[BooMSticK]

I'm not so sure there can be set a simple rule to a complicated case like this luna. There's just to many things influencing objects shade or light - bouncelight, translucecy etc. And I guess thats why number 2 seems more 'correct' since it's the most neutral of the two. Really contrasty shadows can happen for sure in nature, but i guess that number2 just seems more general and therefore more correct.

This, shadows adding to shadows thing is an interesting thought, and I have also played a little with it myself. I just never found a solution that could explain it properbly so I guess I just let it rest back then. Thanks Luna for bringing another great subject to life! (again!)
,boom

[balistic]

Neither is really correct, because the sun isn't the only lightsource in the scene. You've also got a hemispherical lightsource called the sky to consider.

Skylight will creep into the cast shadow of the tree, but the further you get beneath the branches, the less of that light can make it in. So your shadow will have a gradient of value, with the darkest sections being those that can see the least amount of sky (usually right around the trunk).

Put yourself at a point in a shadow and imagine yourself laying on your back, looking at the sky. How much of the sky can you see from that point, and how much of your view is obscured by objects?

For a 3D example, see the base of the diagonal beam in "American Flat":

http://www.bprince.com/AmericanFlat.jpg


Or:



[ April 12, 2002: Message edited by: balistic ]

[antx]

I would also say pic 2 is correct cos the dark side gets it�s light from the environment only and not from the sun directly. And this is true for both, exposed parts and parts in the shade.

If you really want to be certain: You�ve got a camera. Shoot some pics and check it out in PS via selecting certain areas and use the histogram function.

[Bg]

Light can very tricky, but you can simplify it many ways: Comics are a good example, lighting in them has it's own rules which are not tied to realism, and the result is often far from realitic.

The other example is 3d programs, they started with simple and faked lighting, but with global illumination and sub-surface scattering it's getting very close to realism. Even the current standard, direct illumination, gives quite realistic look if used right. Direct illumination can be emulated by human quite precisely in simple scenes but in complex scenes computer takes the lead.

But the global illumination is the real beast:

The render times of GI should give you the idea of its complexity: a scene rendered in GI can take 10, 100, 1000 (or more) times longer than DI! This can never be emulated by humans, even a matte cube (the exercise anyone?-) is unbelievable hard to do paint realistically, let me spoil this... it's impossible. But there's still some other things affecting lighting, this is the part where we can cheat: 3d applications can't do anything else than to obey the rules that were programmed to it, but we, as painters, can add this so called icing to the cake, (that 3d cannot yet do): sub-surface scattering is one of these (feature that is supported by very few renderers, but it will change soon): When there's bone beneath your skin the colour and lighting is different than in case when there's only fat and muscle. Candles, rubber and milk (or anything slightly transparent) are things you can draw rather easily, but 3d program might have hard time rendering one of those.

And afterall the biggest part of lighting is the global illumination (others are very subtle compared to it) and humans are much worse "calculating" it than computers.. so if you want to see what are the correct values, model the scene with simple primitives and render it, it's the second best way to get correct/realistic reference, the best is to set the scene and take a photo, or paint in life

How far off I went of the topic ?-)

Ok one more thing, here's a little example I made:

The light is coming from above, (it's spot very far away, just like sun is and it's the reason for blurry shadow on example 2. 3d apps have thing called diagonal light for sunlight, but in reality sun's rays aren't diagonal, every ray is pointing different direction.. ok this is nitpicking but that's the way it is)

The first example has light attenuation and bounce (like in reality), the second one doesn't.

[Bg]

By the way, Balistic, is that scene lit by DI or is it GI? the very dark shadow sides of small rocks suggest spotlight (looks like they don't get enough bounce light), but the soft shadows GI... anyway, brilliant work if that's done with spotlights and omnis!!

[balistic]

My stuff is all done "by hand", with direct illumination. I find that GI sucks too much control away from me.

[edible snowman]

.

[ April 13, 2002: Message edited by: edible snowman ]

[Steven Stahlberg]

Assuming it's your average sun/sky type situation:
number 2 is definitely the right one, except the 'top' part of the cast shadow is much too bright.

[Lunatique]

Here's something I've been thinking about.

Let's say a limb in 50% in the sunlight, and 50% in the shade(tree), and it's around 3pm in the afternoon.

Now, would the darkest side of the part exposed to the sun be the same as the darkest side that's in the shade??





In pic one, the darkest part in the shade(area B), is darker than the darkest area in the light(area A).

In pic two, they are the same(area C and D).

Which is correct?

I personally think pic 2 is correct, but I'm not sure how to explain to myself why.

[ April 12, 2002: Message edited by: Lunatique ]

[BooMSticK]

is it just me or did we somehow got a little sidetracked here with fancy names like global illum, 3d and stuff like that?

I understood Luna's post as a question if theres a genereal rule how to descibe how an objects shadow affect a surface that's already in shadow!
I believe that it's not a simple thing to do. Maybe it's even impossible since many factors have influence. But stripping down to a basic setup will this be possible?
,boom

[jbirn]

"Now, would the darkest side of the part exposed to the sun be the same as the darkest side that's in the shade??"

No. Out in the sunny area, there would probably be more bounced light hitting the bottom of the limb, and more blue light from the sky filling in the sides as well. These extra shades grade themselves out softly as you head into the shade, they don't stop all of a sudden at the shadow line, though.


"a genereal rule how to descibe how an objects shadow affect a surface that's already in shadow!"

General rule: there are always more shadows.

Even inside of other shadows, there are more shadows. Basically, you need to make 3 kinds of shadows:
1. The "main" cast shadows, that would be crisp and blue on a sunny day.
2. The "contact" shadows where two things touch eachother, like where a rock sits on the ground, just shading darker right under or between the surfaces that touch. Those are still there whether you are inside a main shadow or not.
3. The "secondary" cast shadows, which would be VERY soft on a sunny day, because they are just where the light from the sky and clouds is blocked, or reflected light is blocked, not focused like a sun-shadow. Sometimes they are too soft to really see, but they are there even within the main shadows. The areas that are darkened by secondary shadows would start getting green or brown if you couldn't see the sky at all from where that shadow hits.

-jeremy

[Lunatique]

Hey, this is really helpful. Thanks a lot for joining in the discusssion, everyone.

Hopefully my current painting will turn out for the better with what I've learned here.

I have to say that lighting, more than anything else, is the hardest part about being an artist to me. Harder than composition, value, color...etc. Maybe it's because lighting affects everything else.