PDA

View Full Version : LW 9.5 Lights



Raccoon Bandit
02-23-2008, 02:02 PM
Since the other thread is so into fiber/fur I thought I'd open this one to just concentrate on some of the cool aspects of LW9.5's new lighting system. And I know this first set was already posted there but since this will be (should be) devoted to lights I'll repost and hopefully Secret Squirrel will hop in to with his awesome images which I bow down to.


Heres a quick look at why we like the new Photometric (IES) lights. While not a solution for every problem they really have their uses. The attached image(s) are all using the exact same settings except for light types and their specific settings although the intensity remained the same for them all. Please don't bite my head off on the AA it is only set at 3 with AS on also so I could render quickly (all renders are well under an hour!) with Monte Carlos GI.

In addition to the two hanging lamps theres an HDRI used to help with the dielectric vase, marbles and logo, plus an area light (set to a quality of 3) mainly to show the interaction of the hanging lights with shadows. The GI as I said was Monte Carlos, Interpolated at default settings except for Bounces are set to 2 and Rays to 150.

Raccoon Bandit
02-23-2008, 02:03 PM
This next series shows some more spherical light stuff. For this set I decided to use the MattMD.lwo file that Nicholas (Lor) Boughen uses in his 'Light 3D 8/9 Lighting' books since so many people are familiar with it. The only thing I have done, other then turning off the celshader, is to create a single light with no GI.

- The top image shows an Area light setup with (255,255,255) color, 100% intensity quality = 4, size = 50x50x50 and no ambient.

- The second image shows a Spherical light setup with (255,255,255) color, 100% intensity quality = 4, size = 50m and no ambient.

- The Bottom image is done in PS CS3 and just shows setting the Spherical images blend mode to different on top of the Area image.

Raccoon Bandit
02-23-2008, 02:03 PM
This next set is the fun one. If you take a Spherical light and invert its size, in this case its -200m you have an instant dome light. Better yet turn on Monte Carlo GI and you have a daylight rig setup in less then 20 seconds. No more need for the spinning light trick!

geothefaust
02-23-2008, 02:08 PM
:O

YOU RULE!

Wow. That is neat. I am happy to hear that it's so easy to invert it to create a dome light. Great stuff!!

JBT27
02-23-2008, 02:13 PM
With the first set, how big are the images? ..... is what we see the full render size? And what specifically do you mean by 'well under an hour'?

I am new to using photometric lighting, so maybe I am missing something.....but presumably from what you say the render times are good for photometric lights on a video frame sized(ish) image? But to me it sounds like rendering these is pretty slow and an animation with them will be tedious to put it mildly.

I have to admit that the lighting is of more interest to me than the fur and hair, so this thread is of great interest.

Thanks for posting these! It is very exciting, but I am concerned to understand your comment on render times for those photometrics; it sounds like you are well-impressed with the times, but being new to me, it's starting to sound like a relatively slow render process. Is that right?

Julian.

Raccoon Bandit
02-23-2008, 02:46 PM
JBT27, I can't tell you exactly how long the renders are because I was working on debugging a program I'm writing which when in operation was competing extremely heavily for thread and cycle time. But none of them took more then an hour at 640x480 to render.

Raccoon Bandit
02-23-2008, 02:46 PM
Here's another request for all you masked 9.5 avengers: how about an area light with inverse-squared falloff standing on edge perpendicular to (and just resting upon) a flat ground plane? No antialiasing, please.

So here you go, a single area light sized increased to 2x2x2,set perpendiclarly to a ground plane and just touching. Intensity = 100%, 0% ambient, no AA, no GI. Fall off is set to Inverse Distance^2, Range = 5m (your Range MUST be greater then your size else you will get noise welcome to the world of math!)

Raccoon Bandit
02-23-2008, 02:47 PM
And for the fun of it the second set use Inverse Distance fall off all other settings the same.

Raccoon Bandit
02-23-2008, 02:55 PM
I'll start a larger better render of my Photometric vase scene tonight, again I'll, hopefully, fingers crossed, knock on wood, have me program running all night so I won;t know what the real render time will be but at least you'll have something better/larger to view.

Even better then that theres some GREAT work coming real soon as in they are gearing up to render right now from other super secret critters real soon!!!!

Weepul
02-23-2008, 03:40 PM
So here you go, a single area light sized increased to 2x2x2,set perpendiclarly to a ground plane and just touching. Intensity = 100%, 0% ambient, no AA, no GI. Fall off is set to Inverse Distance^2, Range = 5m (your Range MUST be greater then your size else you will get noise welcome to the world of math!)
Thanks for the test. :) That bit about range and noise isn't really correct, since changing range is exactly the same as changing intensity, for inverse squared at least. Eg. intensity 100% range 1m is the same as intensity 4% range 5m.

I don't suppose you'd be willing to re-test with the lower intensity? Also, what quality/number of samples did you use here? If it was higher than 25 (the max. in 9.3.1, equivalent to quality setting 5) could use use 25 rays/quality 5, so as to compare efficiency? [edit] Render times and system specs would be appreciated too. :D

IMI
02-23-2008, 03:44 PM
Thank you for this new thread, RB. I'm getting burned out on reading about hair and fur. Lighting is far more interesting to me.
Looking forward to your photometric scene!
Suppose you could talk one of those other super secret critters into starting a UV thread too? ;)

JBT27
02-23-2008, 03:50 PM
JBT27, I can't tell you exactly how long the renders are because I was working on debugging a program I'm writing which when in operation was competing extremely heavily for thread and cycle time. But none of them took more then an hour at 640x480 to render.

OK, thanks! This is looking nice, but when and if you have time....more tests please! :)

Thanks!!

Julian.

SecretSquirrel
02-23-2008, 04:02 PM
OK, thanks! This is looking nice, but when and if you have time....more tests please! :)

Thanks!!

Julian.

Any tests in particular you'd like to see?

*Pete*
02-23-2008, 04:04 PM
spherical light seems to have the potential to be the main light to use, replacing the area light in most situations.

thanks for the test and the thread....perhaps have hair/fur threads, uv threads, light threads and GI threads....there is a lot about the new GI that is still a mystery.

IMI
02-23-2008, 04:08 PM
spherical light seems to have the potential to be the main light to use, replacing the area light in most situations.

thanks for the test and the thread....perhaps have hair/fur threads, uv threads, light threads and GI threads....there is a lot about the new GI that is still a mystery.

Agreed completely.
Spherical and GI... well, actually, all the lights.
I'd like to see several versions of a scene lit with all the new lights, individually, along with render times and machine specs.
If it's not to much trouble, that is. :)

tischbein3
02-23-2008, 04:12 PM
@Secret Squirrel
What about IES and caustics ?

Raccoon Bandit
02-23-2008, 05:37 PM
That bit about range and noise isn't really correct, since changing range is exactly the same as changing intensity, for inverse squared at least. Eg. intensity 100% range 1m is the same as intensity 4% range 5m.
Via observation your 98% right. Mathematically I'm right. In the end we only care about observation though so here you go.

Same scene as above only difference is I did bump Light Quality up to 20 and drop Intensity down to 4%.

