Works in Progress

#281

Jo B.

Thanks!

And thanks for your suggestion. Uh… you might want to stop looking deeply into characters faces, though. :wink:

Would you settle for a cheetah?

Jim

#282

A cheetah would be cool, but I’m wondering how you will stylize any animal to fit wit hthese two. I really do hope you make something just to see what it looks like :wink:

#283

How about a Borzoi (ie. Russian Wolfhound) or an Afgan Hound? They would look swell with those upper crusty types.

#284

Here is a little critter that I made a while ago. I would like to hear any comments. One thing that I am wondering is weather he might look good with hair. If I put hair on him though, should I make him thinner?

#285

John,

Cool! That looks made of clay. (Is it?) Nice character to it.

Jim

#286

Thanks Jim, I rendered it with the new photon mapping gizmo. Heres how )mini tut:

Make a room for the character with a sky dome and a floor.

Set that and the characer to about 85% radiance

Have a klieg set up to point at the dome but make the fall off short of hitting anything

Calculate radiosity (I use the max level of light rays and no more than 6 bounces)

Once this is done, light the model with conventional lighting, set the photon samples and final gathering samples up high and press go (don’t use multipass).

Wait.

Wait.

Wait.

You should get somthing like the above render.

You can also render the standard lighting out seperately then composite the two afterwards (a good idea for getting more control and so that you can change your mind about the direct lighting and not have to wait for the radiosity bit).

One thing to note though is that this method means that you won’t get the neat effect of strong light cast un-evenly projecting about the place as the radiosity isn’t calculated with the light hitting the model, but it is reliable and basically works (for me).

I have noticed that photon mapping has changed recently and the sample area has more effect. This is an aspect that I don’t understand too well so you might have to consult Eyevs’s page (hope I spelt his name right).

Photon mapping looks better with an even light as artifacts don’t show. It is enough to get that nice global illumination look but you don’t get much drama.

One thing that I am considering is using the Photon mapped image with no shadows or other lights as a reference for making an illumination map for characters. It would be a kind of occlusion map that basically makes certain recesses etc gloomy. Then you could get a global illumination look in animations.

Has anyone tried this?

I recommend photon mapping as a way of viewing your models afresh. You get a good sense of the mass of the forms compared to normal rendering.

Hey, that must have been a good hundred words and I didn’t bitch once!

#287

LOL, I was going to post the same thing as Jim. Beautifully lit . I also love that character.

Now for his personality…make him evil, MWA HAHAHAHAHA!

#288

Great model and lighting John, I was convinced that was a picture of a little clay model.:eek:
Interesting thoughts too on the photon mapping for animation.

#289

This is the first model out from A:M that “cheated” me that it was a photo. I didn’t know that photons in A:M has so much evolved since I was playing with them last time. Way to go Yves! :slight_smile:

Good Work John! And thank you for tutorial!

#290

Yes, I have played with all of the incarnations of radiosity since v 8.5 (I think that I counted three major versions). Yves has realy done a good job on the latest one and it is the first that I have been able to get decent results from. I have a few issues with it but, like I say, I am not the greatest of experts. I believe that Yves is going to continue working on it and I think that it is in good hands. If you are reading this Yves, your work is v. much appreciated. :beer:

#291

John, you said “I have a few issues with it”

What are they? Are you finding limitations with it?
Thanks-

#292

Well there are a few artifacts that pop up. Sometimes I get a black mark here or there (look at where Nirmals hind left foot touches the ground). Also there is often some visible banding or specs that are hard to get rid of (look just above the horizon). Also, I would like the ability to cast more photons. I tried to make a scene made from a whole load of cubes (over one undred) and there just wern’t enough photons to go around. This caused a kind of randomness of light distribution. Maybe this is a limitation of photon mapping in gereral though. I would quite like to see more types of radiosity solution. Bakable radiosity and GI both still have their place but I guess that there is only so much time to put into that kind of thing.

