Quote - "Any horizontal surface will generate .5" ?
 I take it you mean... "Any surface with horizontal normals will generate .5"

Yes.

Quote - I'm trying to understand not so much the maths, which I think are a bit beyond me, but the concept. The PM:AO YBias part of the shader makes an assumption that the dominant source of light is directly above, so upward facing surfaces will be most affected by AO.

Correct

Quote - This concept seems to be moving the modelling of directional lighting away from the lights themselves, and putting it into the materials.

Which makes me uncomfortable -- I think that's just because I don't understand the need for it. Is it because of the limitations of AO as implemented in Poser?

Absolutely. Look, if Poser 6 light-based AO didn't have so many bugs in it, I'd not be doing this at all in the material. For 2 years, I resisted material-based AO, and face_off (who is more masterful than I am, frankly) kept telling me to stop banging my head against the wall. There is no doubt whatsoever that light-based AO calculations are more correct than material-based AO IN THEORY. However, after I ended up with artifacts for the 15,000'th time from light-based AO, I gave up and started to work with the material-based AO techniques.

I have some evidence that light-based AO has been improved in Poser Pro and Poser 7 since SR3, but I'm still not certain about it. If I could trust it, I would generally not bother with material-based AO.

In general, I would prefer to have no lighting-dependent factors in the shaders. Without a doubt, if you choose a wildly different lighting scheme, you're going to have to go into the shader and adjust the YBias value to match. You would not have do to any such thing with a light-based AO calculation.

In a similar vein, I prefer an SSS calculation that is not based on manually adjusted factors related to the lighting situation. But for several years now, the very best shader systems, those coming from face_off (realskin family of shaders) have not only got material-based AO but also the SSS itself is hard coded into the shader. Face_off provides a script you run to automatically modify the shader SSS parameters when you move your light. I don't like that solution very much for the same reasons - I should just be able to move my light, and the shader calculations take everything into account. To some degree, I think I've been successful in approximating skin SSS without directly encoding information about light directions and intensities. To that end, I think I've improved on face_off's work. But I still don't have a perfect solution to the AO modulation aspect. That's why I added the YBias parameter.

I have so many other things on my plate, I don't have time to investigate whether light-based AO is good enough to use now. Furthermore, I'm trying to make realism for P6 users as well, and so even if a P7 or PPro user would prefer a shader from me without material-based AO, there are still quite a few users who are going to need it, and who will be willing to accomodate the quirky nature of the solution by adjusting parameters in the shader when they adjust lights.

---

Renderosity forum reply notifications are wonky. If I read a follow-up in a thread, but I don't myself reply, then notifications no longer happen AT ALL on that thread. So if I seem to be ignoring a question, that's why. (Updated September 23, 2019)