The Official aweSurface Test Track

wowie

I'd still recommend not using them. But sometimes you still need a distant light, at least right now anyway.

The shadow catcher works with shadows cast from point/spot/distant and area lights. Shadows from HDRI tends to look like AO, so if you need much more defined shadows, just use an area light.

Thank you all for the feedback and bug reports.

Sven Dullah

wowie said:

I'd still recommend not using them. But sometimes you still need a distant light, at least right now anyway.

The shadow catcher works with shadows cast from point/spot/distant and area lights. Shadows from HDRI tends to look like AO, so if you need much more defined shadows, just use an area light.

Thank you all for the feedback and bug reports.

My pleasure:) Thank you for all the support and tips! Now let round 2 begin

Sven Dullah

...a question about caustics, like a glass of water on a table...with pathtracing it should be possible in theory to get that effect, if the rays are allowed to bounce an enough number of times? Are there other limiting factors that makes this practically impossible? I'm thinking mesh density would be one such limit? Any thoughts on this matter?

And while I'm at it, how can the photon shading model be utilized? I know how to do it in shader mixer, but my understanding is that awe is incompatible with those shaders?

wowie

Sven Dullah said:

...a question about caustics, like a glass of water on a table...with pathtracing it should be possible in theory to get that effect, if the rays are allowed to bounce an enough number of times? Are there other limiting factors that makes this practically impossible? I'm thinking mesh density would be one such limit? Any thoughts on this matter?

You don't need multiple bounces for caustics. However, the current path tracer don't resolve cleanly (it actually produces caustic for 'free'). For practical use, photon mapping is still the way to go.

Sven Dullah said:

And while I'm at it, how can the photon shading model be utilized? I know how to do it in shader mixer, but my understanding is that awe is incompatible with those shaders?

Technically, yes. But I think you either need a very small emitter (area light) or use the point/spotlight. I never got it to work though.

Sven Dullah

wowie said:

Sven Dullah said:

...a question about caustics, like a glass of water on a table...with pathtracing it should be possible in theory to get that effect, if the rays are allowed to bounce an enough number of times? Are there other limiting factors that makes this practically impossible? I'm thinking mesh density would be one such limit? Any thoughts on this matter?

You don't need multiple bounces for caustics. However, the current path tracer don't resolve cleanly (it actually produces caustic for 'free'). For practical use, photon mapping is still the way to go.

Sven Dullah said:

And while I'm at it, how can the photon shading model be utilized? I know how to do it in shader mixer, but my understanding is that awe is incompatible with those shaders?

Technically, yes. But I think you either need a very small emitter (area light) or use the point/spotlight. I never got it to work though.

Ok, yes I've noticed some caustics but they are kind of inaccurate, I thought maybe a higher polygon count would help, but you're saying it's limitations within the pathtracer itself?

This is quite nice IMO (the bottle creating some caustics on the table,) click to enlarge;)

another one

wowie

Sven Dullah said:

Ok, yes I've noticed some caustics but they are kind of inaccurate, I thought maybe a higher polygon count would help, but you're saying it's limitations within the pathtracer itself?

Well, no. The path tracer like any other path tracer is capable of producing caustics, but the devs didn't expose any method of handling it so it renders cleanly. I made a practical choice to basically clamp caustic producing paths since they're very susceptible to fireflies. I guess you could say it's a feature of the shader. Path tracing caustics doesn't require high polygon count, unlike photon mapping. You only need high polygon count if you're doing it with polygon models anyway.

The results you're seeing on both renders can still be classified as color bleeding. Caustics have way higher energy and more focused. Hence, increasing the probability of getting fireflies. You can also fake it with transmission shadow, though I haven't tried it in a while so I don't know if it works or not. By that, I mean it should work with thin polygons when you have 'Use Face Forward' enabled.

Sven Dullah

wowie said:

Sven Dullah said:

Ok, yes I've noticed some caustics but they are kind of inaccurate, I thought maybe a higher polygon count would help, but you're saying it's limitations within the pathtracer itself?

Well, no. The path tracer like any other path tracer is capable of producing caustics, but the devs didn't expose any method of handling it so it renders cleanly. I made a practical choice to basically clamp caustic producing paths since they're very susceptible to fireflies. I guess you could say it's a feature of the shader. Path tracing caustics doesn't require high polygon count, unlike photon mapping. You only need high polygon count if you're doing it with polygon models anyway.