SecretSquirrel
02-23-2008, 05:47 PM
800x600
4 Bounce Interpolated Final Gather
1 Pass AA, 0.05 Adaptive Sampling
Sigma2, Dielectric and other Node Shaders
ImageWorld HDRI lighting only

Test Machine: Quad Core 2 Penitum 2.66Ghz, 4GB RAM (iTunes playing in background)

GI Time: 1m 10s
Render Time: 13m 20s

Weepul
02-23-2008, 06:18 PM
Via observation your 98% right. Mathematically I'm right. In the end we only care about observation though so here you go.

Same scene as above only difference is I did bump Light Quality up to 20 and drop Intensity down to 4%.
I'm pretty sure I'm right about the math :hey: at least in 9.3.1. But that's not the issue at hand...

Thanks for the test - it's impressively smooth, but looks strangely blurry...well, I guess I'll just have to wait until I get my hands on it to really see what's going on. :beta:

caesar
02-23-2008, 06:24 PM
800x600
4 Bounce Interpolated Final Gather
1 Pass AA, 0.05 Adaptive Sampling
Sigma2, Dielectric and other Node Shaders
ImageWorld HDRI lighting only

Test Machine: Quad Core 2 Penitum 2.66Ghz, 4GB RAM (iTunes playing in background)

GI Time: 1m 10s
Render Time: 13m 20s

Outstanding !!!

tyrot
02-23-2008, 06:24 PM
800x600
4 Bounce Interpolated Final Gather
1 Pass AA, 0.05 Adaptive Sampling
Sigma2, Dielectric and other Node Shaders
ImageWorld HDRI lighting only

Test Machine: Quad Core 2 Penitum 2.66Ghz, 4GB RAM (iTunes playing in background)

GI Time: 1m 10s
Render Time: 13m 20s


DEAR secret

can you remove those Node Shaders and give a basic Render with same material?

Although pass is So less "1"!! ..AS is really taking care all the jagged edges..I think there is a big improvement overthere...

BEST

IMI
02-23-2008, 06:29 PM
Looks great, Oh Secret One.
I'm gonna have to get me one of those quad cores. My dual core AMD 3.2 ghz is great, but I suspect 9.5 might enjoy a quad core Intel more...

SecretSquirrel
02-23-2008, 09:41 PM
Here's a comparison between 9.3.1 and 9.5. Both have 128 'Rays Per Evaluation' and 'Min Pixel Spacing' set to 3.0.

There is much more control over GI in 9.5, so this is as close as I can get with the settings.

First image is LightWave 9.3.1, render time 45 seconds. Second image is LightWave 9.5, render time 41 seconds!

What you need to note here is how much better the 9.5 image is, look at the smoothness of the floor compared to 9.3.1. But then look at the details on the head model, lots of detail is lost on the 9.3.1 image because the pixel spacing is too high to catch them, but in 9.5 you have control over secondary bounces which help bring out this detail at low 'Rays Per Evaluation', yet render time (in this case) is actually lower, but the quality is MUCH better!

I think example says a lot actually.

P.S. I should point out that there are 1,633,722 polys in this scene, those head models are very dense meshes.

wavemaster
02-23-2008, 09:48 PM
Thats great to hear about the secondary bounces. Can you share any of the other GI options that have been added?

Auger
02-23-2008, 09:49 PM
Nice comparison! :thumbsup:

SecretSquirrel
02-23-2008, 09:55 PM
Thats great to hear about the secondary bounces. Can you share any of the other GI options that have been added?

Hmmmm, probably not at the moment, Jay will probably kill me for saying that much!

:)

wavemaster
02-23-2008, 09:57 PM
I am just fishing for info!:D Thanks for the tests.

Exception
02-23-2008, 11:33 PM
Thats great to hear about the secondary bounces. Can you share any of the other GI options that have been added?


Disk Caching :)

Cageman
02-23-2008, 11:34 PM
Looking really good so far! Keep these lighttests comming! :)

SecretSquirrel
02-23-2008, 11:40 PM
Same scene and settings as the last render, but with a Spherical Light in the middle of the objects.

The Spherical Light sphere is just an object to show the light location and size, it's set to not affect anything in the render but just show up.

Render Time: 2m

harlan
02-24-2008, 12:01 AM
Hmmm.... wouldn't it be cool to have a 'display in render' toggle for things like lights that would enable a representation of them in a render (like your spherical light "object").

SecretSquirrel
02-24-2008, 12:07 AM
I thought so too, a few of us have asked for this feature, but it won't be in 9.5 as far as I know.

SecretSquirrel
02-24-2008, 12:12 AM
Same scene and settings, but lit with a photometric light. I think I'm correct in saying that the shadowing (with regard to shadow softness) depends very much on the IES file you're using. I'll try some different ones and see.

Render Time: 1m 2s

SecretSquirrel
02-24-2008, 12:33 AM
Apparently it does! Here's another one with a different IES file, which isn't the best one for showing off the lovely streaks, but as you can see soft shadowing depends on the IES type.

Render Time: 1m 49s

loki74
02-24-2008, 12:54 AM
Apparently it does! Here's another one with a different IES file, which isn't the best one for showing off the lovely streaks, but as you can see soft shadowing depends on the IES type.

Render Time: 1m 49s

Which really makes sense, if you think about it, since the shadow hardness in real life also depends somewhat on they type of bulb your using...

Very nice tests, anyway--looking forward to this!

Panikos
02-24-2008, 01:01 AM
What about Caustics ?
Nobody mentioned them.
Still ugly & unstable ?

mikadit
02-24-2008, 01:18 AM
Absolutely awesome!

What about the possibility do optain false colour render? I mean, it's possible to edit the IES light output to optain differents colours for the intensity and desity of the lights?

SecretSquirrel
02-24-2008, 01:35 AM
Absolutely awesome!

What about the possibility do optain false colour render? I mean, it's possible to edit the IES light output to optain differents colours for the intensity and desity of the lights?

Not possible in this release, I agree, would be cool though!

Here's the scene again lit with area lights. The decrease in speed is down to the better quality GI in 9.5, area lights are the same. But you should be able to see overall less noise in the 9.5 scene.

First LW 9.3.1: 1m 7s

Second LW 9.5: 1m 39s (but less noise than 9.3.1)

mikadit
02-24-2008, 02:00 AM
In anyway I'm really happy for all those new 9.5' features. I like how the edges are much more defined with the new lighting/rendering system.

mav3rick
02-24-2008, 02:00 AM
now lw has better area light than fprime / g2 :)

flashover
02-24-2008, 02:09 AM
Great tests!!!

arsad
02-24-2008, 03:27 AM
nice images.

Cageman
02-24-2008, 03:41 AM
Looks like LWs renderer is getting closer and closer to the same definition in detail MR has without needing same level of AS as in 9.3. I'm really happy about this! Good work NT! And thanks alot to Secret Squirrel and Raccoon Bandit for these very nice images.

Keep 'em comming! :)

tyrot
02-24-2008, 04:02 AM
dear secret

thanks for the posts. Great renders. Pics leave no doubt, 9.5 GI simply rocks.


BEST

-EsHrA-
02-24-2008, 04:34 AM
cheers for the tests testers.
GI seems better than previous lw versions but for it to rock we need comparisons with others engines.


mlon

IMI
02-24-2008, 05:55 AM
Thanks again, SecretSquirrel.
The IES soft shadows look like shadow maps, which isn't a bad thing, since it's pretty good. But are they full raytracing? Volumetric options? Can they do gobos with images or procedural textures?
(Forgive me if this has been addressed before, but I really haven't had time to properly wrap my head around just what IES light are, along with any limitations.)

