hborre posted at 12:10 PM Tue, 18 October 2022 - #4446928

That looks pretty good. I have to agree, the lacrimal is a difficult area to get right, and many texture maps include specular highlights that interfere with realistic lighting.  I'd be very interested in how you set up the Material Room nodes.

sure

I'm glad to see that the Blinn node has no objections to being used in Cycles.  Typically,  the Translucent node is used with the Add closure node rather than the Mix closure; have you tried that arrangement?  And thank you very much for sharing.  I look forward to trying this out and adding it to my repertoire.  

Wow... Thank you

hborre posted at 1:02 PM Tue, 18 October 2022 - #4446932

I'm glad to see that the Blinn node has no objections to being used in Cycles.  Typically,  the Translucent node is used with the Add closure node rather than the Mix closure; have you tried that arrangement?  And thank you very much for sharing.  I look forward to trying this out and adding it to my repertoire.  

I did not know that. I'll take a look

Honestly, with an AddClosure, insead of a MixClosure, I like very much

And I have a (probably subjective) feeling that adding PrincipledBsdf+TransluscentBsdf is slightly different than TransluscentBsdf+PrincipledBsdf

TBH they are glowing in my render using Add instead of Mix

ghostship2 posted at 2:29 PM Tue, 18 October 2022 - #4446940

TBH they are glowing in my render using Add instead of Mix

There was a time (before your fantastic hair shader, in fact), I used to use the translucent to make the hair shine with back lights and if I remember correctly it was too shiny if the bump effect wasn't enough

You do get a slight variation in the effect depending on how you wire the noodles.  In many of the Blender setups, the TranslucentBsdf node is wired to the top value of the Add shader node.  The way to mitigate the glow is to reduce the Surface value on the PrincipleBsdf to a value much lower than one.  My go-to value is approximately 0.01 to 0.02, I compromise at 0.015 most of the time anyway.  

I have come to the conclusion that point, spot, and infinite lights are no good for creating realistic renders. They don't "reflect" back into the scene from any shader in a realistic way. They don't have a "size" and are a compromise/workaround to get quick renders. I'll stick to area, mesh and background for lights from now on. Ok, here are a few renders of the lacrimal with various background lighting and one using Area lights.

If I may give my layman's critique on how the lights look in the eyes. In the first one, I see several small lights like maybe several small bulbs. The second looks like several casement windows. The third has a flat panel of light. About the only place that you might see is a photography studio. The fourth has a faint bar of light like maybe a dim fluorescent tube. The next looks like windows in the same place as the fluorescent light. the final one again has panel lights again like in a studio. All have lights reflected, or light from the environment

Point lights do reflect and have size. Here are point lights at 10%, 100%, and 1000%.

RedPhantom posted at 7:20 PM Tue, 18 October 2022 - #4446965

If I may give my layman's critique on how the lights look in the eyes. In the first one, I see several small lights like maybe several small bulbs. The second looks like several casement windows. The third has a flat panel of light. About the only place that you might see is a photography studio. The fourth has a faint bar of light like maybe a dim fluorescent tube. The next looks like windows in the same place as the fluorescent light. the final one again has panel lights again like in a studio. All have lights reflected, or light from the environment

yes, you are correct on all of those. They all use the background HDRI "object" to light the scene except for the last one that is using area lights. it's mostly about how the lacrimal looks not about the eye.

RedPhantom posted at 7:31 PM Tue, 18 October 2022 - #4446967

Point lights do reflect and have size. Here are point lights at 10%, 100%, and 1000%.

but they are not showing up in the camera so not really what you want. They are fine if placed into something like a lampshade but placed into a room naked, they don't work for me.

Anyone knows of any teeth fixes for LF and LH?

RedPhantom posted at 7:31 PM Tue, 18 October 2022 - #4446967

Point lights do reflect and have size. Here are point lights at 10%, 100%, and 1000%.

If you don't want a point light to appear in a reflection, reduce the scale to 1%.  At such a small pin scale, it is not visible.

hborre posted at 2:42 PM Wed, 9 November 2022 - #4448812

RedPhantom posted at 7:31 PM Tue, 18 October 2022 - #4446967

Point lights do reflect and have size. Here are point lights at 10%, 100%, and 1000%.

If you don't want a point light to appear in a reflection, reduce the scale to 1%.  At such a small pin scale, it is not visible.

I do that. I will even take it down to .01% at times. It was just a demonstration since Ghostship said they don't show up in reflections.

exzazxe posted at 2:02 PM Wed, 9 November 2022 - #4448809

Anyone knows of any teeth fixes for LF and LH?

What would you like to have changed?

