I want to render a submarine underwater. I want the water to be murky green as in the attached picture. I can get a pretty good effect using the atmosphere settings, but I was wondering if it was possible to do it just with coloured lights. Basically I need something that can render fast in an animation, and the atmosphere set-up I have is going to take literally weeks to render all the shots I need.

And just to be clear, i don't want a surface with caustics and godrays coming down, thisis slightly deeper and darker. And I am not looking for something like in this picture.

Maybe this will be more of the idea of what I am trying to achieve.

Possibly a better shot of what I am after. Just a little bit more green.

You can try with a background plane, one or more planes with transparecy and textures and a little depth fog.
The transparency planes can give you caustics effects and a curved background with some lights the background illumination.
Use few transparent planes, transparency renders much slower.

Have you tried just using depth cue on the atmosphere node?

Those are the settings I used for this picture (which uses the free sub (scaled to 22% ) available at the DAZ archive here http://free.daz3d.com/free_weekly/detail.php?free_id=173 )

And here's a quick anim of the sub emerging from the depth cue effect. 

Perfect. Thanks, setting it up now.

One thing to note - the depth-cue effect in Poser is linear falloff, all the way to 0, which is not realistic. It makes the sub anim above look like the sub is materializing by passing through an opaque portal, instead of merely approaching the camera. True depth-based light transmission (such as fog or murky water) is actual an inverse linear effect, which never goes to zero at any distance, but approaches zero forever.

Some people don't notice things that are at least directionally correct, even if they are wrong in a numerical sense. This is the case with distance falloff. Some people notice, others don't. I notice very badly and it looks to me as bad as rendering without shadows.

If you find this a problem, I can show how to fix it.

I just needed something that will render very fast, and look sort-of acceptable.

This is not for any high end purposes, or for a gallery, but just to demonstrate an effect to people working on an idea.
 
If I wanted to do a high end render or animation I would just use Vue, much easier to get this effect in Vue.

I know the above is not the ideal solution, but it renders fast which is my main criteria for this.

Thanks for the offer though.

I'm talking about the above solution - using depth-cue - not something else.  It simply involves connecting a couple nodes to the Atmosphere node. If you're happy with it as is, that's fine. My point was that with no cost to speed, you can get a quality improvement.

I'd be interested in the asymptotic depth cue, BB.  I've got an elvish submarine in the works which could use it.

Poser 12, in feet.  

OSes:  Win7Prox64, Win7Ultx64

Silo Pro 2.5.6 64bit, Vue Infinite 2014.7, Genetica 4.0 Studio, UV Mapper Pro, UV Layout Pro, PhotoImpact X3, GIF Animator 5

Bagginsbill, yes, then I am definitely interested.

Just a question, what do you have your units set to in Poser? I was trying one of your shaders from one of the other threads, and it wasn't looking right, but I assume it was because I have my units set to something different to yours.

Thanks in advance...

 BagginsBill uses Inches. 

Thanks. I was using the metric system.

(Inches: a measuring system so old that even the people who invented it don't use it anymore, lol)

Oh yes we do.  Well, some of us do.  I use both.  :) 

My Store

My Gallery

I never said it was perfect! :)  I did delete about eight frames at the start of the anim in Animation Shop to make the file size smaller, so it didn't quite just materialize :)

At the end of the day it was just a quick idea that required no node wizardry for people (example me) who don't know the ins and outs of the nodes.  I'd love to have the better version of depth cue too.

And I know, it didn't have GC enabled either! :)

Hi folks. I have done some experimenting with the Depth Cue to get it perfect.

I need to go make something to eat, but meanwhile I thought I'd show you the effect.

Later, I'll post the shader, and an explanation of how it works. You don't need to know how it works to use it. But for those interested, I will show how I constructed it.

First a little explanation of my test scene.

I have two of those subs from Daz. There are also a bunch of poles. The small poles (which are red in the renders) are every 20 feet from the XY plane. Every 100 feet is a large (white) pole.

The camera is in the XY plane.

The closer sub's conning tower is even with the pole that is 100 feet from the camera.

The farther sub's conning tower is 250 feet from the camera, halfway between the 200 foot and the 300 foot pole.

The attached image shows the layout looking from the "Top" camera.

I set up a "traditional" depth cue, with the start distance at 0 and the end distance at 160 feet.

This means that the "transmission" of light from the objects is 50% at 80 feet, and 0% at 160 feet.