I too would like to know if caustics has been improved any in LW 9.5, with the IES, spherical, or in any way at all.

Lightwolf
02-24-2008, 06:15 AM
Forgive me if this has been addressed before, but I really haven't had time to properly wrap my head around just what IES light are, along with any limitations.
Basically a set ot data that describes the physical properties of a manufacturers real life light.
The idea is to be able to use those when doing arch-viz to exactly see the effect of a manufacturers light. They're mainly designed for light planning in architecture, but as you can see can be used for rendering as well.
You basically wouldn't want to change an of the properties, since the intent is to give you real world properties. MAybe LW will allow you to override some of them... but I suppose you might as well use normal lights with gobos then.

Cheers,
Mike

IMI
02-24-2008, 06:22 AM
Ah, thank you Lightwolf. I had read the part about them (IES lights) being some sort of "map" (sort of like a lighting equivalent of digital elevation models?) to simulate actual manufacturer lights, but hadn't considered "the intent is to give you real world properties".

*Pete*
02-24-2008, 06:41 AM
the IES is to lights as the real lens camera is for the camera.

in both cases can you select the real world light/lens and get virtually the same results within LW.

IMI
02-24-2008, 07:16 AM
Thanks Pete. Isn't there an existing IES light plugin for LW 9.x somewhere?
For 64-bit?

Mitja
02-24-2008, 07:24 AM
Yes, there's DP node IES map, but's rather a workaround than a IES light plugin, actually!

dballesg
02-24-2008, 08:00 AM
Hi,

I love the renders, I compare them open them on Firefox tabs and switching between them, and 9.5 excels way long! :)

But I am curious about caustics as well! :)

So a test with caustics?

David

Exception
02-24-2008, 10:54 AM
Not possible in this release, I agree, would be cool though!

SS, don't forget the Light API... surely someone can write a plugin that can do that now... another huge benefit of the new lighting system. The IES and Sphere lights are just a few of the myriad of possibilities.

Ztreem
02-24-2008, 12:22 PM
SS, don't forget the Light API... surely someone can write a plugin that can do that now... another huge benefit of the new lighting system. The IES and Sphere lights are just a few of the myriad of possibilities.

So hopefully we will be able to get geometry based lights without using radiosity.

wavemaster
02-24-2008, 12:25 PM
I was thinking with the new GI options we will be able to get geometry to act more like lights, but I dont know what the options are!

wavemaster
02-24-2008, 12:25 PM
I was thinking with the new GI options we will be able to get geometry to act more like lights, but I dont know what the options are!

wavemaster
02-24-2008, 12:27 PM
can someone say if the gi options are global options or per surface options?

Raccoon Bandit
02-24-2008, 12:27 PM
Its true with the new Light additions to the SDK almost anything is now possible and I don't doubt that after launch you'll start seeing as many light plugins as you do node plugins coming out!

[EDIT] Oh and sorry my Photometric Vase render never finished last night do to my program I was running concurrently hung my system (it did not like sharing cycle time with LW rendering and clogged the gpu <-> cpu traffiic)

Raccoon Bandit
02-24-2008, 12:34 PM
can someone say if the gi options are global options or per surface options?
Theres a lot we could say and heck show, some of which I'd kill to show right now, but alas a lot of goodies are still off limits for us to talk about/show. And don't take that wrong, its actually a good thing not a bad thing. Think about it like this you have all your Christmas presents sitting out in front of the tree at the start of December. Everyones gone so you open everything to see what you got and then wrapped them back up. The rest of December and especially Christmas day would suck.

wavemaster
02-24-2008, 12:49 PM
I hate xmas:thumbsdow
just kidding, I hate not having beta though!

Exception
02-24-2008, 01:00 PM
So hopefully we will be able to get geometry based lights without using radiosity.

That's what the spherical light is, so it should be possible for more things.

dballesg
02-24-2008, 01:09 PM
Theres a lot we could say and heck show, some of which I'd kill to show right now, but alas a lot of goodies are still off limits for us to talk about/show. And don't take that wrong, its actually a good thing not a bad thing. Think about it like this you have all your Christmas presents sitting out in front of the tree at the start of December. Everyones gone so you open everything to see what you got and then wrapped them back up. The rest of December and especially Christmas day would suck.

Well and what about of us that ALWAYS open the gifts wayyyy before???? :)

*Pete*
02-24-2008, 01:27 PM
obviously...some got to open the gifts way before the rest of us...

erhm...besides, christmas is over already, about time to open the gifts and release what ever is hidden inside!!

Hidden Halibut
02-24-2008, 01:28 PM
obviously...some got to open the gifts way before the rest of us...


No, we're the guys that wrap them!
No gifts for us :(

*Pete*
02-24-2008, 01:37 PM
No, we're the guys that wrap them!
No gifts for us :(

I always felt sorry for the people who work as giftwrappers in the bigger shops at christmas time...it must be very boring job.

but you guys have a job that is comparable to the test drivers at the ferrari F1 team...far from being boring ;)

wavemaster
02-24-2008, 01:49 PM
It would make the pain much less if yous would post some frigin images!!!!!!!!!!!:D

Hidden Halibut
02-24-2008, 01:57 PM
It would make the pain much less if yous would post some frigin images!!!!!!!!!!!:D

I'd love to but I have work to do testing some areas... and oh, yeah before I forget, perhaps I can squeeze in my job somewhere this week in between all the testing :)

wavemaster
02-24-2008, 02:01 PM
who can work at a time like this?:beta: :ohmy: :eek: :cursin: :bangwall: 8~

cg_mike
02-24-2008, 02:03 PM
I like that these tests are using all the same statue/head model. I recall seeing it in some of the nodal shader tests awhile back, but I can't seem to locate it in my LW content directory. Is it a content LWO or is it from someplace else?

Thanks!

IMI
02-24-2008, 02:16 PM
I'd love to but I have work to do testing some areas... and oh, yeah before I forget, perhaps I can squeeze in my job somewhere this week in between all the testing :)

Aha! I guess that eliminates you as being a Newtek employee.
Trying to figure out who's who....
There was a comment from you earlier regarding being "nice to fish"... tell me, does the term "fish hater" mean anything to you? ;)

Umm, well, in any event, carry on. :)

hrgiger
02-24-2008, 02:23 PM
Has it been asked yet if the new lights (IES and spherical) work with Fprime? More importantly, has it been answered? Also, do the area light improvements with less noise also translate to cleaner Fprime renders?

Hidden Halibut
02-24-2008, 02:26 PM
Aha! I guess that eliminates you as being a Newtek employee.


I don't think any of us work for Newtek.
And nobody gets paid, as far as I know.

We're doing this for your enjoyment...

(nice politics... ha)

zardrose
02-24-2008, 02:49 PM
Modo photon splits for is big photon, small photon

LW 9.5 It has?:bowdown:

http://bbs.chinavfx.net/attachments/200802/25055334-snap0120960.jpg

IMI
02-24-2008, 02:49 PM
We're doing this for your enjoyment...

(nice politics... ha)

In all seriousness, I'm enjoying it immensely, and am highly appreciative of all of y'all who are involved with this and making these efforts. :)

mav3rick
02-24-2008, 02:53 PM
Its true with the new Light additions to the SDK almost anything is now possible and I don't doubt that after launch you'll start seeing as many light plugins as you do node plugins coming out!