It is good that the current implimentation is as good as it is though. If It has realy convinced people that they are looking at a photo of a clay model then that certainly says something.

#293

This is a robot model that I have been working on. I created it over a few evenings for a film I am working on. There are a couple of really stupid modeling mistakes that I have to correct but otherwise I think he`s done.

In the film he`s been programmed as a stand-up comedian in the belief it will make him popular with humans.

http://www.dalemation.free-online.co.uk/CHESTAWIP.html

Dale.

#294

John Keates: Wow! It’s nice to see these kinds of results with A:M’s radiosity. Give us more! :wink:

Dalemation: Nice robot. Lots of character:thumbsup:

#295

John-you can increase the amount of photons that are sent to the objects. If you go to settings in the pws you can shoot over 100,000 photons if you want. You can increase sample areas to. I think you can increase any part of photon settings. I found that to make this work that the objects should be closed-not used in a open chor. That way photons will shoot into space forever! Not waht you want. I always create a room and throw the objects into it. I think photon maps can be just as good as any other GI renderer. I had to read Yves website like 10 times before I knew how to set up a chor. Theres a lot of power in there. The only problem is that its a huge render time. But after taliing to my friend who uses C4d whenever he uses any GI or radisoity the render times are just as bad-somits not all AMs fault! :>

#296

Hey Dalemation:

I like that robit so far. He’s got good character showing thorugh in that pose.

#297

Originally posted by John Keates
Yes, I have played with all of the incarnations of radiosity since v 8.5 (I think that I counted three major versions). Yves has realy done a good job on the latest one and it is the first that I have been able to get decent results from. I have a few issues with it but, like I say, I am not the greatest of experts. I believe that Yves is going to continue working on it and I think that it is in good hands. If you are reading this Yves, your work is v. much appreciated. :beer:

Thanks John,

I’m always pleased to see what other do with the Photon Mapping. Here are a few notes:

The maximum number of photons you can cast in a scene is 1 million. In some circumstances like a very complex scene, this may not be enough. But on the other hand, there are other ways to compensate.

The main way to compensate is to increase the sample area. Sample area is in 100th of cm. In other words, 100 means 1cm. Usually, you want to set it to a size that will cover all the surfaces with photons. If you find a setting that works well for a scene and add some objects in the scene thus increasing the total surface to cover, you should increase the sample area to compensate. You can use some maths to help you figure out some good starting points for the sample area. But in the past, when I gave those maths to users, they where rather turned off. So if you are not afraid of the maths, I can give them here. There is no calculus involved. Only surfaces estimations and transformations.

It is somewhat possible to get rid of some artifacts, most notably the bandings by increasing the jittering.

The “Calculate Rasiosity” step is used only for the Quick-Render or any render to screen mode. Whenever you launch a render to file, the Radiosity is recalculated. Also, if a radiosity have been already calculated for a scene, either through the “Calculate Radiosity” or because of a render to file, Then the resultant Photon Map is used for quick-renders. So for your example, even if you calculated the radiosity without the main light on, it will not be taken into consideration after you add the main light because a new photon map will be calculated.

Photons are not stored on the first hit. This means that, in your setup, the photons are not stored when they hit the skydome. The first hit illumination contribution is better and more precisely calculated with the ray-tracing engine. In situations where a skylight effect is desired, this poses a problem. To compensate this, I have modeled a hyperbolic mirror that can be placed in front of a klieg and which shoots and stores back the photons on the skydome. I plan to give this mirrir an the mirror-klieg rig when I’m finished with this part of the tutorial.

Yves.

#298

I`ve just finished an animation test on the robot I rigged last week.

http://www.dalemation.free-online.co.uk/ChestaanimWIP.html

(DIVX 1.83MB)

#299

test for set light not finish yet
click on my web site

http://www.mediastation.150m.com

#300

I’d like some feedback pls,
done in 8.5.