okay will try that

 I tried the bigger shadow and I"m very happy with it - thanks.
I have decided I like hdri backgrounds - a lot!
Love esther

 and I like your dome and light and shadow thingy
Love esther

ok, this might be a dumb question, but...

let's say i want to do an indoor scene, and not from a photo.  this tool, wonderfully, solves the problem i always had with IBL, which is that i'm only interested in  CG scenes.  but i'm wondering about the initial lighting.  should it just be the directional lights?  should it be the directional lights with a dim white IBL?  or should i try to approximate the colors of the scene?  or does it not matter, because of how the tool works?

i haven't tried it yet, because i haven't needed to (i like plain backgrounds and empty scenes, actually), but i think i'll be wanting to have some fun with it soon.

no time at the moment--but i've actually made an interior scene and used bill's stuff to CREATE a radiosity IBL map from it.

i then create a new light using that IBL, drop a character in, and my god-it looks fantastic.

more later if you like!

r/.

yes, more, please.  specifically, in that first render, how did you light your scene?

take a look btw at my prev post (page #2) to see it.

base scene lights were strong, contrasty, simple:

1. a point light behind where the character was gunna be placed.  bluish and set so that it falls off quick--this hits the floor and back walls in a cool tone.

2. front of scene was a spot light--yellowish aiming where char was to be placed.  this made my foreground very hot/warm.

3. i added some objects to 'fake' those skylights--they are just self-illuminated boxes (turn ambient up)--they don't cast light but fake it for the IBL map.

place bill's tool (check prev posts---pretty sure you put it where the 'eye' of camera is), render out your HDR map.

1. now make an IBL light--and attach the map, drop in yer char and go!
   hope this helps

Thank very much.  I was just at your site prior to checking the forum here and I noticed your post.  Again I want to thank you personally.  Can't wait to try these.

interesting.  so you can use this IBL tool to make rough area lights.  i hadn't thought of that.  that's good to know.  i wonder how this would work with a Stonemason SF set?  he often has areas set for lights.

i am sure it would work well.

stonemason includes a spherical reflection map based off a render....so this is like making a radiosity IBL to go with!

good luck!  keep base light clean and strong

Wolf, VRay does not come with Max; you have to license it separately for Max, about $1000.

::::: Opera :::::

To be honest, I've not worked out the exact lighting to maximize your simulation of radiosity.

Synthetic gives a good workflow. Use simple direct lighting to generally illuminate the props. Place any "glowing" props you like to create localized "area lighting". Render the IBL, save it, and now load that image onto an IBL light. Render again, save load. I don't know how many times it is worth repeating. However, do note that if your props generate more "reflected diffuse" light than actually arrives (not uncommon with some of the naive shaders that just add diffuse and specular together) then repeating this process will create "LASER" - LOL. Well not really, but you will be causing light amplification through over stimulation of your shaders.

gosh--thanks, bill.

i seriously DOUBT it is accurate, BUT it looks OK! :)

I am just now learning of this and it seems fascinating. I am wandering if you could post a diagram of the calculus behind the process with side notes. 

Another thing, what do you figure would come from me having the attatched IBL light setup in the material room? I have done this to increase the range of lighting.

Sorry about the picture. I figured out how it would be affected. By the way, this is JUST FANTASTIC!! It works better than I could have imagined. You can't go wrong with calculus, it cleans things up nice and neat. I am looking into snell's law once again to try and figure out why the probe is shaped that way. Did you want to give it a handle or is it that long for a reason?

I had just recently figured out what IBL and AO was all about. I realized that it is good for the main purpose of indirect light or Global illumination. Once I had the "overall light" I could add extra lights to simulate Direct Illumination, like the sun or a lamp. Can IBL be used for Direct illumination? How can this probe make it easier and more efficient (Direct illumination that is)?

Could you by chance simulate the same effect in 3ds max or Blender. The diagram would be helpful.