[EDIT] Oh and sorry my Photometric Vase render never finished last night do to my program I was running concurrently hung my system (it did not like sharing cycle time with LW rendering and clogged the gpu <-> cpu traffiic)


you sound like poet.
i hatte books

actually this comment was for post b4 where u talked about some xmass and stuff like that

Raccoon Bandit
02-24-2008, 03:27 PM
No, we're the guys that wrap them!
No gifts for us

You forgot to mention the billions of paper cuts we get wrapping. And a lot of time testing is not fun either, as it means dealing with code that is literally taped together between old junk and new shinny goodness. Plus we sometimes find that some presents get unwrapped, played with and then we re-wrap them. We do lots and lots of re-wrapping. Tank goodness paper is recyclable!

RedBull
02-24-2008, 03:30 PM
Has it been asked yet if the new lights (IES and spherical) work with Fprime? More importantly, has it been answered?

Not until Fprime 3.5 has been updated to work with the new features.

SecretSquirrel
02-24-2008, 04:18 PM
Has it been asked yet if the new lights (IES and spherical) work with Fprime? More importantly, has it been answered? Also, do the area light improvements with less noise also translate to cleaner Fprime renders?

Any light improvements / changes need FPrime to be updated to take advantage of them.

Remember FPrime mimics LW's render engine, it doesn't use it and magically speed it up.

Lightwolf
02-24-2008, 04:41 PM
Has it been asked yet if the new lights (IES and spherical) work with Fprime?
Once FPrime has been updated to use the latest SDK which hopefully allows for the evaluation of the new lights by third parties... why not?
And I assume it has been designed with that in mind.

Also, do the area light improvements with less noise also translate to cleaner Fprime renders?
I doubt that. Afaik FPrime uses its own light evaluation - which is why spots can't use shadow maps in FPrime for example.

Cheers,
Mike

IMI
02-24-2008, 05:03 PM
I think the question we should be asking is, when do we get a LW with an integrated FPrime-esque preview renderer, which, of course, works with all the program's lighting and texturing features.

m*** has one, LW should too. ;)

Lightwolf
02-24-2008, 05:16 PM
I think the question we should be asking is, when do we get a LW with an integrated FPrime-esque preview renderer, which, of course, works with all the program's lighting and texturing features.

Not before a new core is in place. There's just too much stuff happening under the hood that has not been designed with constant querying from a second task in mind.

Cheers,
Mike

IMI
02-24-2008, 05:24 PM
Not before a new core is in place. There's just too much stuff happening under the hood that has not been designed with constant querying from a second task in mind.

Cheers,
Mike


I understand, and agree. Still though, you know they're going to have to do it sooner or later. Sooner would be better for all of us, including them. ;)

Lightwolf
02-24-2008, 05:28 PM
I understand, and agree. Still though, you know they're going to have to do it sooner or later. Sooner would be better for all of us, including them. ;)
I'm sure that if they replace code, they'll keep possibilities like that in mind. But as you know, Rome wasn't built in a day ;)

Cheers,
Mike

Silkrooster
02-24-2008, 05:30 PM
Modo photon splits for is big photon, small photon

LW 9.5 It has?:bowdown:

http://bbs.chinavfx.net/attachments/200802/25055334-snap0120960.jpg
I am seeing the word 'Bucket' Now I wonder...hmmm,
Silk

kfinla
02-24-2008, 05:33 PM
Thanks for the images Secret Squirrel

IMI
02-24-2008, 05:41 PM
I'm sure that if they replace code, they'll keep possibilities like that in mind. But as you know, Rome wasn't built in a day ;)

Cheers,
Mike

Yes! Of course!
I'm not complaining, by far. I'm happy with it as it is... but we all wish for more, don't we? :)

Stooch
02-24-2008, 05:48 PM
Hi, i have been asking NT for a while now to introduce a built in jitter into the lights like the well known noisy channel trick, is this basically what NT did or is this a more "light transport" kind of scheme?

RedBull
02-24-2008, 06:12 PM
I think the question we should be asking is, when do we get a LW with an integrated FPrime-esque preview renderer, which, of course, works with all the program's lighting and texturing features.

m*** has one, LW should too. ;)

I would be happy enough with multiprocessor Viper, and Viper to be exposed in the SDK. These days an 4 or 8 core machine can render a 320x240 preview much, much faster than Viper can, and that really doesn't make much sense.
(previewer slower than final render)

hrgiger
02-24-2008, 06:41 PM
Maybe I should start writing down what will and what won't work with Fprime as changes are made to Lightwave. It's getting harder and harder to keep up.

And if Fprime mimics Lightwave's render engine, why can't it mimic the shaders and lens flares and everything else that doesn't render in Fprime but renders in LW? Just curious is all.

RedBull
02-24-2008, 06:44 PM
And if Fprime mimics Lightwave's render engine, why can't it mimic the shaders and lens flares and everything else that doesn't render in Fprime but renders in LW? Just curious is all.

Communicating between one plugin or shader and another is not easy to do in LW, Nodal made this easier for shaders, but stuff like Lens Flares are a post process and are not exposed in the SDK.

hrgiger
02-24-2008, 06:46 PM
Communicating between one plugin or shader and another is not easy to do in LW, Nodal made this easier for shaders, but stuff like Lens Flares are a post process and are not exposed in the SDK.

I guess what I meant was, why can't it mimic Lens flares themselves (Worley)? I mean, didn't they implement their own motion blur that at the time, was superior to Lightwaves?

Exception
02-24-2008, 07:15 PM
LW 9.5 It has?:bowdown:


Lw 9.5... it no has photons.

It's a different system.

Exception
02-24-2008, 07:17 PM
And if Fprime mimics Lightwave's render engine, why can't it mimic the shaders and lens flares and everything else that doesn't render in Fprime but renders in LW? Just curious is all.

Because Worley hasn't written the code to mimick those parts. Simple as that.

SecretSquirrel
02-24-2008, 07:49 PM
Quick test at a "chunky" designer rug!

Render Time: 8m on a Quad Core 2

(Thought I'd post it here too as the other thread is becoming flooded with crap talk!)

IMI
02-24-2008, 08:38 PM
I guess what I meant was, why can't it mimic Lens flares themselves (Worley)? I mean, didn't they implement their own motion blur that at the time, was superior to Lightwaves?

Lens flare is simulated well-enough in Open GL, in LightWave. Maybe Worley managed motion blur better, but to ask why is X not as good as Y when X and Y are two different entities is.... well....

hrgiger
02-24-2008, 08:41 PM
Lens flare is simulated well-enough in Open GL, in LightWave. Maybe Worley managed motion blur better, but to ask why is X not as good as Y when X and Y are two different entities is.... well....

Except in this case, we can't really compare X and Y when there is in fact no Y.

IMI
02-24-2008, 08:53 PM
Exactly.
Newtek is "X", Worley is "Y". Both are variables. In one case, we have the substantial reality of Newtek's LW 9.5 (where we have the questionable lens flare), and in the other case we have the unrelated Worley (Gods of the Correct Motion Blur), battling each other over who reigns supreme in the court of disassociated contests.

Exception
02-24-2008, 09:04 PM
battling each other over who reigns supreme in the court of disassociated contests.

hahaaa!

