wrong shape, mistakenly do not produce shadows from objects far away, and do not calculate a colored shadow based on what is not occluded. IDL (if it works right) is 10 times better than AO.

I still have some adjustment to make on the settings, so this isn't gallery worthy yet. But it is good enough to see that AO sux.

Compare the shadow shapes of this with my original. I will post both.

This is P8 IDL

Wow, there really is no comparison. The Poser 8 render is definitely the better one.

Laurie

come on BB. AO doest sux. yes IDL and GI is  realisticand yes AO is just a cheat. but in movies they have been using AO for years. if it was so bad they wouldnt be using it.

Quote - No. AO doesn't pay attention to lighting at all. It calculates occlusion by sampling a bounded hemisphere, looking for stuff nearby. When nothing is nearby, it generates a 1. When stuff is nearby, it generates numbers closer to 0.

Why should AO be ignored if a particular light is shining on it?

Of course if a light is shining on an area producing an AO shadow, that light will make it less dark. But it will not make the AO shadow go away completely, because it shouldn't.

but you said

''ice-boy what do you mean there should be no AO where the light comes from?

AO is supposed to be used to block the IBL. The same spot may also have a directional shadow. The directional shadow should correspond to the blocking of directional light, and the AO shaodw should correspond to the blocking of ambient light - usually implemented as IBL.

If you use light-based AO on the IBL, this will work perfectly.''

www.renderosity.com/mod/forumpro/showthread.php

I meant that Poser's AO sux in comparison to Poser 8's IDL. The fact that AO only works over a short distance is the big issue. Look at the underside of the plane. There is little or no occlusion shadow on the top of the fuel tank, despite the fact that the entire fuselage blocks the sun there.

As for the quote, I think you're misinterpreting what I said. I'm having trouble explaining this without math.

Look, the IBL supplies a certain amount of illumination. Call that X. The AO decrease that illumination for that light only, Call that factor K. K varies from 0 to 1. A directional light, for example from an infinite for sun light, also contributes light. Call that Y. A directional shadow can decrease that illumination for that light only. Call that factor J.

So the luminance from the environment, simulated by the IBL should be KX. The luminance from the sun light, simulated by the infinite, should be JY. The total luminance is KX + JY. There is no reason that the KX term should change just because the JY term is strong.

AO is usppoed to be used to block the IBL. Right - I said that and I still stand behind it.

The same spot may also have a directional shadow. Right - nothing I've said lately contradicts that.

The directional shadow should correspond to the blocking of directional light. Right - that's the JY term.

And the AO shadow should correspond to the blocking of ambient light - usually implemented as IBL. Right, that's the KX term.

What you were arguing above is that K should go away if JY is strong. That is not what I said anywhere, and I'm disagreeing with that.

i mean that in poser 8 i dont get the same results like in poser pro and 7.
in poser 8 light based AO looks like if i would use AO on the materials.

I created a test like yours, a sphere on the ground. I used an IBL with heavy AO, and an infinite with raytraced shadow.

I see minor differences between PPro and P8. I do not see an overall change in AO behavior, just minor details in the interpolation of samples.

This is the Poser Pro render.

This is the P8 render.

the setup is 1 plane as the ground. 1 ball, BB's Enviro Dome (screen grab of settings to come), 1 IBL Light (same gradient settings as the Envirodome), 1 Raytraced Light on the same angle as the IBL light.

Screen grabs of settings to follow.

so any ideas why I either get no AO or cartoon block shadows or no shadows?

Hmmm. My apologies, i keep forgetting I'm using SR1 already. Maybe it is fixed?

Also, I did not render with GC.

GC will hide a weak shadow.