The results you're seeing on both renders can still be classified as color bleeding. Caustics have way higher energy and more focused. Hence, increasing the probability of getting fireflies. You can also fake it with transmission shadow, though I haven't tried it in a while so I don't know if it works or not. By that, I mean it should work with thin polygons when you have 'Use Face Forward' enabled.

Ok tks! Good tip about use face forward

Mustakettu85

wowie said:

The path tracer like any other path tracer is capable of producing caustics, but the devs didn't expose any method of handling it so it renders cleanly.

Actually the "why" here is that 3Delight, like, say, Arnold, is still a unidirectional pathtracer, this is why to sample raytraced caustics cleanly and correctly you need a gazillion samples. 

Here's a page from Arnold docs; the explanation in the box works for 3Delight, too:

https://docs.arnoldrenderer.com/display/A5AFMUG/Refractive+Caustics+using+an+Emissive+Shader

 

 

Sven Dullah said:

And while I'm at it, how can the photon shading model be utilized? I know how to do it in shader mixer, but my understanding is that awe is incompatible with those shaders?

The shader mixer photon-mapping camera is a mess, but if it doesn't crash DS for you, it may work. 

AweArea lights don't have "emit photons" on, but my RadiumPT (included with the latest free awe update) very much do.

And aweSurface has all the photon shading models - you'll need "water" or "glass" for your refractive surfaces.

So you have all the ingredients for a yet another experiment =)

wowie

Mustakettu85 said:

wowie said:

The path tracer like any other path tracer is capable of producing caustics, but the devs didn't expose any method of handling it so it renders cleanly.

Actually the "why" here is that 3Delight, like, say, Arnold, is still a unidirectional pathtracer, this is why to sample raytraced caustics cleanly and correctly you need a gazillion samples. 

Here's a page from Arnold docs; the explanation in the box works for 3Delight, too:

https://docs.arnoldrenderer.com/display/A5AFMUG/Refractive+Caustics+using+an+Emissive+Shader

Yeah, but they don't use tricks like Renderman's new Manifold Next Event Estimation.

Mustakettu85 said:

AweArea lights don't have "emit photons" on, but my RadiumPT (included with the latest free awe update) very much do.

Actually, you could check what's inside aweAreaPTAttribs.dsa - line 247.

Shader.setStringAttrib( "light", "emitphotons", "on");

Should be there on the freebie point/spot/distant light too. But I never get stuff even with your RaytracerCaustics render script. So I never tested or focused on it during development.

Sven Dullah

Tks wowie and Kettu I'll start by testing the shader mixer camera, wish me luck

Sven Dullah

https://www.daz3d.com/solitaire-by-meipe converted to awe...rendered at 12x12 pixelsamples and 2048 Irradiance samples. Every surface in the original is a custom shader and renders gray in the viewport, so pretty much remade everything;)

There are some jaggies on the windows and curtains even with 12x12, guess some more DoF would have helped...or postwork...

wowie

Sven Dullah said:

There are some jaggies on the windows and curtains even with 12x12, guess some more DoF would have helped...or postwork...

Looks awesome. If you're seeing jaggies even at 12x12, render at higher resolution. Is this with the old build or the new one?

Sven Dullah

wowie said:

Sven Dullah said:

There are some jaggies on the windows and curtains even with 12x12, guess some more DoF would have helped...or postwork...

Looks awesome. If you're seeing jaggies even at 12x12, render at higher resolution. Is this with the old build or the new one?

Still on the old build;) Getting there...

wowie

Add some bump to add bits of imperfections here and there. Might also want to look into slightly higher roughness with lower roughness at grazing angles for the table on the left.

Sven Dullah

wowie said:

Add some bump to add bits of imperfections here and there. Might also want to look into slightly higher roughness with lower roughness at grazing angles for the table on the left.

Tks, will look into that;)

Sven Dullah

Testing aweSurface with a shadermixer photon mapping cam and various spotlights. The DS standard spot doesn't emit photons and renders black surfaces through the photon camera, not a big surprise. The dz shader spotlight renders the surfaces correctly, but doesn't emit photons, so no caustics.

