|
|
radiosity tests
About Truespace Archives
These pages are a copy of the official truespace forums prior to their removal somewhere around 2011.
They are retained here for archive purposes only.
radiosity tests // Work in Progress
Post by chrisj // Mar 19, 2006, 3:42am
|
chrisj
Total Posts: 239
|
Been a while since I've posted. Don't always get the time but I've been working with radiosity in 6.6, to improve solution times, and try to emulate the latest vray renders. I will move on to vray myself eventually, but only when I can afford a pc capable of running it.
This latest scene was only put together for test purposes, but I'm thinking of adding more details, if I can achieve the quality of realism I'm hoping for.
Chris |
Post by Vizu // Mar 19, 2006, 3:55am
|
Vizu
Total Posts: 628
|
a nice scene that you rendered.
Only the rendertime is to much with radiosity so i never use it.
What you would like do with this scene ? if you do it just for fun, you can upload the scene and the TS7 VRay owner can render it in VRay to see the different between the 2 rendering methode. |
Post by ronrn // Mar 19, 2006, 4:14am
|
ronrn
Total Posts: 10
|
Very nice results. Give us all the juicy details, please. Settings, lights, face count, solution times, etc. I've never been able to achieve results like this in a complex scene with radiosity. |
Post by W!ZARD // Mar 19, 2006, 4:24am
|
W!ZARD
Total Posts: 2603
|
Hi Chris - nice to see you posting again - haven't seen much from you lately, I'd been wondering if you wqere still with us.
I love your image - er... I'll rephrase that: I love this picture that you have rendered :D . I have never ventured into the realms of radiosity but I know it's an interest of yours. This looks great - your interior design flair is showing again!! Good work. |
Post by hemulin // Mar 19, 2006, 9:24am
|
hemulin
Total Posts: 1058
|
That is nice....
I never knew tS6.6 could do something like that natively. |
Post by GraySho // Mar 19, 2006, 9:26am
|
GraySho
Total Posts: 695
|
I've never been a friend of the radiosity solution in ts (lightworks), but you seem to have mastered it :)
Looks very promising. |
Post by Johny // Mar 20, 2006, 6:48am
|
Johny
Total Posts: 672
|
Very nice. How long time need to rendered it? |
Post by chrisj // Mar 20, 2006, 10:09am
|
chrisj
Total Posts: 239
|
Glad to have my efforts appreciated.
Nice to be back in action Wizard. I see you're going from strength to strength with your exterior scenes.
I've been trying to get the best out of radiosity for ages, and have come up with various techniques and adaptions to improve solution times and render quality. Radiosity is prone to all sorts of artefacts, but there are a few hard and fast rules to improve matters. In this scene, I make sure the major surfaces of the room , wall, ceiling and floor, are all one object, and have a high radiosity quality setting. This prevents any intersection artefacts. In fact, booleaning as much as possible together makes a big difference. Bright, smooth, plainly coloured large surfaces are the killer. You don't have to worry about rough, dark, or highly reflective surfaces to the sames extent. If you cant join things together, be wary of objects creating shadowing artefacts onto others, using judicious relative placement. Keeping the light sources to a minimum, especially limiting yourself to a few omnis is the best solution to lighting. A key rule is to strictly limit the range of your light sources. In reality point sources fall of rapidly by an inverse square law, so a standard light bulb shouldn't have a range of more than a metre, this means the solution converges much more rapidly. To create the above image, I used a single infinite light for the direct sunlight, and a single omni for the diffuse illumination from the exterior. The key is the placement of the omni. This single omni is placed above the room, so that it contributes no direct illumination to the interior. The omni illuminates an exterior screen which then directly illuminates the room by radiosity. This effectively creates a large, high quality area light, without the final render overhead (final render time here was 20 mins for an 800MHz PC, although the radiosity solution took some hours). Once again, this omni, which is set to a high brightness, has the fall off controlled to a distance appropriate to the scale of the scene. I soemtimes use a very low intensity (0.1) omni to add a little extra ambient light to the interior, and to help smooth out any irregularities over the larger areas. If you colour any non contributing surfaces balck, with a constant reflection shader it will converge quicker. An finally, enclose your entire scene inside a box. Remember that some of your surfaces may have to have their shadow casting disabled. I've also set the ray limit to a minimum. There you have it, my secret solution.
Hope to complete this one over Easter.
Chris |
Post by chrisj // Mar 20, 2006, 10:17am
|
chrisj
Total Posts: 239
|
Glad to have my efforts appreciated.
Nice to be back in action Wizard. I see you're going from strength to strength with your exterior scenes.