I like that court. Been in it all week!

wacom
02-24-2008, 09:40 PM
Modo photon splits for is big photon, small photon

LW 9.5 It has?:bowdown:

http://bbs.chinavfx.net/attachments/200802/25055334-snap0120960.jpg

I watched a tutorial for another render engine doing arch viz work...and they showed that putting in a slight bevel at the corners of a room (in addition to using a real thickness) help with all sorts of lighting issues- this one included. It gives the render engine somewhere to put those "in between" stray photons etc.

squeegie
02-24-2008, 09:48 PM
I watched a tutorial for another render engine doing arch viz work...and they showed that putting in a slight bevel at the corners of a room (in addition to using a real thickness) help with all sorts of lighting issues- this one included. It gives the render engine somewhere to put those "in between" stray photons etc.
You mind sharing a link to said tutorial? I have been looking for such things ( not other software, but arch-viz tutes)

*Pete*
02-24-2008, 09:54 PM
Quick test at a "chunky" designer rug!


thanks:thumbsup:

thats it then, carpets and grass is possible :D

colkai
02-25-2008, 02:00 AM
In all seriousness, I'm enjoying it immensely, and am highly appreciative of all of y'all who are involved with this and making these efforts. :)
Amen Brother! :bowdown:

dee
02-25-2008, 03:07 AM
Quick test at a "chunky" designer rug!

very nice, thanks!

zardrose
02-25-2008, 04:05 AM
Modo Photon Blend ; Fusion

zardrose
02-25-2008, 04:12 AM
Modo photon splits for is big photon, small photon

MPS 1 + 6 100Ray 4 Bounce = 28s


LW copy it

KillMe
02-25-2008, 04:30 AM
Modo photon splits for is big photon, small photon

MPS 1 + 6 100Ray 4 Bounce = 28s


LW copy it


hmmm the gi is splotchy - why we want that lw producing much better quality than that as far as i can see

Captain Obvious
02-25-2008, 04:41 AM
I guess what I meant was, why can't it mimic Lens flares themselves (Worley)? I mean, didn't they implement their own motion blur that at the time, was superior to Lightwaves?
Because motion blur is the sort of thing where there is a "correct solution." Lens flares are more abstract, so to speak. There is no correct solution in the same sense as there is for motion blur.

mav3rick
02-25-2008, 04:51 AM
I'm sure that if they replace code, they'll keep possibilities like that in mind. But as you know, Rome wasn't built in a day ;)

Cheers,
Mike


oh c*r*a*p they learned me different in school :(

starbase1
02-25-2008, 04:51 AM
Because motion blur is the sort of thing where there is a "correct solution." Lens flares are more abstract, so to speak. There is no correct solution in the same sense as there is for motion blur.

Really? I would have thought there was a specific solution for a specific lens at least - it's all down to internal reflections after all...

wavemaster
02-25-2008, 04:53 AM
I would think so too...

mav3rick
02-25-2008, 04:55 AM
Quick test at a "chunky" designer rug!

Render Time: 8m on a Quad Core 2

(Thought I'd post it here too as the other thread is becoming flooded with crap talk!)



hey animal what means quad core 2 ?
i heard of quad core and core 2 duo

and which quad u got ?

Lightwolf
02-25-2008, 04:57 AM
oh c*r*a*p they learned me different in school :(
Hehe, maybe that's the problem, they should have taught you instead of learning you ;)

Cheers,
Mike

Lightwolf
02-25-2008, 04:59 AM
Really? I would have thought there was a specific solution for a specific lens at least - it's all down to internal reflections after all...
True... but afaik most lens flare "filters" use a much simpler approach. Position a few 2D thingies in the image, stretch, blur them and paint them on top of the image. Presto, instant lens flare (which is why you have single, editable elements) ;)

Cheers,
Mike

Panikos
02-25-2008, 05:45 AM
Lens Flare is a post effect. Internally in LW is an invisible Pixel Filter than you cannot change its sequence of execution, especially if it is behind Transparent objects or before/after Noise Reduction etc

Lens Flare cant fit into FPrime due to its way of operation.
However I'd like to have LW Post Filters over FPrime result, however FPrime internal info arent sent to LW buffers.

zardrose
02-25-2008, 05:57 AM
NEW !
Modo photon splits for is big photon, small photon
:bowdown:

http://zardrose.myweb.hinet.net/22.jpg

Erwin Zwart
02-25-2008, 06:24 AM
awesome, I wish kray could split photons like that!!
:stumped:

I like my photons big and bouncy

safetyman
02-25-2008, 06:25 AM
Are you saying that Modo's renderer is better? Not sure what that's all about. Do you really need to make your text so big to make your point. Please explain.

Limbus
02-25-2008, 06:26 AM
NEW !
Modo photon splits for is big photon, small photon
:bowdown:


All your base are belong to us!

Ivan D. Young
02-25-2008, 06:31 AM
This is a typical post as we get closer to a New version of LW, all the Modo folks and Worley crew come out and start to complain, show how much greener the grass is on the other side, and lament that all the new features don't do this or don't do that, boo hoo hoo. Change is gonna happen, not everything can stay working for ever, and we have to learn to adapt to new tools.

theo
02-25-2008, 06:42 AM
Quick test at a "chunky" designer rug!

Render Time: 8m on a Quad Core 2

(Thought I'd post it here too as the other thread is becoming flooded with crap talk!)

I love the overall effect with the tiniest of caveats: terribly sorry, SS, but the front edge looks a bit eerie... as if these strands are readying the rug for a maul-grade pouncing. I can just hear the click-clacking of rug-legs on the resonant surface of the floor as they slowly drag the rug toward a hapless victim...:D 'Course, this has nothing to do with the great work on the fiber itself... just my twisted mind trying to dig up some Monday morning fear.:D

Keep up the great work!

Erwin Zwart
02-25-2008, 07:06 AM
This is a typical post as we get closer to a New version of LW, all the Modo folks and Worley crew come out and start to complain, show how much greener the grass is on the other side, and lament that all the new features don't do this or don't do that, boo hoo hoo. Change is gonna happen, not everything can stay working for ever, and we have to learn to adapt to new tools.

as an owner of LW931, FPrime3.2, kray1.7, modo301 and messiah3 I welcome any information about which parts of the grass is greener and how exactly. I even like to think this helped every app to improve itself. In the area of GI I know kray did for messiah3 and even some modo. I think it also helped in the push for better GI in 9.2 and now 9.5. I hope in the end all platforms will be serious alternatives to do whatever you need in any of them.

SecretSquirrel
02-25-2008, 07:14 AM
NEW !
Modo photon splits for is big photon, small photon
:bowdown:

Please don't post Modo information in this LightWave 9.5 thread, keep it on topic please.

beverins
02-25-2008, 07:20 AM
NEW !
Modo photon splits for is big photon, small photon
:bowdown:



Modo's render looks the same as the LW one, actually, you just cheated and moved the light to make LW look bad. Why don't we post a Maxwell or Fryrender image, and purr and coo about how these are light simulators?

Oh, and your English there? To give you an idea on how bad your English is, it would be like me typing random Chinese characters.

