Sun, Feb 16, 6:52 AM CST

Renderosity Forums / Poser - OFFICIAL



Welcome to the Poser - OFFICIAL Forum

Forum Coordinators: RedPhantom

Poser - OFFICIAL F.A.Q (Last Updated: 2025 Feb 15 11:01 am)



Subject: A Dummies Guide to Indirect Lighting in Poser 8


momodot ( ) posted Thu, 13 August 2009 at 5:44 PM

file_436863.png

Okay... here 2 white spot lights at 50%  = intensity 100%. Maybe the key is keeping all the lights at 100% in total?



Believable3D ( ) posted Thu, 13 August 2009 at 5:48 PM

Momodot, here's my opinion: The light sets of 6-8 lights were workarounds because Poser was limited. It didn't have any form of global illumination (unless you count ambient occlusion, which is a very poor form). In real life, how often do you see 6-8 lights pointing at something? (I'm not talking about a supermarket with non-pointed fluorescents every few feet; I'm talking about anything remotely resembling the system usually used heretofore in Poser.) Pretty much never; they were only used in Poser out of virtual necessity. If you don't have that necessity because something better and more realistic is on offer....

So I would say, Yes - forget about your old light sets. IMO, of course.

Assuming an actual scene for the IDL to work from (if there no dome and a solitary object, IDL is pretty much pointless), I'm getting better results with IDL with a couple of quickly-placed lights than I was the overwhelming majority of the time when I was either spending hours setting up my own lights, or using paid-for light sets from other people.

______________

Hardware: AMD Ryzen 9 3900X/MSI MAG570 Tomahawk X570/Zotac Geforce GTX 1650 Super 4GB/32GB OLOy RAM

Software: Windows 10 Professional/Poser Pro 11/Photoshop/Postworkshop 3


Whichway ( ) posted Thu, 13 August 2009 at 5:58 PM

file_436865.jpg

Ok, this one looks flawless to me. What do others think? Render time is kind of long at 00:39:04.486 on my dual core laptop, but might be ok for "production". I'll see if I can knock that time down some.

Whichway


Whichway ( ) posted Thu, 13 August 2009 at 5:59 PM

file_436867.jpg

Settings for the slow but very good one above.


momodot ( ) posted Thu, 13 August 2009 at 6:23 PM

****Whichway, I believe I would think that was a photograph if I saw it out of context.




Whichway ( ) posted Thu, 13 August 2009 at 6:45 PM

momodot - :biggrin:


pjz99 ( ) posted Thu, 13 August 2009 at 8:22 PM

It looks pretty good, but there are still some faint blotches in the upper square towards the foreground.  It may not show up easily on all monitors, you may want to look at it with some contrast increased adjustments in your image editor.  The occlusion shadow looks very good though.

My Freebies


pjz99 ( ) posted Thu, 13 August 2009 at 8:27 PM

Quote - My tests are telling me I need to give up the old six and eight light sets I have been using and to work with only one to three lights but I am also getting the feeling there is some sort of interaction between RayTrace Bounce settings and Indirect Light Sample settings... I haven't worked it out :)

You can still use multiple lights, but since Indirect Lighting will always brighten the output more than old ways of lighting, you will want to tone down intensity settings pretty much everywhere.  The fantasy suit render I showed earlier used one main light at 100%, upper right foreground, and one fill light  at 40% lower left background.  The nearer your subject is to one or more bright-colored surfaces, the more light will be bounced onto it from those surfaces, so your sphere in a corner test exaggerates the light.  You may also want to try turning on HSV Exponential tone mapping with an exposure of 2.2, it's partly intended to compensate for that "blowing out" to full white that you got on the first try.

My Freebies


pjz99 ( ) posted Thu, 13 August 2009 at 8:28 PM

Quote - So I would say, Yes - forget about your old light sets. IMO, of course.

Yeah pretty much all old-school Poser light sets are garbage now.