Sorry about all of the questions, it is just so very interesting.

i'm not bagginsbill (or even a  lesser substitute), but the following are problems i've read about and observed in the forums with Poser IBL being directional light::

it doesn't affect specular properties
it doesn't cast shadows
it's very inaccurate and low resolution, so you won't get the same precision as with a directional light

i've been using Synthetic.'s iterative process, and starting with a plain IBL of 10 to 20% with my directional lights, and making a new IBL, then using that to generate the next IBL, and keeping my directional lights.  i think that process has worked pretty well for me.  the only downside i'm experiencing so far is that i can't capture anything in the center of the stage.  soon i'm going to make an image of a woman in a red dress in a white room, and i'd love to see the red on the walls and such.  there's probably a way to do this, but i right now i'm working on another image.

On the question of an IBL being able to cast shadows, bagginsbill has stated that, yes, it can produce a shadow. I was skeptical but tried it and it is true. IBL can cast a shadow. I didn't find the shadow very satisfactory, so I have turned shadows off for the HDRI lights that I'm creating.

let me put it differently: it can't cast shadows based on its source and therefore light.  iirc, it casts shadows as if it were an infinite light.  so you could kind of sorta fake the direction of light in an outdoor IBL.  not so great if you were using the IBL for an indoor scene, or anything with multiple light sources.  imho, that's pretty much the same as not doing it, because it doesn't cast shadows in the direction and intensity that it lights.

CB and LL:

Nice. My cult grows. LOL

You guys got it right - must have been paying attention. IBL can make shadows, but they behave like an infinite light. Good for outdoor IBL with a sun in it. Not good for indoor multi-light scenarios. Better to use IBL for fill and use actual Poser lights for distinct, strong light sources.

Suthers:

A math cultist? I will prepare some diagrams. (I will first write a program to generate said diagrams easily. I've been wanting one for a long time. There is some really fascinating stuff in math for CG, but most people cannot visualize it like I do in my head. I need a very fast math function interactive diagramming tool. Have played with a few free ones - I see flaws. Will write my own. Be back later).

In the meantime, here's a clue, and you should be able to work out some of the rest. Viewing the radial cross section of the probe as a function f(x), the derivative f'(x) = tan(x).

Some thought provoking questions for you:

What is f?

What is the value of f(pi/2)?

Why is f(pi/2) interesting or important?

How long should the probe really be?

What did I do about it and why?

Speaking of lights...

Let's say you have 3 lights in a scene. All 3 Infinite. When and why would you recommend turning on shadows for more than 1 of those lights?

I examined Bagginsbill's light set. He had 1 infinite and 1 IBL. Shadows were turned on for the Infinite.

I examined another light set with 3 Infinites, 2 out of the 3 lights were turned on. In which scenario, is it best/necessary for good lighting to turn on more than one shadow?

In general, the use of an infinite or spot  or point light is simulating a real light source (or a concentrated reflection of a light source as from a mirror) and should have shadows enabled.

The IBL should have AO shadows enabled, but Poser light-based AO is buggy prior to Poser 7 SR3. That is why I suggest material-based AO for the general public. Mat AO is also faster in some cases (because you can skip it where you don't need it) and gives more control (because you can vary the parameters across the scene.)

So the overall answer is - all lights should have shadows.

@ SSAfam1: It really depends on the particular situation of the scene as set up.  You create one dominant light with shadows turned on and then find certain areas on your figure or scene lacking highlights or detail.  By creating additional fill lights pointing towards those areas, you cast accenting illumination without introducing extra unwanted shadows.  The scene should be thoroughly examined through several renders to find the right balance pleasing to you and technically correct.

Fill lights are usually spot lights, and are there for artistic reasons, as hborre says.

If you're using an infinite light, it should have shadows, otherwise, you probably should not be using an infinite light.

The original question was about 3 infinite lights. I say shadows on all.

hborre and Bagginsbill

Understood. :)

Thank you both.

I do have a question about lights that I've been wondering about since I began Poser back in 1998. I've downloaded a heck of a lot of light sets from free stuff over the years and have purchased some. In some light sets, there have been lights present that are not even turned on. Now I ask, why have a light in the set up that isn't even turned on?  Could there be some esoteric reason for that or did the artist simply goof and forget to turn on a light while creating the light set?