If you count the poles, the last one you can see is the one at 140 feet. The pole at 160 feet (and everything behind it) is invisible. The sub seems to materialize out of nowhere at 160 feet.

This is not how it should look.

Click for full size.

Now I do the math. As before, the transmission is 50% at 80 feet. But, with the correct asymptotic falloff, the transmission is 25% at 160 feet, not 0. At 250 feet (where the second sub conning tower is) the transmission is about 11.5%. At the tail of the second sub, about 360 feet away, the transmission is about 4.4%. The transmission never goes to 0.

Another absolute work of beauty BB, that last pic looks very real.
If I may indulge you for 2 things:
1- What the heck does a asymptotic look like in Poser terms, nodes, etc?
2- How did you get such realistic caustics? Again in Poser terms, nodes, etc.

Can you post pics if you are using nodes for this please?

Comitted to excellence through art.

Will do - but first I have to get in my car, go to my apartment, and eat something. I'm starving!

Attached Link: http://www.renderosity.com/mod/forumpro/showthread.php?message_id=2741994&ebot_calc_page#message_2741994

bill posted sub caustics method in 2006.  as usual, find by searching poser forum,
username bagginsbill.  inverse square falloff plugged into volume colour/density channel
is more recent addition to his earlier technique**. ** whilst carrara can calculate wave-generated
caustics quickly, mark bremmer (carrara expert) recommended method similar to bill's.
**
**

Hi all - just a quick note to let you know I haven't forgotten about this. It got me thinking about some other things, such as how the light should decrease with increasing depth, as well as with distance, and how the scattering from particles should appear.

Here's a little demo of where I'm at. The water surface is 200 feet above the bottom (ground). You can see sky through it. Notice the area of total internal reflection. But! There's an anomaly due to the fact that reflections of the distant areas, which should be dark, come from the surface which is much closer. The Atmosphere effect does not respect the total distance that info has traveled - it only deals with the direct distance between the surface and the camera. I'll have to fix that with some more math - seriously complicated math.

book

We are not impatient!! We can wait for perfection, lol, and we demand nothing less. :-)

Well, after another week's work on it, I'm happy to say I think I've got this pretty well nailed.

It took a rather large amount of work with algebra, trigonometry, and calculus. I don't often find much need for calculus in shader work. But to get the correct behavior of the attenuation and scattering of light caused by tiny particles, I needed it. I had to do line integrals in a 3D field.

78 nodes later, this is the result. Click for full size.

Now the question is - does anybody want to see how I did it, or do you just want the shader?

Note: I really should sell this shader as it is truly a masterpiece, even for me, but what the heck. I'll give it away. There's probably not a very big market for this effect.

In the previous render, the camera was at a depth of 75 feet below the water surface. (The ground is 100 feet down.)

In this render, the camera is only 20 feet below. Observe how the scattered light is arranged differently, because of the camera position.

Quote - Now the question is - does anybody want to see how I did it, or do you just want the shader?

I for one want to see the workings behind the final shader.
Maybe do a PDF so that it's easier to reproduce the calculus equations.

I am not too old to learn. I would like to see how it was done, although I am not guaranteeing that I will understand it the first time.

In this render, the camera is only 20 feet below. Observe how the scattered light is arranged differently, because of the camera position.

Impressive results. This makes the render that holds the eye,, captivates.

I wonder about the reflection of the subs under the surface of water. If you get this much "caustic" effect below, shouldn't the reflection be almost destroyed? The reflection there truly is a moving target...

This is "groundbreaking" for poser. I hope you post some rendertimes.

Bagginsbill,

If you made this a product, and included the ground, water surface, shader, lights, etc, all ready and set-up in a PZ3, that all the user has to do is to add his underwater object, be it a submarine or an Orca or underwater city, then I for one would buy this.

I think one of the great drawbacks of Poser is the shortage of ready made scenes for rendering. It doesn't help having 20 boats in your runtime, and no place to put them, or having aircraft and no decent aerial scenes to put hem into.

I always default to using background pictures for aerial scenes, which is not ideal, but the results I get with most skydomes are not as good as I get with background pictures. For most surface water scenes I just render in Vue, because the water scenes are ready made and available, in Poser I would have to create most of the scenes from scratch.

To me, the challenge in Poser is to animate and pose characters, not to spend hours creating the scene in which I want to animate my characters. There are hundreds of indoor scenes, sci-fi interiors, etc, for indoor scenes, but for outdoor scenes, Vue is always the thing I have to move to.

Just my two cents.