My Freebies


Whichway ( ) posted Thu, 13 August 2009 at 8:54 PM

Quote - It looks pretty good, but there are still some faint blotches in the upper square towards the foreground.  It may not show up easily on all monitors, you may want to look at it with some contrast increased adjustments in your image editor.  The occlusion shadow looks very good though.

Ok, thanks.I'm wasting more time trying to save a last little bit of render time  than I'm gaining. I'll shift gears a bit and try to improve the render instead. I think it's  approaching optimum. How does it feel for render time to you? You should get a good factor of two improvement over my machine if the previous tests are any clue.

Whichway


pjz99 ( ) posted Thu, 13 August 2009 at 9:06 PM

What model of CPU do you have?

My Freebies


lkendall ( ) posted Thu, 13 August 2009 at 9:21 PM · edited Thu, 13 August 2009 at 9:23 PM

Excuse my ignorance on this, but I notice that several experienced users in this thread point out that smoothing polygons can cause problems for indirect lighting. I have Smooth polygons turned off right now in my Render Settings.

I also noted here that transmapped hair can really slow down rendering with IDL and Raytracing. I can confirm that. I left my computer rendering a scene with two figures in a room with four lights and some props. With modest IDL settings, the render stalled for about 12 hours on the transmapped hair on one character, but had no problem with the other because the transparency maps were turned off (the hair looked wrong but it rendered fast). I saw that turning off Raytracing on hair could speed things up.

Now my question, when I went to Properties for the hair to turn off Visible in Raytracing, I also found that Smooth polygons was turned on for the hair. Which setting takes precedent (the Render Setting for Smooth polygons or the item's Properties setting)?

LMK

LMK

Probably edited for spelling, grammer, punctuation, or typos.


bagginsbill ( ) posted Thu, 13 August 2009 at 9:27 PM

Both have to be on for smoothing to take place.

And the problem with smoothing is not with IDL. It is with raytraced shadows. It is not a new P8 problem. It has always been there.


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)


bagginsbill ( ) posted Thu, 13 August 2009 at 9:30 PM

PJ - I totally misunderstood the tone mapping math. The 2.2 is not like GC 2.2, although there's nothing wrong with 2.2. I just wanted to point out that 2.2 is not special the way it means something important for GC. It's quite arbitrary actually.


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)


pjz99 ( ) posted Thu, 13 August 2009 at 9:38 PM

Thanks, I don't really have the faintest idea of what the exposure setting "should be".  2.2 seems to work well enough though.

My Freebies


lkendall ( ) posted Thu, 13 August 2009 at 10:17 PM

Thanks. :)

Another question.

I notice that some figures and items have the SAME image loaded in several 2D-image map nodes, and plugged into different channels on the root node. Why not plug all of the channels from the root node into the same image node instead of two or even three nodes with the same image?

LMK

Probably edited for spelling, grammer, punctuation, or typos.


pjz99 ( ) posted Thu, 13 August 2009 at 10:42 PM

I think the technical term is "noobitude".

My Freebies


Whichway ( ) posted Thu, 13 August 2009 at 11:46 PM

pjz99 - The CPU in particular is Intel Core Duo CPU T9300 @ 2.50GHz 4.00 GB RAM.

The system is a Dell XPS M1530 with 15" 1920x1200 screen, Vista 32-bit.

Whicway


Whichway ( ) posted Fri, 14 August 2009 at 3:53 AM

file_436893.jpg

Last in this series, I think. About 1hr 31min render time.

Whichway


Whichway ( ) posted Fri, 14 August 2009 at 3:53 AM

file_436894.jpg

Settings for same.


pjz99 ( ) posted Fri, 14 August 2009 at 9:22 AM

There is one spot that shows a transition that isn't perfectly smooth, but it wasn't perfectly smooth in the render I did earlier at nutty quality:
http://www.renderosity.com/mod/forumpro/media/folder_9/file_436567.jpg

Something that concerns me is that your sphere hovering between the cards is hardly catching any light, or it's getting averaged out by the large sample size.