The reason (hopefully) is to give the user some simple variations.

It would be like me giving you a shader option with an initial value of 0, which does nothing. You could turn it to 1 and it does something.

For example, I gave out a light set that had a rim light already set up behind the figure (tricky thing for some people apparently), as well as two main lights - one for a dramatic facial closeup and one for a full body portrait. All you have to do is turn one off and turn the other on to switch.

Why not two light sets? Well suppose you wanted to slightly adjust that rim light while working on a full body shot. Then later you want the corresponding facial closeup. If they were in separate light sets, any modification you made to the rim would be lost when you load the closeup light set.

So by combining the two scenarios, you take advantage of the common factors which you can then adjust ONE time instead of two.

Of course maybe I'm being overly generous. It may be with some older sets containing 25 or more lights that people don't even know what the hell they were doing with them.

I'm like that with more than 30 nodes in a shader. I have no idea what I was trying to do with them anymore. That's when I go to matmatic instead of the material room.

nodding I see what you are saying, bagginsbill. But I suspect that in some of those really old light sets that it was simply that they forgot to turn a light on in their set or forgot to delete the light all together. Maybe the creators of those light sets intended to do as you just explained and didn't put that in their readme. Of course, some of the lights I still have in P4 (practically fossils), did not come with readme files.

i never got why would you buy light sets. your scene is set up by you so you and only you know how to set up the lights.

just to say....

it's probably due to what i consider a bug in Poser.  there's a reason there's a "delete lights" script. i pretty much always do my own lights, so i haven't checked the situation variables involved, but when you apply a new light set, the old lights usually get turned off instead of replaced.  some people seem to have written their light setting properly and it does replace them, but a lot of lights (especially in PPP times)  just overlaid your present lights.  do that a few times, even with your own light sets, and what you've got is a huge number of lights that aren't doing anything.  add in the past use of 10+ lights to imitate global lighting, and you've got a light set that's kind of hard to manage.  iirc, there was a script in free stuff to clean up your lights.  or maybe more than one.  i can remember the thumbnail for it.  it gets rid of lights that are off.  it might still be there.

Quote - i never got why would you buy light sets. your scene is set up by you so you and only you know how to set up the lights.

well, that's how i feel.  but i will say the exception is the same as the one where i might need lots of lights.  if you have a prop or scene with a significant number of light sources, and in Poser there's no way to make illuminating materials... well, you might turn to someone else to create the effect of, say, the lights going down Stonemason's Skyline hallway or his GSFC.  long ago, i tried to light the latter with points, and really i should have tried about 4 or more point lights with particular settings to duplicate their effect.  now, i might try this IBL tool, but i'm not sure Poser's IBL is that precise.

 I use semideiu's scripts for the deleting and creating lights from shaderworks at runtimedna.
Love esther

You forgot to disable shadow catch only. You don't want the built-in shadow catcher.

Also, your transparency falloff should be 0.

Or, just go get my official shadow catcher and figure it out :)

http://sites.google.com/site/bagginsbill/free-stuff/shadow-catcher

 I'll do both.  Thankyou  very much!
Love esther

i was searching in this thread the answer but i couldnt find it.
i have in the render settings GC turned on. poser PRO. what should i do now about the IBL image? are the settings 1 or 2,2? 

Best bet, settings for IBL image should be 1.  With GC turned on for the render settings, your image will double up if left at 2,2.

Hi ice-boy.

I assume you're asking if the lens should have a 2.2 in it or not. If you are asking something else, forgive me for all that I'm about to make you read. I'm answering this question: There is a 2.2 in the GenIBL-Lens shader - what should that be if I'm doing a GC render in Pro?

This is very confusing, and I had to think this through and experiment for over an hour to verify my answer.

The answer is it depends on whether or not you are going to save your new probe in HDR or LDR format. If HDR use 1.0. If LDR use 2.2, same as always.

