please can someone have a look at my scene?

[gerardo]

There are some questions that you must to answer, even before asking for help, Rob. You need to decide which will be the mood/atmosphere for your scene, to have clear what you want to obtain before you begin. Once you know this, you will be sure that everything what comes later will fit or not with your vision. This is first than realism. Btw, you don't need accurate realism to achieve an agreeable and overall believable look.
There are several ways to achieve this task, however. For indoors, some people prefer to use conventional solutions (like radiosity and such), what is good if you can afford the render times or you are beginning with lighting process, since that allow you to understand better the light behavior (together with observation of real situations of light); FPrime is a nice tool to get almost instant interactivity in your settings (even with MonteCarlo Radiosity); but at least for me it isn't enough fast yet, and some times you can get quicker results with LW Interpolated Radiosity (check the great Otacon's scenes) or even with other tricks.
I've taken a look at your scene, you don't include the textures so after F9, I get this:

http://www.geocities.com/gerardstrada/1.txt

What surprise me is your lights setup, managed with a lot of criteria, however the results doesn't make it justice. This is due to your Interpolated Radiosity settings and the use than you give it, I think. We use Interpolated Radiosity in these cases, just to reinforce some of radiance and occlusion effects provided mainly by the right use of area and point lights. This means that we still need to fake some type of GI aspects with "bounce lights". Interpolated Radiosity just fills some areas that these tools doesn't, what implies to work with no so high values with this Radiosity solution.
To isolate this phase, I prefer to work without textures (standard grey); look the same scene with the same number of lights (located slightly different), 65% in Intensity Radiosity and a wide evaluation spacing:

http://www.geocities.com/gerardstrada/2.txt

Looks better, I think (and render times is more faster too) but is far of being believable. Why? because we've worked without horizon.
I'll try to take advantage of your lighting setup and figure out a mood that match it. For that I begin from scratch using the first frame of your scene. I'll respond some implicit questions that I'm sure you'll know how to recognize:
First we need to have clear our atmosphere and mood, so let's say we want a summer afternoon (4:30 - 5 pm), a blue sky and a widespread garden outside. We choose a interesting place for our key light (a small area light which will work like a sunlight):

http://www.geocities.com/gerardstrada/3.txt

We change it to a warm color and bluish shadows:

http://www.geocities.com/gerardstrada/4.txt

Then we add a cenital pointlight (shadows OFF) to simulate the interior bounce light:

http://www.geocities.com/gerardstrada/5.txt

We add other area light (shadows OFF) in the opposite window (bluish color):

http://www.geocities.com/gerardstrada/6.txt

And other pointlight to simulate the bounce of sunlight from the floor to that zone of the wall:

http://www.geocities.com/gerardstrada/7.txt

Until here we can add some radiosity solution, but I don't like it because render times and poor control; so I'll use the technique that I shared in this thread:


http://www.cgtalk.com/showthread.php?t=176195


To simulate three important GI aspects:
Occlusion effect
Radiance effect
Scattering effect

So we add SG_AmOcc:

http://www.geocities.com/gerardstrada/8.txt

And FI'sLocalAmbient gradients:

http://www.geocities.com/gerardstrada/9.txt

Instead of FLA gradients we can use gradients in color channel too, but I prefer FLA gradients for walls.

[gerardo]

Note that all surfaces are grey, but indeed we are getting some nice colorations and mood.
The result looks like this without lights and gradients colors:

http://www.geocities.com/gerardstrada/Bw.txt

Then we can make some color correction and level adjustments in post, to match our vision:

http://www.geocities.com/gerardstrada/Color.txt

In this case, render time with this technique is 8 times more faster than Interpolated Radiosity.
I PM you the scene file; you can add some others gradients for more accurate result but I think will be easy for you to go from there. Btw, if you are interested in a deeper approach of this "bounce light" effect with gradients technique, pick up some of the next issues of HDRI3D Magazine :)



Gerardo

E

[mb20]

hi gerado.

thanks so much for all you help here. Really great help and I've learnt alot from it already. If you could send the file to vintageguitars101@hotmail.com that would be fantastic

do you know where to get hold of FI's ambient gradient plugin too please??

thanks alot

rob

[gerardo]

I've just sent you the scene file. In that scene we could reduce even more the render times if we used spinning lights for the keylight and replaced all "bounce lights" by gradients. There are several ways of solving this part. In this case I've preferred to do it this way; however, more important than to obtain a quick render time, is to observe the light behavior minutely to obtain a believable result, for what is advisable to play a lot at the beginning with Radiosity; later, when that part is already more mastered, you can worry about render times :)
Btw, you can get the FLA shader here (is free):
http://f23.aaacafe.ne.jp/~fisjunk/plugin/plugin.php
From Flay.com:
"Allows you to set ambient light settings per object, overriding Lightwave's own ambient light function"



Gerardo

[mb20]

hi gerado. thanks alot for this. I have been reading alot of your other threads. Its going to take some practice but Its great that you are willing to help other people on the forum so much with these great tips and tricks. I hope your planning on writing a book sometime :) sure to be a best seller

best regards

rob

[mb20]

very much in progress but any crits on lighting or texturing is very welcome

cheers for looking

rob

http://www.threedy.com/site/forum/attachment.php?attachmentid=76164&stc=1

[gerardo]

Looks good! but I strongly recommend you try first with some real GI solution. Anyway, if you want to simulate it, perhaps you want to consider these suggestions:

http://www.geocities.com/gerardstrada/lvnroomsuggestion.txt
:)



Gerardo

[mb20]

hi.

A small update. I made the scene lighter and worked on some textures. still having many probs. Any suggestions. crits will be most appreciated

cheers

rob

http://www.threedy.com/site/forum/attachment.php?attachmentid=76922&stc=1
(http://www.threedy.com/site/forum/attachment.php?attachmentid=76886&stc=1)

[gerardo]

Hi Rob,
I get this message from your link:
"Invalid Attachment specified. If you followed a valid link, please notify the webmaster"



Gerardo

[mb20]

hi gerado.

sorry bout the broken link. this one should work

cheers

rob

http://www.threedy.com/site/forum/attachment.php?attachmentid=76922&stc=1

[gerardo]

I encourage you seriously to try first with some real GI solution, since is very difficult to represent something that one don't know how it should be. Simulate GI to speed up the render with tricks is the last step :)
Your last test is more bright, what may looks better, but also lighting is too flat; think that your opposite window is too small to give you that brightness level, due mainly to sunlight direction. Try to recover your radiance and occlusion effect (use Viper to check your gradients and objects reference, maybe you need to diminish your LocalAmbient intensities and increase the ambient occlusion effect). I think your first test is a good starting point to adjust this:

http://www.geocities.com/gerardstrada/lvngroomsuggestion2.txt



Gerardo

[CGTalk Moderation]

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.