My Freebies


ice-boy ( ) posted Fri, 14 August 2009 at 9:24 AM

why is there a reflection on the plane?


pjz99 ( ) posted Fri, 14 August 2009 at 9:26 AM

In what pic?

My Freebies


ice-boy ( ) posted Fri, 14 August 2009 at 10:15 AM

in the last pic with the green ball.


ice-boy ( ) posted Fri, 14 August 2009 at 10:16 AM

can someone tell me what settings to use for test rendering with IDL?

what settings for a normal render?

and what settings for a good render?

thanks.


bagginsbill ( ) posted Fri, 14 August 2009 at 10:43 AM

Quote - why is there a reflection on the plane?

Don't be confusing. In Poser you know what Reflection is and if you use that word we think you're talking about reflection. I see no reflection - I see color bleed - diffuse inter-reflection, aka IDL.


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)


ice-boy ( ) posted Fri, 14 August 2009 at 10:49 AM

file_436926.JPG

i mean this


bagginsbill ( ) posted Fri, 14 August 2009 at 10:56 AM

You circled more than one thing. You circled an occlusion shadow on the large rectangle, a sphere, and a small rectangle.

Which surface are you talking about?


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)


ice-boy ( ) posted Fri, 14 August 2009 at 10:59 AM

so this is real geometry?


pjz99 ( ) posted Fri, 14 August 2009 at 11:24 AM

Oh yes.  That was chosen to force the need for >1 GI bounces to be really obvious.  Insufficient GI bounces means that those parts of the scene are not illuminated at all.  A few people kept saying that forcing GI bounces to 1 was a great workaround, this scene is one example of why it's not.

My Freebies


ice-boy ( ) posted Fri, 14 August 2009 at 11:27 AM

ohhh i thought that this was a big artifact .

thank god its not .


Whichway ( ) posted Fri, 14 August 2009 at 11:33 AM · edited Fri, 14 August 2009 at 11:36 AM

{Sorry, nap time demanded it's due.}

Yes, it is real geometry. The light on the small card gets there by bouncing off the floor or big sphere, then off the big card, then off the small card and into the camera, for the shortest route. Some of it bounces off the small ball before hitting the small card. Other intricate paths are possible, up to seven bounces. Some light leaks between the two cards giving the gap. Light can sneak all over the place, which is why it  takes a lot of CPU time to track it all down. Nothing has any intrinsic texture on it.

:biggrin: It had better not be one big artifact! I worked pretty hard to get it looking like that.

Whichway
 


Whichway ( ) posted Fri, 14 August 2009 at 12:03 PM · edited Fri, 14 August 2009 at 12:09 PM

Quote - Something that concerns me is that your sphere hovering between the cards is hardly catching any light, or it's getting averaged out by the large sample size.

One thing, I made the small sphere red to tag its bounced light - that may be throwing you off. I could do this one more time at these settings with the balls white again if people can stand it. I'll try to catch the red dot phase as well so you can see where the real evaluation points are. They are being placed by the max error limit tracking of the base algorithm; Irradiance Sample Size in not playing any role here. The way to get more accurate light on the small ball would be to lower max error, which means a tiny increase in the second Irraniance Cache value, say from 99.0 to 99.125. 

[Side comment: When working for very high accuracy, this is a very touchy control. The Max  Error = 100 - IC_setting, which means for very demaning work, all the useful range of control is compressed into the last little part. Cutting the error in two means cutting the remainder of that equasion in half, and that soon gets to be a small fractional change. You'll have to type it in as the slider is nowhere near a subtle enough control for the job. To make matters worse, the number of real evaluation points grows roughly as the square of the change, e.g., cutting the error in half will quadruple the number of real evaluation points and that's where all the CPU time goes. Fortunately, I think this is turning out to be a very demanding test scene. A simple figure study I'm trying out seems to be doing fine with second IC down at 88, or 12% Max Error. The texture hides the diffuse illumination error to a certain degree and there are three lights in a studio setup so there is more direct light most places to swamp the error of the bounced diffuse.]

 
ice-boy: Just a little bit more patience, please. I think we've got this thing on the run and will be able to gives rules of thumb for trying out soon.