The test scene is a plane with a mid gray rough surface, a sphere with awe glass preset 1 with a deep blue transmission color, no diffuse, and an environmental sphere with indirectlight/occlusion off. Surprisingly enough I managed to render with a shadermixer caustic spotlight and raytracer final;) The initializing stage took 20min and the actual render 2-3 min. Not very pretty though

Then created a photon mapper spot,rendered slightly faster...

Sven Dullah

Also tested with the awe area light through the photon mapper cam, no caustics, only takes longer to render. Think I need to forget this idea, was worth a try

Edit: Hmm would it be worth trying Kettu's light shaders? Anyone tested them with the photon mapper cam?

Still think the aweSurface with proper transmission settings looks best so far;)

khorneV2

hi,

i have a question about iblm and awe lights and surfaces. Render times are high. I guess I saw something about in the threads but i am lost in the references.

So, can someone tell me if there is/are a special procedure / special settings to use both ?

thanks in advance

regards

Sven Dullah

khorneV2 said:

hi,

i have a question about iblm and awe lights and surfaces. Render times are high. I guess I saw something about in the threads but i am lost in the references.

So, can someone tell me if there is/are a special procedure / special settings to use both ?

thanks in advance

regards

I've only used them together because 1)there was no awe shadowcatcher, but now we have one, and 2)garibaldi hair won't render with scripted pathtracing.

What's in your scene, is it HDRI lighting and/or something else? I'm not at my DS computer right now so don't recall exactly what settings I've used;) I can have a look  at it later, if you provide some more info on your scene content

@wowie  Will there be an official DIM update or do we have to install the new build manually?

wowie

Manual for now. There some changes I've made since uploading so it looks like the DIM version will be something like 1.11 or something like that.

Sven Dullah

wowie said:

Manual for now. There some changes I've made since uploading so it looks like the DIM version will be something like 1.11 or something like that.

Ok!

wowie said:

Mustakettu85 said:

wowie said:

The path tracer like any other path tracer is capable of producing caustics, but the devs didn't expose any method of handling it so it renders cleanly.

Actually the "why" here is that 3Delight, like, say, Arnold, is still a unidirectional pathtracer, this is why to sample raytraced caustics cleanly and correctly you need a gazillion samples. 

Here's a page from Arnold docs; the explanation in the box works for 3Delight, too:

https://docs.arnoldrenderer.com/display/A5AFMUG/Refractive+Caustics+using+an+Emissive+Shader

Yeah, but they don't use tricks like Renderman's new Manifold Next Event Estimation.

Mustakettu85 said:

AweArea lights don't have "emit photons" on, but my RadiumPT (included with the latest free awe update) very much do.

Actually, you could check what's inside aweAreaPTAttribs.dsa - line 247.

Shader.setStringAttrib( "light", "emitphotons", "on");

Should be there on the freebie point/spot/distant light too. But I never get stuff even with your RaytracerCaustics render script. So I never tested or focused on it during development.

Would it be possible to implement an emit photons on toggle button I simply suck at editing scripts and the likes Or am I only dreaming of being able to produce some believable caustics with pathtracing?

Oh btw I found out a nice thing about the area light, you can actually remove limits for the intensity scale channel:) Very handy if you want a microscophic emitter with very high intensity (think laser)

 

wowie

Sven Dullah said:

Would it be possible to implement an emit photons on toggle button I simply suck at editing scripts and the likes Or am I only dreaming of being able to produce some believable caustics with pathtracing?

The way it is now, lights are always flagged to emit photons. I think mustakettu's script that I bundle don't actually call the photon hider, which is necessary for photon mapping. She used a different script to enable that.

Sven Dullah said:

Oh btw I found out a nice thing about the area light, you can actually remove limits for the intensity scale channel:) Very handy if you want a microscophic emitter with very high intensity (think laser)

Remember exposure is already exponential. Both intensity scale sliders are added together, so technically if you use 8 for both, you're setting the intensity to the 16th level (65K) the original amount. It might just be best to use a linear or no falloff instead. Work real nice if you do want a narrow beam of light, though I don't think it's quite as narrow as a laser beam.

khorneV2

HDRI + IBLM, i saw some post about redudant Ambient occlusion and i would like to have a trick about best settings

Sven Dullah

wowie said:

Sven Dullah said:

Would it be possible to implement an emit photons on toggle button I simply suck at editing scripts and the likes Or am I only dreaming of being able to produce some believable caustics with pathtracing?