That's a great point, Helgard. While developing the shader, it was obvious that the Atmosphere alone wasn't convincing. I had to make a ground prop and a decent vegetation shader for it. I had to make the caustic light shader, and the water plane shader, and a custom IBL. Putting that all together in various combinations would, indeed, be much more valuable than the Atmosphere shader alone. So that's what I'll do.

richardson - You make a good point, and I think I can remedy that.  Note that the caustic effect is just something I threw in for good measure, and is not a feature of the Atmosphere effect. It's simply a procedural mask applied to the main light, and not based on physics, but just a quick cheat. I don't have a lot of reference material to go by. The interesting thing about the scattered light effect is that I didn't plan it - I didn't look at reference images and then set out to imitate that. I simply set up a fairly accurate assessment of the physics, and how it looks just came out. On the other hand, the caustics are totally faked. They are going to do what I tell them to do, no more, no less, so I need to know more specifically what they do so I can program that into the approximation.

If you guys know of some good references, that would be helpful. Particularly ones involving deep water and large objects. Most of the images of submarines I see are actually renders and I suppose if they are convincing, it would be fair to copy what they look like. But I'd prefer real photos.

Oh - forgot to say - render times are the same as not using the effect. It's really cheap. Those renders were under 8 minutes on my craptop.

When things get interesting, one of two things happen; my "craptop" lol crashes or I have to go to work...

Attached Link: http://www.dualheights.se/caustics/

Quote - Most of the images of submarines I see are actually renders and I suppose if they are convincing, it would be fair to copy what they look like. But I'd prefer real photos.

I would think that most real photos of submarines underwater would be murky and uninteresting. They don't venture into those shallow, clear tropical waters that make for attractive photos very often. By a similar token, although I don't know the physics, pretty effects like caustics drop off quite quickly with depth. You might need them for the jug of water in a sunbeam (the new CGI standard image, now replacing reflecting spheres over water) and for a swimming pool, but not for a submarine unless it was close to the surface.

There's a free (fake) caustics generator at the link, for anyone who doesn't know about it. I've seen its output being sold at stock photo sites, but that's another matter entirely...

When looking for reference images, don't look for submarine. Underwater nature pictures, like whales and sharks, etc, taken by nature photographers are generally far better than photo's taken by military photographers.

http://poseidonwebdesigns.com/Dolphins/whaleindex1.htm

Two years ago I did a lot of work with the team that made the movie U-571, they were going to make a follow-up about a Gato class submarine in the Pacific (I had to model the submarine), so I did a lot of research on submarines, but as I wasn't going to render any of the shots I didn't do the research on the underwater rendering. Unfortunately their financing fell through and the movie never got made... missed my shot at the big time, lol.

Quote - I always default to using background pictures for aerial scenes, which is not ideal, but the results I get with most skydomes are not as good as I get with background pictures.

I use a fake curved background that is a plane.
"fake curved plane"?????
Well..., it is a big vertical rectangle with one face that has several tiny horizontal rectangles intersecting it from behind. How it works? It is still a plane, but adjusting the vertical position and the number of horizontal faces sharing the same edges the plane has the same normal as it was a curved cylinder.
This is very much better than a sky dome or background cylinder, it doesn't produce texture distortion, is easily scaled (after all is nothing more than a rectangle) and produce the same illumination as a curved background does. Changing the light angle you can have a highlight in the horizon or middle sky.
The only missing thing that would be great is a light that is able to illuminate only one object and you would have total control of the background illumination.
Take a look at my gallery.

Kawacki, I see what you are saying, but what I meant was something like the attached picture. To do something like this in Poser with a skydome and a ground plane is very hard, and most probably easier to do with a background picture.

Also, any sort of aerial dogfight scene, where you want to be able to animate aircraft from above, below, behind, etc, with moving camera's, is almost impossible in Poser, and much easier to just move to Vue to do that.

one would suggest helg give bill $100 worth of freebies from helg's store as remuneration.

Helgard,

I don't get why you said that. I don't have time to render an animation, but here's a still, using a full 360 degree sphere.

Perhaps you're using crappy low-res images? Or perhaps you focal length is too long?

Click for full size. I can get the same quality in any direction. If that isn't quite convincing and you want to see another angle, let me know.

Bagginsbill,

I know it is possible, I said in my post that it was just hard to do. It is easy to open Vue, add a terrain, an atmosphere, and you are ready to start animating dogfights. I have done scenes like this before in Poser, and I know Poser can do it, but it takes time to find the right hi-res pictures, etc.