Before I explain anything else, let me make this point really clear. Do not use GC shaders with a GC render in PRO. If you're going to let PRO handle GC, then every shader you use should be generating linear output. Otherwise, the math gets really crazy, and the correct value for GenIBL is neither 1 nor 2.2, but rather 4.84. I don't even want to go there.

I am also going to answer this question: what should the incoming image gamma be set to when doing a GC render in Pro?

So hopefully I'll be covering everything you need; how to make a probe, and how to use a probe.

OK, first, let's talk about probes. Not necessarily ones you make with this tool, but rather IBL probes in general.

When you load a probe image into the IBL, you want it to be feeding linear data. If the incoming image is already linear, you can just plug it in and it will work, as long as PRO uses an Image Gamma = 1.0. So if you have a linear image, and you have render GC off (a linear render - old style like P7), you're fine. If you have render GC on, you must make sure that Pro is using the image as-is, so you have to set the incoming texture gamma to 1.0 for images that PRO will believe are sRGB, not linear.

By default, PRO believes all HDR and EXR images are linear, and all LDR images (JPEG, GIF, TIF, PNG) are sRGB. So watch yourself with LDR images and Render GC enabled. (Also, watch for HDR images that are actually sRGB. I've seen at least one from Dosch.)

If your incoming image actually is sRGB, you want to make sure that a 2.2 anti gamma is applied to it, to make it linear. If Render GC is off, Poser will not do this for you, and you will have to user a Color_Math:Pow node and plug in the number 2.2 in Value_2.

Summary:

linear image, linear render - just connect the image.
linear image , GC render - select 1.0 for image gamma.
sRGB image, linear render - use a Pow(2.2) node.
sRGB image, GC render - select 2.2 for image gamma OR select 1.0 and use a Pow(2.2) node.

OK? That's how to use probes.

Now how to make probes.

I'm going to make the suggestion that it is best if probes are always linear. Because you'll notice that in our little how-to-use summary, we never need a node if we're dealing with linear images. Also, linear is what the IBL actually expects, regardless of a linear render or a GC render.

So!

Here's the deal. When you do a GC render with HDR optimized output turned on, two images are produced internally. Otherwise, only one is made. You can see one of these on your screen. That is the sRGB 8-bit image. There is also an HDR linear image produced. If you save the render as HDR or EXR, it is this linear image that is saved.

My GenIBL-Lens has an anti-gamma 2.2 in it. This is because prior to Pro and the option to do a GC render, the only info we get from reflect and refract is what would normally be put out on the final image. That means the info is in sRGB format. But we want to produce a linear format render. So I do an anti-gamma 2.2 as the last step; sRGB ^ 2.2 = linear.

Now if you're doing a GC render, then the data coming from reflect and refract is already linear! That's one of the great features of a GC render - you're operating strictly on linear data. All the math works right. So it would seem that my last step is bad: linear ^ 2.2 = garbage, right?

Except that the GC render adds an extra gamma correction at the end. So the true math is this:

(linear ^ 2.2) ^ (1 / 2.2)

Recall that (x ^ y) ^ z = x ^ (y * z).

So the above becomes:

linear ^ (2.2 / 2.2) = linear ^ (1) = linear

Hah! So the 2.2 works out, but for a completely different reason!

However, that hidden HDR image that PRO makes is different. PRO does not apply that last gamma correction to it. In this case, you want to stop the lens from doing that. So you should go into the GenIBL-Lens shader and change the 2.2 to 1.0.

Now of course, your sRGB image, which is what you see on the screen in Poser, is going to look different. Pay no attention to that. Just save as HDR and the correct information will be there. I have verified this by loading the resulting HDR image back and rendering it without GC (doing a linear render). It looks the same as the LDR sRGB image.

In summary:

linear render, LDR output - use 2.2
GC render, LDR output - use 2.2
linear render, HDR output - use 1.0
GC render, HDR output - use 1.0

Quote - Best bet, settings for IBL image should be 1.  With GC turned on for the render settings, your image will double up if left at 2,2.

Nope. We cross posted. Read me. :)

BB, you have been read.  Thanks.

Quote - Hi ice-boy.