The way it is now, lights are always flagged to emit photons. I think mustakettu's script that I bundle don't actually call the photon hider, which is necessary for photon mapping. She used a different script to enable that.

Aah ok, that explains why I didn't get any error messages when using the awe arealight with the photon mapper cam:)

wowie said:

Sven Dullah said:

Oh btw I found out a nice thing about the area light, you can actually remove limits for the intensity scale channel:) Very handy if you want a microscophic emitter with very high intensity (think laser)

Remember exposure is already exponential. Both intensity scale sliders are added together, so technically if you use 8 for both, you're setting the intensity to the 16th level (65K) the original amount. It might just be best to use a linear or no falloff instead. Work real nice if you do want a narrow beam of light, though I don't think it's quite as narrow as a laser beam.

Yup, that might be a better solution.

Here is some more caustics testing with the same scene, turned on the environmental indirect light/occlusion and played with intensity settings...all rendered through the photon mapping cam...

Using a small area light emitter:

Using a caustics spotlight:

Same thing, decreased caustics intensity:

Sven Dullah

khorneV2 said:

HDRI + IBLM, i saw some post about redudant Ambient occlusion and i would like to have a trick about best settings

...so I did some testing but  did not come up with anything really useful, sorry...checked all hidden shaderbuilder parameters inside IBLM and could not disable the ambient occlusion. Also tried turning off GI for the aweSurfaces, looked like crap. Two testrenders:

pure awe render without a shadowcatcher plane (haven't installed the new build yet), rendertime 1min50sec:

 

IBLM render with the IBLM groundplane, rendertime 4min50sec:

Both rendered with raytracer final pixelsamples at 10x10.

So I don't see a reason for combining the two, unless you need to render the garibaldi hair or using the IBLM groundplane. I also rendered the IBLM version in regular 3DL progressive mode, shaved of 30 sec, so not much of an option. What you can do to shorten rendertimes when using the IBLM groundplane is to scale it down as much as possible without clipping the shadows. And regarding the double occlusion, you can play with the IBLM shading strength, maybe lower it to 85% or thereabout, won't make it render faster though.

IMO the fastest way with IBLM is to use DS default or UBerSurface shaders and render in regular 3DL with progressive enabled. This scene would probably have rendered in a minute with the default shader.

If somebody has a better tip, please share!

ETA: You can also select the IBLM environment sphere, go to the surface pane and disable diffuse, will speed up rendering and then the IBLM light intensity won't affect the background.

wowie

Sven Dullah said:

pure awe render without a shadowcatcher plane (haven't installed the new build yet), rendertime 1min50sec:

Sven Dullah said:

IBLM render with the IBLM groundplane, rendertime 4min50sec:

Hmm, am I reading this right? AWE with GI is faster than IBL Master? I assume irradiance samples is set to 128?

Sven Dullah

wowie said:

Sven Dullah said:

pure awe render without a shadowcatcher plane (haven't installed the new build yet), rendertime 1min50sec:

Sven Dullah said:

IBLM render with the IBLM groundplane, rendertime 4min50sec:

Hmm, am I reading this right? AWE with GI is faster than IBL Master? I assume irradiance samples is set to 128?

Yep that's right:) Except for the shirt (512). The groundplane naturally adds some time to the IBLM render.

mgtskynet

Hello guys, is it possible to turn of occlusion without also having to turn of lighting ?

It really kills render times. Especially when using with ibl master or uberenvironment.

wowie

mgtskynet said:

Hello guys, is it possible to turn of occlusion without also having to turn of lighting ?

It really kills render times. Especially when using with ibl master or uberenvironment.

For UE, that will be ambient only mode. But really, with AWE Surface there's really no need to use either.

mgtskynet

wowie said:

mgtskynet said:

Hello guys, is it possible to turn of occlusion without also having to turn of lighting ?

It really kills render times. Especially when using with ibl master or uberenvironment.

For UE, that will be ambient only mode. But really, with AWE Surface there's really no need to use either.

Okay, thank you wowie, do you have any tips on how to use occlusion on opacity maps without it tanking render speeds so much ?

Right now i have to disable occlusion on pretty much any hair or plant that uses opacity maps.