Whichway


Whichway ( ) posted Fri, 14 August 2009 at 12:15 PM

file_436948.jpg

The studio render finished while I was typing the last replies, so here it is. Comments on render quality solicited. In particular, note the shadow under his right arm and pec. Without IDL, this area would be completely black.

Whichway


Whichway ( ) posted Fri, 14 August 2009 at 12:22 PM

file_436949.jpg

Here are the settings. Note the comparatively much lower values of the second Irradiance Cache and of Samples, and the lower value of bounces. This scene is *much* less demanding than the test scene, and so allows for much looser settings and *much* faster rendering than would be needed for the settings used on the test.


momodot ( ) posted Fri, 14 August 2009 at 12:47 PM · edited Fri, 14 August 2009 at 12:55 PM

So is there a way to reduce the strength of the IDL fill light effect? It seems a little strong to me IMHO... I have had to blend non-IDL renders with IDL renders 50/50 to reduce the effect which is a pain since the lights for the two renders have to be at different intensities... setting the diffuse-value on everything to 0.8 helps a great deal but still it doesn't strike me as the most intuitive way to bring down the "light in the shadows" of an IDL render.

I just feel that the contrast is too low on the IDL renders right now... like it was with too strong IBL. I want my shadows filled a bit like I'm getting bounce off the floor walls etc. but to my novice eyes the shadow fill in the standard IDL renders is much stronger as though reflectors, white cards, or even slave flashes are responcible for the fill rather than what IRC I would call "ambient bounce"... a lot of fill is coming off even darkly colored floors... is there a simple way to just turn down the effect a bit?

My favorite shooting IRC was 80/20 bounce... I would aim eighty percent of my flash at the ceiling angle to come down right on the figure with 20% bounced straight off the flash onto the subject as fill. It gave the impression of overhead 'existing light". With a fixed rig I would light with a flash high off the camera on an arm and fill in from the sides with mini-slave flashes... all that is besides the point though... the issue is that I really love the IDL but I want the overall ambient quality toned down a bit so I have more contrast... are there any direct render settings that can boost my contrast or mainly just turn that indirect light down a notch or two?



ice-boy ( ) posted Fri, 14 August 2009 at 1:32 PM · edited Fri, 14 August 2009 at 1:33 PM


momodot

do you see on the right in the middle  '' intensity'' ?

 this is how you control it.


Whichway ( ) posted Fri, 14 August 2009 at 1:59 PM

file_436958.jpg

All white version of the final render is finishing up. Meanwhile, here is most of the real evaluation point distribution. I think there are enough points on the small ball to get it's illumination right given it's small area on the screen, but others may disagree.

Whichway


Whichway ( ) posted Fri, 14 August 2009 at 2:02 PM

file_436959.jpg

And here is the full white render. Settings are identical to the ones posted earlier for the final tinted version.

Whichway


momodot ( ) posted Fri, 14 August 2009 at 2:46 PM

Thanks ice-boy. Way cool. I had not tried the D3D script yet... did not see that option on the Render Settings panel in P8 but will look and try the script in any case :)

Whichway... that is cool. Have you tried it with different surfaces on the ball? E.g. clay, plaster, skin, chrome... be interesting. Maybe something funny like wax. I really think I would take that for a photo.

Has anyone tried to make a P8 IDL render with a simulated glass object? I wonder if the light will transmit through an object with refraction and what happens with SSS etc.

I am curious... what is the render overhead of using an object with ambient now? Can we start using objects with ambient instead of traditional lights?



