Thank You for the skin, it's a huge improvement. With all the great shaders I seem to be picking up lately, the skin was falling behind.

Richardson, I'd be tempted to delete that Blinn node, since it's not actually doing anything.  Also, Blender_9 looks like it could be deleted too.

BB, I am sold on your maths behind the fresnel_blend.  However, unable to sleep last night I ran an experiment.  I compared fresnel_blend and edge_blend, and got an identical result....down to the last pixel.  I am trying to better understand this.  I thnk part of the confusion (for me) is that the two nodes have opposite color in the inner and outer colors.  The render in this image is the fresnel_blend.  The scene is lit from your envsphere - no lights.  The diffuse_color of the rendered sphere is black, so all the color is coming from the reflection node.

Pls IGNORE my above 2 posts!!!!!

The fresnel_blend needs to plug into the reflection_value, rather than the reflection background color. 

My bad!

Fresnel_blend

With NO blend (of either type), because the lightsource is 360 degrees around the object, there is no reflection definition.

The fresnel_blend reduces the camera facing poly reflected specular to about 1/3 of that at the rim.

I can see this definitely helping if you have large area lightsources.  The other bonus is, if you have large area lightsources, you can reduce the Reflect node quality a heap, and reduce the IDQ, saving on rendertime in both cases.

So hats-off to BB - I'm sold.

Nice one Paul.

I'm experimenting with a new virtual light. I notice that parts of the body closest to the emiter can get burned (yellow bloom). Too far and it goes dull. So many variables. Maybe a morphable contour... hmm

While I really like the results of just emitters on the final render, I find it to be REALLY hit and miss, even not acting the same way from render to render...maybe too much to use it exclusively for now.

Laurie

Laurie - look great.  If you don't have any lights in the scene (and I assume there are no lights in the scene), then the Blinn will not work - it will output black.  You need to add reflection nodes to the jacket leather (as you've done on this skin).

The render to render results will differ slightly because the irradiance cache gets recalculated.  If you use high enough cache settings it should be pretty consistent.

Quote - With NO blend (of either type), because the lightsource is 360 degrees around the object, there is no reflection definition.

Quite. It looks like you've set up that Reflect returns a constant value 2 from the environment and then you have the Reflection_Value set at .5 intensity, yielding 1. You could skip the whole lighting/reflection thing altogether and just plug whatever blend you're testing into Ambient or Alternate_Diffuse. Poser will just draw the node then.

Quote - The fresnel_blend reduces the camera facing poly reflected specular to about 1/3 of that at the rim.

Not 1/3.

You're not taking GC into account. I wrote the formula a page or two back.

k = ( (ior-1) / (ior+1) ) ** 2

The Fresnel effect of IOR=1.4 facing the camera is (.4 / 2.4)**2 = .02777. Gamma correcting that you get .02777 ** (1/2.2) = .1961. In 255 scale that is .1961 * 255 = 50. I measure the middle of your ball at 50, so that's all matching up with theory.

But it's not 1/3 even if you ignore gamma. At the edge it is 1. I mean the very thinnest edge. You'd need to render zoomed in a lot more at the edge to see that.

I am playing with SR3 and using no IC. Noodle on that.

Which irradiance cache is off, raytrace or IDL? Is "off" = 0, or 100? I still find this confusing.

IDL IC=100. The IC value is how accurate do you want it. The less accurate you want, the more the cache is used instead of calculating new IDL samples. When you set accuracy to 100, no cache is used at all, and the whole IDL precalc is skipped.

Yes but, people dont have SR3.

Quote - Yes but, people dont have SR3.

Yet. :-)

Patience !

That is spectactular. I love your morphs ;). Too bad we're kind of limited with hair, both transmapped and dynamic (which has it  faults).

Laurie

For me to get this, I had Diffuse at .5  and Amb at 7. I know this does not mean much without exact trans and rotation, scale... But proximity is an issue

Opps I didn't see you again. I must learn to refresh more often. I'm having a lot of trouble with hair and these proximity shadows. It's strange.

Quote - OK. I'm off the heads for a bit. I like parts of this but you can see what a limited range light emitters have. I guess the easy fix is to go big, high ambient and far... to average out the ray lengths for more consistency, right?

If I'm understanding your concern, you have to realise that when using emitters as sole light sources, you are working more with real world lighting than ever before in Poser.  Poser lights weren't exactly physically standardized.  So to achieve the same kind of results visually you may be used to, or liked, will require some thought.  The smaller the emitter is in size relative to the scene, the tighter and brighter the reflectivity and shadow will be.  And you'll have to adjust ambient values accordingly.  So it's like relearning the wheel in  sense in Poser.  This is something I grew used to in other apps like 3dsmax, where area lights were commonplace in renderers like Vray.  Once you get the hang of it, however, you may never want to use standard Poser lights again.

