Well schools started and i'm preparing for the science fair, and i'm gonna use A:M for it, how? 'm gonna simulate realife situations, i'm goona take picture of diffrent rooms, and other places. then i'm gonna make a simplisitc verstion of those situations. and i'l make 3 diffrent renders of the situation one with all the actual lights that are in the room( or scene ) with no radiosity one with radiosity and one useing the real lights and other lights as well to simulate it. but i can't really start on it yet, you see theres some more info i need.

1. how is it suposed to be measured, my teacher said it would be ok to show how many lights i used or how many bounces they were in the scean.

2. there needs to be a variable, she said that haveing the 3 diffrent pictures would be ok(as i said)

3. there need to be a hypothesis or problem.

I don't really know how to implement this in the presentation, i also would like it if anyone could give some links on lighting i could use. i'm also gonna show diffrent lighting styles as well like 3 point lighting( with i will pint out is exactly what they use to light the characters in the newer jimmy newtron shows). so can anyone help me plz?


>MAGE

I'd say that the purpose of your experiment would be tyring to figure out a method to recreate realistic lighting in a CG scene.

Your hypothesis might be that it can be done with something like photon mapping.

You could take measurements of the room, the positions of lights, their color, and their brightness. You'll also need to get the colors of the wall as well.

Though this is your project, so I should let you make the major decisions.

wow thx for the help i think this can work. though i think the hypothesis should be for something more like its better to use photon mapping instead of useing mutiple lights. an i never thought of color and postition measurements. your a big help, right now i'm going some work with cornell box.

still dose anyone have any links i could use? i know about yeves site but i need some sites on how lights work on real life and things like that.

well i started the project and my first render was the cornell box as an example to show. also and all my images i'll be putting my charcter in it. i also used this operatunity to test raidiosity with reflections as well.

but i have a qeustion. should the walls be effecting the other models that much?!
i'm also gonna have to rerender this due to a issue with the hair. but i dunno how to fix it for some reason the lights are having anery negative affect on him.

i also have anothr question can a raidiosity scean render right with one light? for some reason when i render this one scean with one klieg light its stays at 0% for 8mins and once its finaly done(at about 40mins) it dosn't even look like it was raidiosity.

EDIT: one more thing when i rnder it with 3 lights it looks just fine, but weird in some places '.

and can any one xplain how in the world this happend?!

well i did another render and the problem seems to be fixed,but i don't like the light, its not soft at all, for some reason it does this when i use a bulb light can anyone help?

Yes they should; however, you need to give your image a gamma correction so there is less contrast in the colors. I suggest a gamma correction of 1.5 to 1.8. You can save the image as an EXR file, and adjust the gamma in there if you want more freedom. This is with a gamma of 1.59:
[attachmentid=19693]

I would also suggest using more Jittering for that shot. 1000% should be fine.



Radiosity does work with one light. I do it all the time. Make sure that you have Ambiance Occlusion turned off, as well as any ambiance post effects.



You need to increase your Sample Area.



Make sure that your light casts more than 1 ray, that it has 100% Darkness, and make the bulb the size of the light bulb inside of the lamp.

What type of light did you use for the lamp? If you used a bulb, than the problem is photons. Some of your photons became trapped inside the lamp cone and eventually escaped with an incredible amount of stored brightness resulting in super bright photon stamps which lead to the mean old artifacts. To prevent that from happening in the future, you should use a klieg with 100% width softness and make sure that the light cone does not strike any surface; I think Yves said to make the fall-off distance half-way between the light and the nearest lit surface. Another way to get rid of those pesky artifacts is to use extream amounts of jitter.

I hope that helps

i cahn't use a klieg for some reason when i use a klieg it stays at 0% it stayed there for 37mins doing absolutly nothing!!!

>MAGE

You must be setting up the choreography incorrectly. Kliegs were used in your Cornell Box render, and that was rendered. Look at that setup and see if you can mimic it to fit within your scene.

well actualy it works when i use 2 kliegs, but only then.

well i final got it to work but i had to use 3 lights. 1 bulb(o%intensity) and 2kliegs(same exact cordinates and settings) just to get it to render. and it came out with this, i'm satise fire except for the weird stuff on the edges.

well i rerenderd the cornell box and set the gamma to 1.59 and recived nasty results as shown here, maybe i'll try lower...much lower

hmm thats weird this one with nor raidiosity look fine.

hey, i dicided i'd up the look of the project by making a little mascot to go with it. i just showed it to my teacher and she loved it!

every one meet mobo!

It seems that your project is diverging into more of a "CG lighting" project instead of a recreation of an actual environment in CG.



Render out your scene using the EXR file format, then use A:M Composite to add gamma to the image. This way, you'll have more precision in adjusting the gamma correction in the scene.



Fine as in how? If the scene had the same gamma correction as the radiosity one next to it, I would report that as a problem. If you mean that it looks accurate, I would have to disagree with you. The walls would be casting a large amount of light onto the surrounding surfaces, and that is not visible in the render without radiosity.




Nice model; however, very reflective models will not show much in the way of radiating light. The only thing that will come off of Mobo's surfaces will be the color of the light at its intensity at an oposite angle, so you'll get caustics which will not brighten the scene very much.

hmmm, mabe i could down the reflectivity a bit.

and i mean that it dosn't look like its raidiosity other, all super bright and freaky looking.

and ok i'll try rendering to that.

yeah it looks like that now, but i havn't had the chance to make a simplified vertion of real life areas, like my science class for instance.

Though its s teep learning curve, if you want to render light, I recommend Radiance.
http://radsite.lbl.gov/radiance/HOME.html
It uses real-world units - watts/square meter, lumens etc.
You can download specific lighting manufacture's IES files - which have a measurement of the light distribution from a fixture.
It's free and runs on Mac OSX, linux, or a paid version for windows.
I model in Hash and export OBJ files to render in radiance...
some samples I've done
http://rastermon.com/3Dillus/rad11.htm

caveat - I have'nt used Hash since version 9 (2002), so it may have more capabilites, but I doubt if they are real units - such as the wattage of your bulb.
Good luck

Animation:Master has photon mapping now.

There are a lot of controls for lights in A:M such as color, intensity, fall off, fall off softness, width, and attenuation; all of these affect the radiosity in the scene. Some of the best things that improved radiosity were the ability to have multiple ray casts per light, and the addition of multipass.

The Sponaz example from Radiance look very splotchy. A:M uses jittering to make the radiosity cleaner looking. You should check out the newer versions.

oh man i forgot about this post, i'm sorry. you see i just moved a few days ago and at my new school theres no science fair.....sorry, i'm dissapointed my self.

I should upgrade - photon mapping and hdr would be cool. Let alone, I have to run Hash 2002 in OS9.

Radiance has many settings similar to A:M's sampling percentage with experiment you can remove splotches.
I'm still learning it.

Not meaning to be rude and interject another App on a Hash forum, but would there be a way to model an accurate discription of a 100 watt lightbulb for example? Perhaps that's not Hash's domain - art not science.