I assume you're asking if the lens should have a 2.2 in it or not. If you are asking something else, forgive me for all that I'm about to make you read. I'm answering this question: There is a 2.2 in the GenIBL-Lens shader - what should that be if I'm doing a GC render in Pro?

This is very confusing, and I had to think this through and experiment for over an hour to verify my answer.

The answer is it depends on whether or not you are going to save your new probe in HDR or LDR format. If HDR use 1.0. If LDR use 2.2, same as always.

Before I explain anything else, let me make this point really clear. Do not use GC shaders with a GC render in PRO. If you're going to let PRO handle GC, then every shader you use should be generating linear output. Otherwise, the math gets really crazy, and the correct value for GenIBL is neither 1 nor 2.2, but rather 4.84. I don't even want to go there.

I am also going to answer this question: what should the incoming image gamma be set to when doing a GC render in Pro?

So hopefully I'll be covering everything you need; how to make a probe, and how to use a probe.

OK, first, let's talk about probes. Not necessarily ones you make with this tool, but rather IBL probes in general.

When you load a probe image into the IBL, you want it to be feeding linear data. If the incoming image is already linear, you can just plug it in and it will work, as long as PRO uses an Image Gamma = 1.0. So if you have a linear image, and you have render GC off (a linear render - old style like P7), you're fine. If you have render GC on, you must make sure that Pro is using the image as-is, so you have to set the incoming texture gamma to 1.0 for images that PRO will believe are sRGB, not linear.

By default, PRO believes all HDR and EXR images are linear, and all LDR images (JPEG, GIF, TIF, PNG) are sRGB. So watch yourself with LDR images and Render GC enabled. (Also, watch for HDR images that are actually sRGB. I've seen at least one from Dosch.)

If your incoming image actually is sRGB, you want to make sure that a 2.2 anti gamma is applied to it, to make it linear. If Render GC is off, Poser will not do this for you, and you will have to user a Color_Math:Pow node and plug in the number 2.2 in Value_2.

Summary:

linear image, linear render - just connect the image.
linear image , GC render - select 1.0 for image gamma.
sRGB image, linear render - use a Pow(2.2) node.
sRGB image, GC render - select 2.2 for image gamma OR select 1.0 and use a Pow(2.2) node.

OK? That's how to use probes.

Now how to make probes.

I'm going to make the suggestion that it is best if probes are always linear. Because you'll notice that in our little how-to-use summary, we never need a node if we're dealing with linear images. Also, linear is what the IBL actually expects, regardless of a linear render or a GC render.

So!

Here's the deal. When you do a GC render with HDR optimized output turned on, two images are produced internally. Otherwise, only one is made. You can see one of these on your screen. That is the sRGB 8-bit image. There is also an HDR linear image produced. If you save the render as HDR or EXR, it is this linear image that is saved.

My GenIBL-Lens has an anti-gamma 2.2 in it. This is because prior to Pro and the option to do a GC render, the only info we get from reflect and refract is what would normally be put out on the final image. That means the info is in sRGB format. But we want to produce a linear format render. So I do an anti-gamma 2.2 as the last step; sRGB ^ 2.2 = linear.

Now if you're doing a GC render, then the data coming from reflect and refract is already linear! That's one of the great features of a GC render - you're operating strictly on linear data. All the math works right. So it would seem that my last step is bad: linear ^ 2.2 = garbage, right?

Except that the GC render adds an extra gamma correction at the end. So the true math is this:

(linear ^ 2.2) ^ (1 / 2.2)

Recall that (x ^ y) ^ z = x ^ (y * z).

So the above becomes:

linear ^ (2.2 / 2.2) = linear ^ (1) = linear

Hah! So the 2.2 works out, but for a completely different reason!

However, that hidden HDR image that PRO makes is different. PRO does not apply that last gamma correction to it. In this case, you want to stop the lens from doing that. So you should go into the GenIBL-Lens shader and change the 2.2 to 1.0.

Now of course, your sRGB image, which is what you see on the screen in Poser, is going to look different. Pay no attention to that. Just save as HDR and the correct information will be there. I have verified this by loading the resulting HDR image back and rendering it without GC (doing a linear render). It looks the same as the LDR sRGB image.