It comes back to what I was saying earlier, that there is a lack of pre-made ready to render scenes into which you can just drop ships and planes into and animate. Each time you want to do it, you have to create the entire scene, find the images, etc. In Vue those scenes already exist in the provided sample scenes.

I am not saying Poser can't do it, I am saying that there is a lack of scenes like the one you show above.

It comes down to time. If I have to animate something, do I want to spend time modelling the thing I want to animate. No, I just buy it. So if I need to animate something in a scene, do I want to spend time creating the scene from scratch.

Like with the original start of this thread. I wanted to animate a submarine in Poser. In Vue this would consist of opening an underwater scene, putting the sub into it, and moving the sub. In Poser, as you have demonstrated, you have to create a shader for us, which you said yourself doesn't work properly without the lights, ground, water surface, light gel, etc.

I know that everything is possible in Poser. If it hasn't been done, someone will find a way to do it. But for the average user, he wants to do one thing, for example, animate a dogfight. That means he has to find a sky dome, find the pictures to map onto it, etc. If he knows he can do it in Vue without these hassles, which will he choose?

Helgard

Ok, let me just justify a bit, before I start sounding like a lazy git, lol.

I have to render 900 pictures at 800x800 pixels. I have the objects I want to render. For each image I want to render, I have to create a scene. For some simple objects, like a car, it is easy to load a model like one of Stonemasons sets and render the car. But if I want to render a ship, I have to create the entire scene from scratch, because there is no pre-made water scene. Then I have to render a submarine, another scene to build. Then I have to render a plane, another scene to build. etc etc.

So I was just saying that Poser lacks pre-made scenes. If you look in the market place, how many of the products are scenes? Lots of things to put into scenes, but very few scenes.

I'm curious about the reason for wanting to render in Poser then, since you already have Vue and it's certainly better at one-click terrain creation.

Note: If you have a need to render in Poser, but want to use a Vue environment, why don't you one-click the environment in Vue, then render it out as an equirectangular projection. Then you can load that onto my Environment Sphere in Poser.

Bagginsbill,

These are product images for my site, I do not believe it is fair on the customer to render all the promo images in Vue, if the products are Poser products. It gives a false impression of what the model looks like.

Helgard

Right - figured there was a reason. That makes sense.

So - how about you export the equirectangular image from a Vue-generated world?

*So - how about you export the equirectangular image from a Vue-generated world?

*Yep, trying that now. The first two attempts weren't too good, but that was me not paying attention to what I was doing or rendering the images too small. When I do something that looks decent I will post an image.

Attached Link: http://www.oceanlight.com/html/gray_whale.html

Actually,,, anything is possible. Even the big clear subsurface reflections but we expect dark and murky when it comes to subs. Caustic effects are way more fun..

Thanks for the links to pictures and to that caustic generator program. That sidetracked me a bit while I explored how best to use that effect.

It looks to me like the caustics are most pronounced near the surface, and less so as depth increases. So I changed the main light to behave that way.

Sorry about the delay on the "how to".

So I'll just get started with the explanation. I don't have it all prepared and it will take quite a few posts. But I think it will be interesting. It was for me.

So to begin, we need first to understand how the Atmosphere's Depth Cue effect works.

There are two things it deals with. Both are caused by tiny particles suspended in the water. One is attenuation - light becomes weaker as it travels farther through the water. The other is scattering - as light strikes the particles, instead of continuing to travel forward, it bounces off the particles with a color that comes from whatever color the particle happens to be.

Let's understand attenuation first.

Attenuation is usually expressed as a fraction or ratio. Given an initial luminance coming off an object, after traveling through a certain distance, it is reduced before it reaches an observer. The ratio of the starting and ending luminance is the attenuation.

If we represent the emitted luminance with the letter I, and the final observed luminance with the letter F, the attenuation is defined to be the ratio F/I. For a given distance, this ratio is always the same, regardless of the initial luminance level.

So we can basically understand that attenuation is a function of distance (as yet unknown), that produces a characteristic ratio:

F/I = Attenuation(distance)

Clearly the observed luminance, F, can never be more than the starting luminance, I. Otherwise we're talking about light amplification, rather than absorption and scattering. So we know that Attenuation(x) can never be more than 1. And in fact it can only be equal to 1 when x = 0, i.e. when the light hasn't yet travelled any distance at all.

Let's ignore scattering for the moment, and just try to understand how (Depth Cue) DC implements attenuation, and why that is a problem.