Attached Link: http://www.wolfiesden.com/bryce/tut/old/adv-lighting3.html

Since all the talk lately is of lighting in scenes and such I thought I'd do some testing. Wolfie's Den at the link above did a tutorial on lighting in the real world. I decided to replicate these and do some comparisons to real lighting. The first picture is a single radial light, all light calculations are done from a single pinpoint. As far as I know there are no natural or manmade lights sources that only emit light from essentially a point the size of an atom. Notice the hard crisp shadow. This took 33 seconds on a 1.67 Ghz AMD (works like a 2.0 Ghz).

Now real light bulbs have filaments and light is emitted all along the length of the filament. Some are just straight lines (actually curly but form a straight line) and some form U shapes. This is a straight filament using 4 radials very close to each other. Not much different than the single point except for a very slight softening of the shadow edge. I was expecting a bigger difference but looking at a similar light bulb(clear) in my house this is true. So for these types of lights a single radial will probably work just fine. Render time 1:56

Next we move to the frosted light bulb. These bulbs emit light from the full surface of the glass and diffusing it a bit giving a softer light.

Simulated this by using a sphere of 66 radials the size of the glass. Then the glass ambience is set to 100%. Transparency can control how much light is let through, I left it at 100% transparent.

Notice the softer shadows.

Render time 22:22

Message edited on: 09/07/2004 19:37

Interesting experiment. How about doing the first single radial again, but set it to have softshadows? So we can compare render time and effect?

---------
Phillip Drawbridge
Website 
Facebook

Now for a flourescent light tube. This type of light also emits light from the surface of the glass and gives a really diffuse light. Notice the shadow is almost non-existant at the edges and almost a thin dark line where all the shadows intersect. Ths light was created with 168 radials, 8 radials around the tube duplicated 21 times along it's length. Render time 25:32

Ask and ye shall receive. I had to rebuild the scene so it may not be exactly the same as above as far as position and such.

1 radial set to 50% softness

Render time 7:35 Given the time difference perhaps a single radial set to soft shadows would be better and achieve the same thing.

Message edited on: 09/07/2004 20:11

Did another render this time with the shadow setting to 25 to get closer to the shadow softness of the frosted light. Render time increased to 10:35 So the softer the shadow the faster the render it seems.

The soft shadow setting appears to look much better than the 66 radial setup. just my two cents. And also, the Sunlight on a clear day tends to radiate from a "single" point and thus gives very sharp, defined shadows.

Attached Link: http://freespace.virgin.net/hugo.elias/radiosity/radiosity.htm

Ok back to what I was getting at. Using the link above I decided some time ago to try and repeat the scene in bryce. I only had marginal success with a single light source and actually had to add 2 light at the windows to get more light to "bounce" inside, even using TA. The above picture is a revamp of that scene adding lights as needed to get realistic lighting. Still had trouble with the ceiling without adding lights below the floor not casting any shadows (not included in these renders). Scene consists of 1 sun, 2 radial pinkish lights at the hotspots on the floor, half sphere of light blue radials (about 60 radials or so I think) to simulate the diffuse light from the sky and finally 3 flourescent lights as built in the above tests. This is only the second or so pass, total render time with TA was going to be over 6 days.

True Slakker I forgot to mention that. The sun even though the light is actually emitted from it's surface (some millions of square miles) it is so far away the light might as well be coming from a sigle point and that light is paralell.

This one is without TA and looking into the room from the window as I wanted to see if the blue diffuse light was getting to the back wall. IMO it's actually a bit too much but lower settings didn't get the blueness near the windows where it should have been brightest blue. This is sevral passes but still was going to be almost 8 hours to render. I do like the play of light in this one but again the ceiling is not receiving enough light.

In conclusion with the current Bryce to get realistic play of light will take a lot of lights and an enourmous amount of render time. I would love to see besides the faster render time that AS says DAZ has planned. The ability of Bryce to do actual radiosity and actually bounce light several times off of objects. I would also like to see a material setting to set the amount of light that material reflects back into the environment. Black surfaces would reflect very little to no energy back into the environment while white ones would reflect maybe 95% back (some is always absorbed by the material).