Stooch
02-25-2008, 07:28 AM
Hi, i have been asking NT for a while now to introduce a built in jitter into the lights like the well known noisy channel trick, is this basically what NT did or is this a more "light transport" kind of scheme?

did i use invisible font or something?

colkai
02-25-2008, 07:51 AM
Oh, and your English there? To give you an idea on how bad your English is, it would be like me typing random Chinese characters.
Me fail English - that's un-possible! (random Simpsons quote inserted for levity) :)

Phil
02-25-2008, 07:56 AM
Quick test at a "chunky" designer rug!

Render Time: 8m on a Quad Core 2

(Thought I'd post it here too as the other thread is becoming flooded with crap talk!)

Looks nice. I'd be tempted to stick the camera in amongst the fibers and look up at the table :D

mav3rick
02-25-2008, 07:56 AM
im not good in english either but man i dont get a word from that crazy modo guy

beverins
02-25-2008, 08:10 AM
Look, I'm all for people with weak English to share on this board. English is hard to learn. I'm far from being a grammar nazi, so I understand. But his post was both an attempt at starting a flamewar AND nonsensical. :cat: I think what he was trying so desperately to show was that Modo also has IES lights in some form. From what I have seen so far, though, Lightwave's model is superior to Modo's at the moment.

Anyway, back to the topic at hand - 9.5.

I like the IES lights, nice stuff there. Are the settings envelopable in regard to the IES settings? Can someone try a NEGATIVE IES light to see what it looks like?

Raccoon Bandit
02-25-2008, 08:14 AM
I see plenty of native English speakers who are everything other than masters of their own written language.

You talking about me again? :help:

No crashes last night, both apps played nicely together so heres the larger Photometric Vase render. More stuff to come later this week.

Raccoon Bandit
02-25-2008, 08:16 AM
Can someone try a NEGATIVE IES light to see what it looks like?

Have not tried them yet but surely will today and show what happens.

Wade
02-25-2008, 08:16 AM
You talking about me again? :help:

No crashes last night, both apps played nicely together so heres the larger Photometric Vase render. More stuff to come later this week.


Very nice! :thumbsup:

Mitja
02-25-2008, 08:21 AM
With IES lights, how do you match the real light intensity in LW? I mean: real light: 100W - LW light: 100% ?

*Pete*
02-25-2008, 08:24 AM
id like to see an inverted spherical light with falloff, halfway sunk though a plane so that it would create a ring of light, with different falloff ranges.

Lightwolf
02-25-2008, 08:25 AM
With IES lights, how do you match the real light intensity in LW? I mean: real light: 100W - LW light: 100% ?
Normally you don't... they're a part of the IES file.

how they translate into LWs lighting, which hasn't been designed with real world lights in mind... I dunno. I believe Denis made a few experiments in that direction...

Cheers,
Mike

*Pete*
02-25-2008, 08:26 AM
With IES lights, how do you match the real light intensity in LW? I mean: real light: 100W - LW light: 100% ?

most likely no, i quess we will control the light as before.

Mitja
02-25-2008, 08:31 AM
Normally you don't... they're a part of the IES file.

how they translate into LWs lighting, which hasn't been designed with real world lights in mind... I dunno. I believe Denis made a few experiments in that direction...

Cheers,
Mike
I saw Denis experiments (conversions between lux, watts, lumens & co), but it was a bit confusing, I didn't understand a comma! The situation was different than with the new lights, anyway. With "normally you don't..." you mean that IES lights in other 3d applications don't have any intensity setting?

Lightwolf
02-25-2008, 08:36 AM
I saw Denis experiments (conversions between lux, watts, lumens & co), but it was a bit confusing, I didn't understand a comma! The situation was different than with the new lights, anyway. With "normally you don't..." you mean that IES lights in other 3d applications don't have any intensity setting?
Two ways: Either they set the intensity of the light when you load it - depending on the IES data. Or they use it internally and allow you to multiply it using the "normal" lighting controls.

Ideally you don't need to understand it. Use the file and get the right lighting for that type of light.

Cheers,
Mike

Mitja
02-25-2008, 08:36 AM
most likely no, i quess we will control the light as before.
Hope we'll have at least a conversion formula! The best would be an input directly in watts...

colkai
02-25-2008, 08:42 AM
Ideally you don't need to understand it. Use the file and get the right lighting for that type of light.

Cheers,
Mike
Sorry if this is a dumb question, but does that mean you don't have 1 IES file for the fitting, but 1 for that fitting and each permutation of bulb wattage.

e.g. Halogen GU10 35w, Halogen GU10, 15w, Halogen GU10 50w
..rather than 1 file where you set the intensity for that light some other way?



(He who is pretty clueless about all this ;) )

Mitja
02-25-2008, 08:43 AM
Use the file and get the right lighting for that type of light.

And what if one wants to do a lighting simulation?

Just a question: in layout - is there an array command?

*Pete*
02-25-2008, 08:44 AM
Hope we'll have at least a conversion formula! The best would be an input directly in watts...

i doubt it would serve a purpose, percent or watt are basically the same..the "strenght" of the light on a real photograph depends very much of the camera settings...and often even by how used your own eyes are to the light, ever noticed how dim rooms tend to "brighten up" the longer you spend time inside?

Perhaps some day LW will have a solution for this, but untill then i quess the best is to "light by estimation"

meatycheesyboy
02-25-2008, 08:45 AM
hey animal what means quad core 2 ?
i heard of quad core and core 2 duo

and which quad u got ?

I'm going to guess he meant Core 2 Quad, not Quad Core 2. Intel's newest naming scheme makes it easy to mix things up. As to which one he has, there are only a few available, if I had to guess, I'd say probably a Q6600. But that would be a guess.

wavemaster
02-25-2008, 08:46 AM
Plus it all depends on your surfacing...

Lightwolf
02-25-2008, 08:59 AM
Sorry if this is a dumb question, but does that mean you don't have 1 IES file for the fitting, but 1 for that fitting and each permutation of bulb wattage.
Yes.
One file for every product the manufacturer ships (if they provide the files that is).

Cheers,
Mike
Edit: And no, it's not a dumb question...

Raccoon Bandit
02-25-2008, 10:35 AM
The fun part is that you can accumulate thousands of IES files in minutes. I'm not sure what your definition of fun is but if you like lots of options they are surely and freely available to you.

Andyjaggy
02-25-2008, 10:36 AM
Theres a lot we could say and heck show, some of which I'd kill to show right now, but alas a lot of goodies are still off limits for us to talk about/show. And don't take that wrong, its actually a good thing not a bad thing. Think about it like this you have all your Christmas presents sitting out in front of the tree at the start of December. Everyones gone so you open everything to see what you got and then wrapped them back up. The rest of December and especially Christmas day would suck.

I'm hoping this means gradient falloffs on all the lights :D

Magnus81
02-25-2008, 11:14 AM
Does anyone have any info or renders on some of the character tools? Secret Squirrel? Raccoon Bandit? Anything?

Magnus81
02-25-2008, 11:20 AM
Darn! I opened the wrong thread. I meant to post this in the other thread. Sorry guys. I must have taken my stupid pills this morning.

kopperdrake
02-25-2008, 12:06 PM
I love the overall effect with the tiniest of caveats: terribly sorry, SS, but the front edge looks a bit eerie... as if these strands are readying the rug for a maul-grade pouncing. I can just hear the click-clacking of rug-legs on the resonant surface of the floor as they slowly drag the rug toward a hapless victim...:D 'Course, this has nothing to do with the great work on the fiber itself... just my twisted mind trying to dig up some Monday morning fear.:D

Keep up the great work!

