This discussion reminds me of exterior architectural renders, where people wonder why all the interiors beyond the windows look black or very dark. It's how things are, but the brain "thinks" it should look just as lit as the exterior.

What changes? A theory might be the intensity of the mid-point of the shadow gradient is affected by "strength", not the most or least occluded, but what is half-occluded. !!!!

I conclude that there is no way to get any light into the point of contact if AO is turned on and properly configured to operate all the way into the deepest part of the crack.

In fact, from my understanding of physics, this is correct behavior. In this geometrical situation, no light can reach the point of contact between the two cylinders. Very small amounts can reach very close, but the point of contact gets no direct or bounced light.

Real lips have SSS, so the light can reach the point of contact from under the surface, entering on the front, not the point of contact.

This is why we need SSS.

I could continue this experiment with IDL instead of IBL+AO, but I suspect it will result in the same. There is no direct way to get light into the crevice.

My wife insists I must go work for her now.  Hehehe.

Quote - > Quote - Here's demonstration. Depth mapped lights, with shadow maps at 1024, min bias back up to the default of 0.8.

With min bias set that high, it is likely that the lip geometry just isn't casting a shadow there any more.

I think you're missing the point. I was showing that even at lower quality settings, the DM shadow is better than the RT shadow on this point, until I get the RT settings way down in the Min Bias. I can set the Min Bias on the DM wherever I bloody want - the DM will never ever ever make a blotch like that.

Quote - > Quote - It is not physically correct.

Poser's shadows are not anything close to physically correct even under the absolute best circumstances.

I was responding to ice-boy's insinuation that the ugly black mark was correct.

Quote -
Do you have a 100% spotlight shining directly into the character's face, maybe 1 foot above the character's head?  If not, where are your lights (screenshot would be nice)?

Pardon my frustration, but how many times do I have to say this? I have these results with a huge variety of angles. I showed several with the spotlights directly in front of the face. Not 1 foot above. Directly in front. Others, with various angles. (I can't show screen shots now, because I've changed the scene for every test. I'd have to recreate everything.)

Bagginsbill, your tests are interesting - but other than some of the weird things that AO is doing (perhaps because I don't fully understand AO) - and as you can see from the thread, I shut AO off first in the render settings, then in the materials but got the same problem - but as I say, other than some of the things AO is doing, your results are pretty much what I would expect. They are not the same thing as what is going on with my renders at the lips. Sure, the black is very black. That is not the issue. In every case, you're showing renders where there is a transition. Mine is a felt marker-like line, AO or no AO. That cannot be right and is impossible to be right. What you're showing is that the point of contact is completely black. But the point of contact between upper and lower lip is not 1/4"-1/2" thick. You can reference any photograph or mirror you want. You'll never see this, ever.

Moreover, my problem results are not reproduced by DM. Okay, poorer quality setup; fine. They're also not reproduced by RT at .05 settings. They're not reproduced by DAZ Studio with raytracing.

This isn't simply a violation of an erroneous mental preconception. It's a bad render.

My solution is simple. I keep the Min Bias much lower (hoping the other artifacting doesn't get too bad... I think I could probably get away with a 1 setting if necessary). But it's frustrating that I don't know a way to set up the default new light with my preferred settings. It's easy to overlook changing stuff like that until you've already rendered, plus, it's just a PITA.

I may have greatly misunderstood Bagginsbill's posts, but they appear to boil down to, "It's a shadow."  The ones where he's demonstrating AO behavior don't have to do with your specific complaint, which is about raytraced shadows - which he covered here:
http://www.renderosity.com/mod/forumpro/showthread.php?thread_id=2779249&page=2#message_3501830

Thanks. I understand his observations completely. But he's not reproducing what's happening. (And anyway, he has the light source above the objects there.) And as I said again, were it accurate, reducing Min Bias to .05 would (EDIT: would NOT) make it pretty much go away.

sorry, but i don't see the problem myself.
http://seedmagazine.com/images/uploads/attractive_article.jpg
http://img1.ak.crunchyroll.com/i/spire3/b244c4602bfe518343314db046afffef1224361421_full.jpg

depth-mapped shadows are just wrong, and with a small map size (especially without out zooming) they really don't shadow crevices on a face properly.  even 1024 wouldn't necessarily be enough if the shadow cam covers a whole scene or even the whole figure.  basically, you can't use them as your comparison, because physically, they're just completely wrong.

it sounds like what you're fighting with is spread or blur.   what's your shadow blur on your raytraced shadows?  when you use AO (if at all), what's your distance?  imho, Poser's raytraced shadows do have a problem with being too differentially blurry.  while they can blur nicely at a distance, the close shadows (like on the lips) are too sharp at almost any setting.  but that's just to my eyes; i haven't actually done tests to match photos.

also, since you haven't mentioned it, i'm assuming these are all at the same blurring amount.  even if it's not at 0, you might try turning it up.

the more important problem i've encountered with ray-traced shadows and blur is grain.  stewer, is there any chance of getting light shadow sample control, the way we have for AO, in SP1 - or SP 3 or 4 for that matter - or Pro?  if not, would it be possible (without revealing secrets or taking up tons of your time), to explain why this is difficult?

I find myself repeating myself.

I've had these problems with no blur.

I've had them with plenty of blur.

(Blur makes no sense whatsoever as an explanation, anyway, since there is no blur there, regardless - just a solid black [or red, in the case of the overexposed images] mark.)

I don't have them with DM lights no matter what the settings are. I don't have them in DAZ Studio with raytracing.

And yet again, if the hard ugly mark were right, it would still be there at .05 min bias on RT lights.

I find myself shocked that anyone can look at those images and think the lip mark is normal. No decent quality photo reference will give you that, never mind real life. Even the image you linked, although a very poor quality image (35 kb), isn't as extreme. You can still see a bit of a gradient as you approach the dark line, even though I'm certain the image has been postworked for high contrast (which would make the issue worse).

I prefer the dark shadow over the glowing nostrils in the shadow mapped version

What I would do without going into rocket science math nodes etc, would be:

Render two versions, one with shadow maps, one with raytraced shadows

And take the best part of both as postwork in the Gimp, Photoshop, etc etc.

Easier.
Faster.
Less frustrating.

So maybe it is perceptually wrong. Maybe it is physically-mathematically-etc correct. Who knows. But meanwhile, you get an image as you like it.

I'm already getting images as I like them, now, since following Stewer's advice to try .05 Min Bias. If I do get other artifacts, I'll experiment with how far I can edge it up and still have the lips acceptable. I did one angled render that suggested to me .1 may be okay, though not as good as .05.