Heya Maxxx,,

..what it takes these days to get you to post. ;) You see, we have been putting the screws to this. I  get what you are saying. I'm just trying to see what the limits still are in Poser.

I think multiple panels might help. The other thing I wondered was,, if you have an enclosed room, could you  not have a second reduced room invisible to camera that was 8x white to produce a "fake" lightsource? I'll test this once I get a render done.

No ones adressed the shading under hair props here yet. This, not sure how to control it.

And backfacing (ear) scatter works great with emitters!

I used to use emitters all the time with Luxrender, but it's trickier in Poser!

Laurie

Quote - Heya Maxxx,,

..what it takes these days to get you to post. ;) You see, we have been putting the screws to this. I  get what you are saying. I'm just trying to see what the limits still are in Poser.

I think multiple panels might help. The other thing I wondered was,, if you have an enclosed room, could you  not have a second reduced room invisible to camera that was 8x white to produce a "fake" lightsource? I'll test this once I get a render done.

No ones adressed the shading under hair props here yet. This, not sure how to control it.

And backfacing (ear) scatter works great with emitters!

Hmm.  Not sure I get the fake lightsource theory.  Where, then, would this light be generated from in the visible scene?  Real world lighting requires light to come from somewhere, even if it's just diffuse light bouncing around a room from the sky outside, through a window.

I'll have to look back on the previous posts in this thread.  Been away for a while, and it has obviously taken on a life of it's own since I last  looked. :-)

Quote - I used to use emitters all the time with Luxrender, but it's trickier in Poser!

Laurie

Much trickier.  Poser isn't quite meant to emulate real world physical light behavior like Lux was designed to do.

Problem is, this contrast is the stuff that I'm attracted to. You have a nice specular range , followed by the skin, then a shading where countering light sources do not quite meet. And then a bouced light rim effect from surrounding skin.

Richardson - amazing render!

From my limited knowledge on this, increasing the raytrace bounces will reduce the edges you have pointed out (but also reduce the contract a bit - so you might need to reduce the scene GC a bit (to 1.8-2.0).

such an inspiring and interesting thread here! thank you all!

From my limited knowledge on this, increasing the raytrace bounces will reduce the edges you have pointed out (but also reduce the contract a bit - so you might need to reduce the scene GC a bit (to 1.8-2.0).

Paul, I had  that at 2.0 gc*. Wait till you see this render...

This is awesome BB. I have been playing around with this, and you hit the nail on the head here with the fresnel setup. I did one and posted it in my gallery that is sort of an off shoot of this for sweaty skin. I am still working on the shader tree for it. I was aiming for sort of a two layered approach for the fresnel skin and the sweat. I am sure my shader is not perfect, but I am getting there. I cut the gamma back on the render and turned up the texture saturations a bit to get what I was after.

I need to play around with turning the IC off as well, if it rendered that fast, it must not be to hard on the system with IC off. (I am sure that it depends a lot on the scene, bounces etc.)

As soon as I finish some other stuff up I need to play with this a little more. :)

Awesome stuff here... Lots to learn...

Damn again. I cannot duplicate it.

Thanks BB..;))

I would be great if we had backscatter. And absorbtion depth ;).

Laurie

Before you try the IC off - I was using SR3, which is not released. Pre-release software. If you have SR3, then by all means, try it.

and all for 149.9tfi... Hey, one thing about big emitters. They emit biiiig reflections. Even with soft up at 3.

Quote - Before you try the IC off - I was using SR3, which is not released. Pre-release software. If you have SR3, then by all means, try it.

I just discovered something very interesting BB.

Turn off IC, then set the bucket size to the longest dimension of the render. All the cores will work on the same bucket at once. Not sure if I will run out of memory or not yet. It is still climbing but I have plenty left before it starts paging...I am doing a 600x1000, bucket size set at 1000.

It is already displaying an SSS pass, but is still working on it according to the progress bar. (about 1/2 way thru the bar the SSS pass popped up in the display) The SSS pass looks 100% better with IC off. That in itself is a huge improvement.

The final pass appears to be adaptive, getting 170x170 chunks even with it set to 1000.

This is probably my last attempt at this render. The others I rendered in background; this I rendered in Queue Manager. And as was being reported when PP2012 was first released, I'm seeing differences between the two.