Whichway ( ) posted Fri, 14 August 2009 at 3:30 PM · edited Fri, 14 August 2009 at 3:42 PM

momodot - > Quote -

Whichway... that is cool. Have you tried it with different surfaces on the ball? E.g. clay, plaster, skin, chrome... be interesting. Maybe something funny like wax. I really think I would take that for a photo.

Has anyone tried to make a P8 IDL render with a simulated glass object? I wonder if the light will transmit through an object with refraction and what happens with SSS etc.

I am curious... what is the render overhead of using an object with ambient now? Can we start using objects with ambient instead of traditional lights?

I haven't tried different surfaces. There is a limit to how much of this I can do on my own. The basic test scene is available at http://cid-b233dcaeefa9709c.skydrive.live.com/self.aspx/Public/Poser8%20samples/Simple%20GI%20Test.zip and the settings I've used are posted on this page or the one before. It would be great if you tried those surfaces for us and posted the results. Besides, you'd start getting your own experience playing with these knobs.

There is no real SSS in Poser 8.

Using objects with ambient as effective area lights was done early on, but I admit I can't find the posts now. Maybe bagginsbill or someone else remembers where they are. I didn't follow things closely at the time. It's another interesting study you or someone else might take on. :biggrin:

Whichway


pjz99 ( ) posted Fri, 14 August 2009 at 3:31 PM

Quote - And here is the full white render. Settings are identical to the ones posted earlier for the final tinted version.

That looks pretty good then.  Nice job :)

My Freebies


pjz99 ( ) posted Fri, 14 August 2009 at 3:33 PM

Quote - The studio render finished while I was typing the last replies, so here it is. Comments on render quality solicited. In particular, note the shadow under his right arm and pec. Without IDL, this area would be completely black.

The only thing that jumps out as out of place is the very dark shadow on the pant strap at his right hip.  Is the strap just standing away from the body a bit and making an actual shadow, or is that a splotch type artifact?

My Freebies


Whichway ( ) posted Fri, 14 August 2009 at 3:39 PM

It's actually standing away from his hip. I haven't figured out how to fit clothes very well yet. Hence my concentration on beefcake. :biggrin:

Whichway


pjz99 ( ) posted Fri, 14 August 2009 at 3:43 PM

That's fine then, looks good otherwise.

My Freebies


Whichway ( ) posted Fri, 14 August 2009 at 3:44 PM · edited Fri, 14 August 2009 at 3:47 PM

Thank you! :biggrin: :biggrin:

I'll have to order my next tests, but a higher resolution version of the studio shot with timing will be on the list if no one objects.

Whichway


momodot ( ) posted Fri, 14 August 2009 at 9:15 PM

Man, I tried that test .pz3 no way it can render on this old machine of mine... even turned down to draft in hung.

Big question... where is the 3D3 Render script you are using... the one I download from the 3D3 website has no P8 settings, only P7 settings.



grichter ( ) posted Fri, 14 August 2009 at 9:21 PM

from the scripts drop down menu Scripts>partners>Dimension3D>RenderFireFly

Gary

"Those who lose themselves in a passion lose less than those who lose their passion"


Believable3D ( ) posted Fri, 14 August 2009 at 9:21 PM

It's not a download. It's in the main horizontal menu at the top: Scripts > Partners > Dimension3D > Render Firefly.

______________

Hardware: AMD Ryzen 9 3900X/MSI MAG570 Tomahawk X570/Zotac Geforce GTX 1650 Super 4GB/32GB OLOy RAM

Software: Windows 10 Professional/Poser Pro 11/Photoshop/Postworkshop 3


momodot ( ) posted Fri, 14 August 2009 at 9:24 PM

Thanks!



Privacy Notice

This site uses cookies to deliver the best experience. Our own cookies make user accounts and other features possible. Third-party cookies are used to display relevant ads and to analyze how Renderosity is used. By using our site, you acknowledge that you have read and understood our Terms of Service, including our Cookie Policy and our Privacy Policy.