Just a thought about the rug, but if you can add balls to the fibre strand vertices, can you add a fibre-clump to a fibre strand vertice? Á like parenting a group of fibres to a fibre?

Íf this was the case you could create a real 'woolly' fibre strand with several fibres that is attached to the base vertice of each of those clumpy single fibres on your rug.

Does this make sense? Ánd apologies for the weird terminology - not sure what the various 'things' are called yet :D

Wickster
02-25-2008, 12:06 PM
The fun part is that you can accumulate thousands of IES files in minutes. I'm not sure what your definition of fun is but if you like lots of options they are surely and freely available to you.
Yeah tell me about it...5000+ IES files and counting.

Lightwolf
02-25-2008, 12:22 PM
I'm hoping this means gradient falloffs on all the lights :D
If not on release... there is a light SDK. I'm pretty sure you'll see plenty of creative light options being released.

Cheers,
Mike

mav3rick
02-25-2008, 12:27 PM
You talking about me again? :help:

No crashes last night, both apps played nicely together so heres the larger Photometric Vase render. More stuff to come later this week.


how sweet
a flower :)
i am all wet now :P
hahahah


no really nice render

SecretSquirrel
02-25-2008, 12:49 PM
What I'm finding with Photometric lights is that I just pick some that I like the look of and use them, whether they are 'correct' for that lamp type / application.

But normally you don't have to adjust the LW light intensity as that comes from the IES file.

But if it is too bright / not bright enough, you can reduce / increase the LW light intensity to compensate.

It's a little different to how you normally use lights, it will take some experimentation when you finally get your hands on them. In some ways, it's best to think of them as a light effect rather than a 'physically accurate light'.

It'll make sense when you see it.

SecretSquirrel
02-25-2008, 12:56 PM
Nothing special really, just playing with GI, 2m 12s on a Pentium Quad Core 2 Extreme 2.66GHz

Backdrop lit, no lights.

Mitja
02-25-2008, 12:56 PM
But normally you don't have to adjust the LW light intensity as that comes from the IES file.

Thanks for this heads-up!

Mitja
02-25-2008, 12:58 PM
Nothing special really, just playing with GI, 2m 12s on a Pentium Quad Core 2 Extreme 2.66GHz
And iTunes in background? :hey:

SecretSquirrel
02-25-2008, 01:03 PM
And iTunes in background? :hey:

Nope, University Challenge is on TV in the background though! :)

Raccoon Bandit
02-25-2008, 01:17 PM
Someone asked about seeing negative valued Photometric lights. I did one better heres a series of renders of a simple cornel-esque box setup showing various intensities of an Photometric light. The box has a single point light in it set to 25% intensity with no ambient and no GI. The only reason for this light is to show the effects of negative values.
I added a single photometric light very near the back wall and simply rendered with various intensities as indicated on the images. The only thing done in post was comping all images into one sheet.

Earl
02-25-2008, 01:21 PM
Very nice, thanks Raccoon. That helps out w/ understanding the IES lights.

DangerKitty
02-25-2008, 01:27 PM
Secret Squirrel -

Nope, University Challenge is on TV in the background though!
AHA - Secret Squirrel the game is up. The agents of KAOS (Kitties Against Operational Squirrels) are surrounding the UK as we speak :devil:

eyelandarts
02-25-2008, 01:30 PM
Nice Racoon, do IES lights work with volumetrics? If so, what do they look like?
Thanks.

Stooch
02-25-2008, 02:07 PM
Hi, i have been asking NT for a while now to introduce a built in jitter into the lights like the well known noisy channel trick, is this basically what NT did or is this a more "light transport" kind of scheme?

*Pete*
02-25-2008, 02:11 PM
Hi, i have been asking NT for a while now to introduce a built in jitter into the lights like the well known noisy channel trick, is this basically what NT did or is this a more "light transport" kind of scheme?

most likely they will not credit you for that ;)
if its a noice channel trick, i hope its adjustable.

JeffrySG
02-25-2008, 02:36 PM
Yeah tell me about it...5000+ IES files and counting.
The problem is to sort through all the IES files to find the fixtures you really like and use... I would imagine that in the end we'll all have about 10 - 20 IES files that we use most of the time.

Steamthrower
02-25-2008, 02:39 PM
Recessed downlight, parabolic troffer, linear pendant, track head, bollard...there's really only a few when you get down to it. All of the fixtures in a family have almost identical photometrics.

SecretSquirrel
02-25-2008, 03:09 PM
Repeated... I want to know it, too.
There is a renderer that supports soft shadows with Directional Lights using a sort of Spread Angle. I guess you're refering to that? Would be cool. :)

My furry lips are sealed, but all I can say is this ... :D ;)

Hehe!

Raccoon Bandit
02-25-2008, 03:28 PM
With all the nuts you have shoved in your cheeks how can you have sealed lips?:neener:

Maxx
02-25-2008, 03:40 PM
With all the nuts you have shoved in your cheeks how can you have sealed lips?:neener:
That's just ... wrong. But then again, I'm mentally 7 :D .

About the caustics - I'm assuming that these haven't seen any updates during the last few GI updates due to NT knowing that the lighting system was going to undergo the change we're apparently seeing in 9.5. So, does that mean that we'll be seeing some useful caustics as of 9.6?

Anyone? Bueller? Bueller?

Andyjaggy
02-25-2008, 03:53 PM
well the screenshot chuck showed us had the caustics panal on it and it looked like the same old stuff we have always had. Which is a shame because LW caustics suck.

SecretSquirrel
02-25-2008, 04:08 PM
To put you out of your misery, at present (because who knows with the speed the development team are knocking out updates!) Caustics are the same as previous LW, and are not implemented for the new lights (yet).

I could be wrong, but that's my understanding.

(Hope it's okay to mention this! Runs for cover up a tree!)

serge
02-25-2008, 04:18 PM
With the new Light API, is it possible that third parties can make caustics plugins?

Iaian7
02-25-2008, 04:21 PM
LW Caustics are, I agree, practically unusable. However, with the improving radiosity (which intrinsically supports caustic effects), might we see something completely different? Perhaps the lights themselves could be integrated with the radiosity engine, so it's not just bounces and luminous geometry that are focusable by refractive surfaces?

Though there are advantages to having a separate system (specifically if one does not need radiosity, or simply the ability to tune the systems individually), would we even need the old caustics system any more, if the lights were (somehow) integrated with the radiosity engine?

Just my non-programmer thoughts on the subject.

Thanks to all the alpha / beta testers, I'm loving the previews of 9.5!

Magnus81
02-25-2008, 04:23 PM
With new API is it possible to swap light properties with other lights. For example, giving an area light the falloff and projection map of a spotlight?

Maxx
02-25-2008, 04:23 PM
To put you out of your misery, at present (because who knows with the speed the development team are knocking out updates!) Caustics are the same as previous LW, and are not implemented for the new lights (yet).

I could be wrong, but that's my understanding.

(Hope it's okay to mention this! Runs for cover up a tree!)
Thanks for the update on that - hopefully we'll see it soon.

wacom
02-25-2008, 04:43 PM
You mind sharing a link to said tutorial? I have been looking for such things ( not other software, but arch-viz tutes)

Well it was for archviz in XSI from digital tutors. The video was not mind blowing by any stretch of the imagination, but it was somewhat informative, and more thorough than 90% of their bunk training.