I've been trying to get the best out of radiosity for ages, and have come up with various techniques and adaptions to improve solution times and render quality. Radiosity is prone to all sorts of artefacts, but there are a few hard and fast rules to improve matters. In this scene, I make sure the major surfaces of the room , wall, ceiling and floor, are all one object, and have a high radiosity quality setting. This prevents any intersection artefacts. In fact, booleaning as much as possible together makes a big difference. Bright, smooth, plainly coloured large surfaces are the killer. You don't have to worry about rough, dark, or highly reflective surfaces to the sames extent. If you cant join things together, be wary of objects creating shadowing artefacts onto others, using judicious relative placement. Keeping the light sources to a minimum, especially limiting yourself to a few omnis is the best solution to lighting. A key rule is to strictly limit the range of your light sources. In reality point sources fall of rapidly by an inverse square law, so a standard light bulb shouldn't have a range of more than a metre, this means the solution converges much more rapidly. To create the above image, I used a single infinite light for the direct sunlight, and a single omni for the diffuse illumination from the exterior. The key is the placement of the omni. This single omni is placed above the room, so that it contributes no direct illumination to the interior. The omni illuminates an exterior screen which then directly illuminates the room by radiosity. This effectively creates a large, high quality area light, without the final render overhead (final render time here was 20 mins for an 800MHz PC, although the radiosity solution took some hours). Once again, this omni, which is set to a high brightness, has the fall off controlled to a distance appropriate to the scale of the scene. I soemtimes use a very low intensity (0.1) omni to add a little extra ambient light to the interior, and to help smooth out any irregularities over the larger areas. If you colour any non contributing surfaces balck, with a constant reflection shader it will converge quicker. An finally, enclose your entire scene inside a box. Remember that some of your surfaces may have to have their shadow casting disabled. I've also set the ray limit to a minimum. There you have it, my secret solution.
Hope to complete this one over Easter.
Chris |
Post by chrisj // Mar 21, 2006, 1:59pm
|
chrisj
Total Posts: 239
|
Night version. Just for fun. Apologies for the background. |
Post by Mike // Mar 21, 2006, 3:00pm
|
Mike
Total Posts: 121
|
Really really nice! :banana: |
Post by behzad // Mar 21, 2006, 3:59pm
|
behzad
Total Posts: 173
|
As good as v-ray:banana: |
Post by W!ZARD // Mar 21, 2006, 6:33pm
|
W!ZARD
Total Posts: 2603
|
Glad to have my efforts appreciated.
Nice to be back in action Wizard. I see you're going from strength to strength with your exterior scenes.
Chris
Cheers Chris - and you are going from strength to strength with your interior scenes!:D You are rapidly becoming our radiosity Guru. The insights you've gained from your experiments are invaluable and will no doubt save a lot of time for others exploring Radiosity is tS.
It's good of you to share your techniques. |
Post by GraySho // Mar 21, 2006, 9:50pm
|
GraySho
Total Posts: 695
|
I like this even more than the daylight scene. The Monitor should emit a little bit more light (material light lowest intensity?), the rest looks stunning. |
Post by stoker // Mar 21, 2006, 9:58pm
|
stoker
Total Posts: 506
|
Real Nice!!;) :D :D |
Post by djaram // Mar 21, 2006, 10:40pm
|
djaram
Total Posts: 73
|
hm... I wonder how long it took to render. But it is really impressive.
Great Job! |
Post by Vizu // Mar 21, 2006, 11:35pm
|
Vizu
Total Posts: 628
|
... not for render, more how log it needs for calculate this scene.
should be helpful when we can download the scene to test a bit with it. |
Post by KeithC // Mar 22, 2006, 5:42am
|
KeithC
Total Posts: 467
|
Very nice, I like the night image as well.
-Keith |
Post by hemulin // Mar 22, 2006, 9:18am
|
hemulin
Total Posts: 1058
|
Should there be some reflections from the TFT on the table. Oh, great image btw! |
Post by Alien // Mar 22, 2006, 1:36pm
|
Alien
Total Posts: 1231
|
Should there be some reflections from the TFT on the table.
There are, though admittedly I only noticed it after I re-adjusted the gamma correction in my display settings [somehow it had managed to reset to default since the last time I did it].
Oh, great image btw!
Agreed. :cool: |
Post by chrisj // Mar 23, 2006, 11:32am
|
chrisj
Total Posts: 239
|
Once again, thanks for all the positive feedback.
I hope the techniques I've mentioned can be of use to those of you pushing the limits without vray. If anyone can make furhter suggestion, I'd be very interested to learn more.
As suggested by vizu, I've posted the scene on the download page of my website.(see profile) Any problems, please let me know asap.
Looking forward to seeing the results from a few of you.
cheers
Chris |
Post by hemulin // Mar 24, 2006, 10:11am
|
hemulin
Total Posts: 1058
|
Should there be some reflections from the TFT on the table.
There are, though admittedly I only noticed it after I re-adjusted the gamma correction in my display settings [somehow it had managed to reset to default since the last time I did it].
Ah yes I see them now (the second image). However I think the image on the TFT should be reflected more. |
Post by 3dpdk // Apr 1, 2006, 6:09pm
|
3dpdk
Total Posts: 212
|
WOW Chris!
As usual, your images are incredibly photorealistic. I keep a copy of the kitchen you did a while back just to study from time to time.
I assume you are apologizing for the view outside the window and also assuming (after your lighting explanation) that you must use a flat image to cut the calculation time. It does look a bit like a projected image ... hay ... maybe it's an office buried deep inside some high rise and has no outside view and it IS a projected image ... maybe? |
Post by chrisj // Apr 2, 2006, 4:07am
|
chrisj
Total Posts: 239
|
welcome back Paul.
glad you like it. not getting the time for so much these days, and my imagination is a tad dry at present. cant wait for my long summer holiday to recharge the enthusiasm.
the background is a simple mapped plane. The image is of the Melbourne skyline, found on google. It's slightly distorted because of a radiosity artefact, and a fault with the render engine I've discovered.
Chris |
|