I don't know why the skin artifacts are there either (splotchiness etc), but they were also present in the background renders, although not quite as noticeable because the background renders were smaller.

I don't know why the skin artifacts are there either (splotchiness etc)

Erickl88 on page 1 "Tho, I noticed, that the bigger you make the emitters, the lesser the splotchy-mess appears to be in the end".

Or,  BBs 6th post to this thread.

B3D,

Sorry,,, I see you got that answer already earlier on.

Quote - Sounds like something to use if you've got 32GB of RAM, Scott. :)

I maxed out about 3 gig actually, I need to see if Poser can be foreced not to auto size the buckets.

Looks like it will need more bounces and careful diffuse settings to get it to come out better. A little to dark in the shaded areas. The contrast and the areas like the fingers on the bottle came out a lot better as well. I need to run this again and time it.

I'm playing with SR3 and did an overnight render, which actually finished in 3 hours 10 minutes.

The block floating over the scene is not light-bulb size, but it is pretty small - about the size of one of the chests. It's ambient value is 60. I arrived at this number after directly photographing a light bulb and comparing its radiance (is that the right word?) to the objects it was illuminating. It required an exposure 60 times smaller to photograph the bulb into a similar photographed brightness as the other stuff in the room. (Which means that real-life diffuse value is around 1/60!!!)

Usually, something that small and hot would splotch the heck out of the render. But this was without IC.

There is some noise on the back-left wall - need more samples. I did this with 5000 samples. Probably need 20000 samples and take around 12 hours.

Seems we're getting into LuxRender territory.

Quote - Before you try the IC off - I was using SR3, which is not released. Pre-release software. If you have SR3, then by all means, try it.

Talking about no IC in SR3.  So that's basically just Brute Force computation?  Brute Force is the slowest, but most accurate computational method of rendering global illumination.  It's very accurate, especially if you have many small details in the scene.  Did you try it with something like strand-based Dynamic Hair?  Wonder how well it can handle that.

Quote - ... Did you try it with something like strand-based Dynamic Hair?  Wonder how well it can handle that.

Hmmm...wonder if we might be approaching Luxrender speeds (or lack thereof) then...lol.

Laurie

Yes, brute force. And that's why I mentioned LuxRender territory, similar appearance along with similar render time.

Quote - wow...lot better. She looks great :)

Hey! You are easy to please...  I'll just go back to my backwhippin on the other pc. I should be able to duplicate maxxx's scene. I'm just not getting the blue/gray tone of his render. I wonder if he desatsurated his color maps. This one has too much red as is.

I thought I mentioned that originally.  Prob not.  90% saturation. :-)  Sorry I'm so late to the party, you probably have all this figured out by now and duplicated it perfectly.  I've moved on to some more complex scenes, and ran into other issues with the IDL.  But I won't pollute this thread with that.  Some great work being done here.  Lots of things solved I had no idea about, and just got "lucky".

Quote - > Quote - ... Did you try it with something like strand-based Dynamic Hair?  Wonder how well it can handle that.

Hmmm...wonder if we might be approaching Luxrender speeds (or lack thereof) then...lol.

Laurie

True, unless you have Luxrender plugged into a good network.  I've seen speeds upwards of 1.02 Ms/s on the network here.  That's approaching Octane render speeds using GPU.  Of course, I don't own this network, I just work on it, so I can't do all my renders from it unfortunately.  But I digress. This is all talk for a different thread.

All great info here.  Loving this tread.  I have learned a lot.  One thing kept bugging me though and I couldn't pin it down till just this morning.  When it comes to these emitters what seems to be getting missed is the shape of the emitter.  So I decided to do some test renders to see if what I was thinking was true.  Turns out it does have an effect, esspecially on shadows.

What I did is make 3 props as emitters all in the same location roughly the same size,  in a closed room.  One Round/elliptical, one as a One-sided Square/rectangle, and one cube/rectangle.  Turns out, although minute, the shadowing is different.  Which makes sense as each would direct light off of itself in straight lines in different ways.

Normally this wouldn't be that much of a bother, but if you have refective surfaces it adds another variable of how the reflected room is illuminated.  Add in the color of the walls and you have ANOTHER variable.  Quite interesting stuff!

The ball does a good job of reflecting the whole room, so if you look closely you can see the differences on the room lighting.

I guess the point of this is to show that when we do this type of lighting, we should also consider the shape of the emitter if we are to try and figure out realistic lighting senerios for our individual scenes.

Pics to follow.