Anyway- yes...as far as I know mr does this too (hence max and min radius etc.) but the key here is that, while it's cute and nice to have the computer do that work, it would still help it if you gave it some place to stick those photons that want to be on each triangle but just can't. Putting in a certain amount of bevel helps alleviate this problem in many instances. I don't see why it shouldn't be true for any render engine with some form of interpolated radiosity/FG solution.

(Until literally today there was no good easy way to thicken many walls etc in XSI (until Kim Aldis saved our butt)! The video I mentioned above shows how painful this can be...as the guy basically says sorry to the audience and then goes about manually putting the thickened walls in! Then again, it did show off the awesome tweak tools abilities...but no thickner? Shows how LW is ahead of the game in some regards when it comes to straight up modeling- it's when you start to have to use that stuff to animate...)

tischbein3
02-25-2008, 04:54 PM
thanks for the info, Secret Squirrell.

Chuck
02-25-2008, 05:32 PM
Folks, a moderated sticky thread for posting v9.5 images and animations is now present at:

http://www.newtek.com/forums/showthread.php?t=80543

I've launched it with this message:


This thread is for sample images and animations from v9.5. No other posts are allowed on the thread. Any commentary can be posted in the other available discussion threads on the v9.5 topics. Off-topic posts will be deleted; repeated off-topics posts to the thread will get the offender a temporary or permanent ban.

We do encourage that new v9.5 images also be shared on the existing threads or new threads for full dicussion, but for those who just want to see the images without commentary, this will be the place.

lardbros
02-25-2008, 05:45 PM
Folks, a moderated sticky thread for posting v9.5 images and animations is now present at:

http://www.newtek.com/forums/showthread.php?t=80543

I've launched it with this message:


Good stuff... thanks Chuck! Means we all don't have to sift through crappy messages like this one i'm writing here. Although have been enjoying all the banter from the Secret Mammals!

SecretSquirrel
02-25-2008, 05:53 PM
With the new Light API, is it possible that third parties can make caustics plugins?

With the new light API, anything is possible, only limited by how good a coder you are! I said it before and I'll say it again, NT have done an awesome job of implementing this! It really is a thing of beauty!

Panikos
02-25-2008, 06:47 PM
It was about time.
I remember I requested this some years ago, LW7.*

Weepul
02-25-2008, 07:30 PM
With the new light API, anything is possible, only limited by how good a coder you are! I said it before and I'll say it again, NT have done an awesome job of implementing this! It really is a thing of beauty!
Looks like I need to learn C really soon! :D (Or I could keep begging those who DO know how to write LW plugins to implement my ideas... :hey: )

Pomfried
02-26-2008, 06:54 AM
Looks like I need to learn C really soon! :D (Or I could keep begging those who DO know how to write LW plugins to implement my ideas... :hey: )


Uuh, thought I would just jump in here, since it's been a while I last posted :P
And yeah, I'm really waiting for the new light API as I would like to try out if I am able to make a light type out of one of the node ideas you gave me (hint hint, I guess you know which one that is :) BTW, so you know I'm not dead: most of the new nodes are done I guess, and bug fixes are there as well, though a few things are left to be coded. And well... The biggest node of all these is still waiting to be coded... you know... the aniso one :) but enough of that OT stuff).

What I really wonder though, is to what extent we are given control over the lights? Like do we have to calculate caustics ourselves and things like that. Hm.. time will tell (or a dev from NT? :bowdown:)

Weepul
02-26-2008, 05:01 PM
Uuh, thought I would just jump in here, since it's been a while I last posted :P
:p


And yeah, I'm really waiting for the new light API as I would like to try out if I am able to make a light type out of one of the node ideas you gave me (hint hint, I guess you know which one that is :) BTW, so you know I'm not dead: most of the new nodes are done I guess, and bug fixes are there as well, though a few things are left to be coded. And well... The biggest node of all these is still waiting to be coded... you know... the aniso one :) but enough of that OT stuff).
I'd be happy to give some input while you're working on any of those...particularly that one light type. :thumbsup: From what I've seen, that one's probably going to have a few different incarnations from different programmers, as popular as the concept will (likely) be. And I've another light type in mind, though I'm not entirely decided how its accompanying data files would be generated in LW... :D

Pomfried
02-27-2008, 06:48 AM
I'd be happy to give some input while you're working on any of those...particularly that one light type. :thumbsup:
Hehe yeah, that's no problem at all. I will hopefully finish school in a few months and then (if I have enough interest in programming at that time :D) I will have more time for all my coding needs :)


From what I've seen, that one's probably going to have a few different incarnations from different programmers, as popular as the concept will (likely) be.
Yep, sounds very likely. The question is, whether I should still include it as a surface node, at least until 9.5 is out of beta :stumped:


And I've another light type in mind, though I'm not entirely decided how its accompanying data files would be generated in LW... :D
Uuuh, that sounds interesting already :D

sadkkf
02-27-2008, 11:37 AM
I watched a tutorial for another render engine doing arch viz work...and they showed that putting in a slight bevel at the corners of a room (in addition to using a real thickness) help with all sorts of lighting issues- this one included. It gives the render engine somewhere to put those "in between" stray photons etc.

Do you have a link to that vid?

sadkkf
02-27-2008, 12:12 PM
D'oh. Never mind. I missed your last post.

A Mejias
02-28-2008, 12:28 PM
It would be real helpful to have side by side time and quality comparisons with 9.3 renders, when posting GI and other light type examples here and in the Sticky thread. In many cases it's really hard to tell what the improvments are with out it.

Thanks!

kcole
03-03-2008, 02:02 PM
Any way we can get the 9.5 image thread unlocked long enough for me to subscribe to it? :D

Medi8or
03-03-2008, 02:14 PM
Any way we can get the 9.5 image thread unlocked long enough for me to subscribe to it? :D
Enter thread, click "thread tools" right above first post, and you find "Subscribe to this thread".
I think that works even if the thread is locked.. :)

Silkrooster
03-03-2008, 04:29 PM
Just tried subscribing to the image thread. It won't work. Never thought about subscribing to it when it wasn't locked. Oh well easy enough to find.
Silk

robk
03-03-2008, 06:21 PM
Folks, a moderated sticky thread for posting v9.5 images and animations is now present at:

http://www.newtek.com/forums/showthread.php?t=80543

I've launched it with this message:

The only problem is Chuck, that since you posted this message the only pictures/anims have been posted are a rehash of the pictures that were already posted already. I don't think anything has been posted for about 5 days.
At least when we were hassling the 9.5 picture posters we were getting something to look at. An since your post NADA!!!!!

IMI
03-03-2008, 06:38 PM
Well, it being locked might not help as far as people uploading new piccies. ;)
When and why did it get locked, anyway? The original message said that comments wouldn't be allowed, but that implies it was opened. And it was open, at one point.
It was also said that people who broke that rule might be temporarily banned, also implying it would be open. If it were locked, there'd be no chance of that happening.
Was it being abused when I wasn't paying attention, or have the incognito beta-critters been told to post no more images?

DragonFist
03-03-2008, 07:38 PM
Cause peeps didn't follow said rules and posted comments. Actually, I don't know if that IS the reason. I just know that it happened and am making an assumption.

ftowns
05-10-2008, 06:37 AM
very nice images