Interesting thread, Aldaron. I think you'll find that the render time estimate given by Bryce for your TA scene above is a big underestimate. I've noticed with the advanced render options in Bryce5, particularly True Ambience, that the final pass takes an absolute age compared to prediction in complex scenes. I'm currently rendering a scene that was predicted to be about 12 hours, but will actually be more like 36 hours. In your scene you have about 450 shadow casting lights. This is really asking for trouble with True Ambience, and I suspect the render would actually take several weeks to complete. But the main benefit of using True Ambience is to reduce the number of lights needed to fake light bouncing (light arrays and such like). Using the two together misses the point of both, and combines the worst of both worlds. I believe that True Ambience is designed to achieve a similar job to what radiosity engines are designed to achieve, although the specifics of the process are different (and not as effective). I know my views are not popular amongst the forum 'heavyweights', but I maintain my position with good reason. I'll post more on this in a couple of days. As regards materials behaving in the manner that you mention in your last post, Bryce already has them. Under True Ambience (suitably set up) darker materials 'reflect' less light than lighter ones. Black will 'reflect' nothing; white will 'reflect' the most of all.

True PJF but it doesn't "reflect" it back into the scene far enough. The falloff is way too short. Another point is yes I could probably reduce the number of lights needed but not sure I would get the same effect and probably more banding (which you can't see in these renders). Don't get me wrong, I like TA a lot it just doesn't do enough of what I want. I want light to bounce several times and not just in certain circumstances like you thread that you posted some time ago (which I have yet to reproduce). I'll be putting up a render soon of the tube light with a lot fewer lights and soft shadows and compare it to the above.

Ok here is a tube light using only 3 radials set to 50% soft shadows. If I remember correctly Bryce 4 doesn't have soft shadow settings? If so then obviously those with Bryce 5 will want to use lights with soft shadow settings. Render time 9:42

Dang forgot the image :)

Aldaron - actually, the sun is NOT a pin-point source. Try this experiment: Go outside on a bright, sunny day, when the sun is at an appreciable angle from the vertical (ie, not local noon). Stand somewhere where you can see your whole shadow. (If that happens to be the middle of a road with lots of fast traffic, I suggest moving the experiment elsewhere..) Now look at your shadow. Near the feet, the shadow will be crisp. As the shadow "moves" up the body it will get less distinct. The further away the surface that recieves the shadow, the less distinct... This is because your body is being hit by light rays from across the whole surface of the sun. Since ALL those rays are straight (not parallel - just straight), where they overlap after passing beyond the your body's edge, the amount of shadow is reduced. Plus, of course, the atmosphere spreads light around too. Which latter is something that many sci-fi artists forget: they are told that "in space (or on the moon) shadows are crisp, because there's no atmosphere". Not so. Although there's no light scattering from atmosphere, if the light source is close enough to be discerned as a disk, there will be shadow softening. The contrast will be much greater though. Errm... I'm rambling.. I'll shut up now... Cheers, Diolma

Aye, I'm interested in seeing the finished render in your light-test room... But I'll have to agree with PJF, this scene doesn't really call for True Ambience... Still, I think it was a lighting experiment, right? And you're trying to figure out how far light will bounce in Bryce and whatnot...? I think it's a great scene setup to test such a thing, but alas, Bryce doesn't bounce light. At one point, PJF (who I consider a "heavyweight", regardless!), myself, and a few others experimented heavily in True Ambience and the like... We were able to make light bounce with TA, but since there are no TA LIGHTS, per se, it is nearly useless as a tool... I DID figure out how to bounce a spotlight off a mirror, but alas, it wasn't something you could do in Bryce. It is easily faked, however. Hopefully they'll give us some cooler options in Bryce 6, give us some new toys and techniques to play with!

Wow, a real blast from the past. Proves that ebots will haunt you forever. "We were able to make light bounce with TA, but since there are no TA LIGHTS, per se, it is nearly useless as a tool..." While TA can't help in bouncing hard light off of mirrors, it is far from useless. It "bounces" diffuse light quite well, and in this context all Bryce lights are "TA lights" since they can all provide the initial "energy". True Ambience was only partially developed when Bryce5 was released, and there are quite a few problems with how various objects and surfaces react to it (non-smoothed polygon meshes being notorious). If DAZ can fix those (or some) for Bryce5.5, then True Ambience may see more action. "who I consider a "heavyweight", regardless" If you're calling me a fat git - you're right.

Attached Link: http://freespace.virgin.net/hugo.elias/radiosity/radiosity.htm

BTW if you wondered where I got the idea for the room.....^

Attached Link: http://www.renderosity.com/messages.ez?Form.ShowMessage=1445452

LSD, check this thread then tell me Bryce can't bounce light :)

"experimented heavily in True Ambience and the like... We were able to make light bounce with TA, but since..."

Now let's see ya render an animation with this effect? Not gonna happen... Good to see you still alive, PJF! What have you been up to, anyway?

Message edited on: 01/29/2005 13:02