In summary:

linear render, LDR output - use 2.2
GC render, LDR output - use 2.2
linear render, HDR output - use 1.0
GC render, HDR output - use 1.0

to much info for me to handle.

i explained it like that: when baginsbill(you) made the GC shaders in poser 7 you never said we should gamma correct the IBL in the material room. so i had the GC shaders on my figure but never on my IBL. just the image. so if i now renderin poser pro with GC turned on i need to set it in the IBL material room to 1.
i will never use bad IBL images because i use your images or i make them myself with the IBL generator.

but now that i think about it..........am i right? he he

Quote - in poser 7 you never said we should gamma correct the IBL in the material room

I never said anything about how IBL should be done because it isn't changed by using GC shaders. In particular, if you're using GC shaders, you're clearly not using render GC, which means that there is no chance Poser (P7) is going to change the interpretation of one of my IBL probes. So why mention it?

Quote - so if i now renderin poser pro with GC turned on i need to set it in the IBL material room to 1.

Correct. IBL images that I make or you make with GenIBL are linear already, even if they are LDR (JPEG) images. That's why I said this above:

linear image, linear render - just connect the image.
linear image , GC render - select 1.0 for image gamma.

thanks i got it now.

different questions.
can i make from a normal reflection image map a HDR IBL? lets say i know in the image what is brighter ? should i use photoshop ? 

can we use the IBL generator if its not in the middle of the room? can we move the dolly?

Sure you can move the dolly camera to anyplace you want.

However, don't rotate it. If you do, the "front" of the IBL will not be lined up with where Poser places the image in world space.

When you are working in an interior, moving the probe will change the perspective regarding what is large (close) and what is small (farther away). Regardless of where you place them, center or not, IBL is a compromise. The actual ambient lighting near the ceiling, for example, is supposed to be dominated by the ceiling, but IBL doesn't work that way.

We can only optimize (maximize correctness) of the ambient lighting for one position in 3D space. All other positions will receive the same ambient illumination and this is not how reality works. But it's better than nothing.

I did an experiment in Poser to improve this situation. I hoped to create multiple probes, sampled from different points in the scene, and then use a multi-image light shader that would interpolate between these probe images based on actual position of the rendered object. In theory, this would have gone a long way to improve the realism, allowing a single IBL to produce different lighting at various points in the scene. Unfortunately, Poser defied me. The P (position) node does not work in an IBL shader. So I was unable to perform the interpolation.

If Poser were to implement a proper shader SDK for lights and materials, I could fix it.

i am asking because i want to make specular maps for the figure. so that way it doesnt have to be in the middle.

Quote -

We can only optimize (maximize correctness) of the ambient lighting for one position in 3D space. All other positions will receive the same ambient illumination and this is not how reality works. But it's better than nothing.

I did an experiment in Poser to improve this situation. I hoped to create multiple probes, sampled from different points in the scene, and then use a multi-image light shader that would interpolate between these probe images based on actual position of the rendered object. In theory, this would have gone a long way to improve the realism, allowing a single IBL to produce different lighting at various points in the scene. Unfortunately, Poser defied me. The P (position) node does not work in an IBL shader. So I was unable to perform the interpolation.

If Poser were to implement a proper shader SDK for lights and materials, I could fix it.

i think in 3ds max and maya they basicly project the IBL on simple geometry. so then its more realistic and more correc.t

I don't understand what you're saying about specular maps for the figure. Why would a specular map need to come from a different viewpoint? The Sphere_Map node implements reflection angle lookup - it does not use the simple surface normal the way diffuse mapping does. So even with exactly the same map, you get different results from a map plugged into the Sphere_Map node.

Why do you think max and Maya project the IBL onto anything other than an infinite sphere?

Are you saying you get different lighting from to props in different places in the scene? If so, how does it know how far away to make the virtual environment? Realistically, the floor/ground is only 5 feet away, whereas the sky is miles away. How would it know where to put these things?