Liquid_Ice posted at 12:06 PM Sat, 12 November 2022 - #4449022

exzazxe posted at 2:02 PM Wed, 9 November 2022 - #4448809

Anyone knows of any teeth fixes for LF and LH?

What would you like to have changed?

Thanks for responding, I suppose what I should have asked after were teeth or mouth internal morphs rather than 'fixes'. They seem to be very spaced widely apart and the scale transforms don't seem to affect the spaces, I find the tongue very hard to pose. 

I realise that the most likely solution is to export and model them and edit externally but it sure would be handy if someone already did that. 

BTW sorry if interrupted the previous conversation on lighting technique

I did that to some extend, the tongue is a problem, especially if you use the phonemes ( especially A) the tong often gets wonky. I also found out the teeth are skewed in the mouth. 

I just discovered a 'teethapart' morph for La Femme Pro .

With the tongue I first Z translate the tongue out of the mouth then apply any bend and twist before ztranslating it back into position, its easy to use the morph tool on tongue parts.

Seems I answered my own question

Yes but the inside mouth is still not great. It goes through the chin. And also the teeth from the 3/4th view is looking weird. I have fixed that but that can also raise issues with different morphs.

This is SAV's Monica Bellucci character for V4

One of the oldest Hair Figures I have (Radiant Jaguar, 2006), still great, thanks to Ghostship's Hair shader ❤

Top, dress and necklace are dynamic props.

@Y-Phil

That looks nice!

ghostship2 posted at 7:28 PM Tue, 22 November 2022 - #4450051

@Y-Phil

That looks nice!

Thank you Mister.
Oh and you've participated to Poser13's picture, if I understood correctly?

I'm just going to leave this here.

(Hint: this picture has a single spotlight. No HDRI. Learn your materials and render settings, everyone! :D )

Very good ! Except the lemon slices that are aligned and evenly dispatched, you should add some chaos so it will be perfect. and breadcrumbs on the table :)

Awesome, Afrodite-Ohki wow

lsauvage posted at 8:29 AM Tue, 29 November 2022 - #4450588

Very good ! Except the lemon slices that are aligned and evenly dispatched, you should add some chaos so it will be perfect. and breadcrumbs on the table :)

What breadcrumbs?

The render is beautiful, indeed.  It demonstrates that you don't always need to rely on HDRI illumination as a principal light source.  It should support the render lights, not compete with them when used.

lsauvage posted at 8:29 AM Tue, 29 November 2022 - #4450588

Very good ! Except the lemon slices that are aligned and evenly dispatched, you should add some chaos so it will be perfect. and breadcrumbs on the table :)

I used the models the way they came - just changed the materials and added some subdivision to some of them ;)

Thanks everyone!

@hborre: with this I learned that blown up lights with a good amount of bounces make up for good sunlight :) this spotlight is at 400% intensity, and the entire scene looks very dark (and fake) with lower render settings and no caustics:

These are experiments (and promos) for an upcoming script by wolfnom and I to help with render settings for Superfly :)

But isn't that exactly what HDRI is good for? Reflected light coming from all directions to all parts of the illuminated object. Also to places which are turned away from the light source. Exactly what a large number of bounces does. Only more effective, faster.

As far as I know, the basic idea was: shoot the surrounding room of a scene once in HDR and then shoot the objects in an empty room with an HDRI (High Dynamic Range Image) as "light source".

This way, only the actual object has to be rendered, not the surrounding area. And you can take the shots anywhere. You don't have to be at the place where the scene actually takes place (a modified technique is also used today in film productions: with LED screens as a light source around the object to be photographed).

About your image above: The render time for an equivalent shot using an HDRI as the light source should be a fraction on the same hardware. To get exactly the same result, however, you would need an HDR image of the same room you used when rendering with Bouces.

Complicating matters, however, is the problem that the Cyles implementation in Poser 11/12 doesn't handle HDRIs well. They quickly create fireflies depending on the image. Depending on the HDR material, you may not be able to get rid of them. No matter how many bounces you render with.

This is completely different in Blender. There it works perfectly and many times faster than in P11 and P12. So fast, in fact, that Cycles can be used as a full-fledged preview even with relatively low graphics card performance - just in those cases when Eevee isn't enough.

Rendering with many bounces instead of HDRIs is probably only good for Poser 11 and 12 (we will see how it is in P13). It should not be taken as a general recommendation!

@Ohki
Almost forgot to mention: Really well done render!

adp001 posted at 7:29 PM Wed, 30 November 2022 - #4450677

Rendering with many bounces instead of HDRIs is probably only good for Poser 11 and 12 (we will see how it is in P13). It should not be taken as a general recommendation!

What if I want to render in a scene for which I don't have an HDRI?

What if I want to move through a space instead of remaining at the center point of the HDRI?

HDRI's give good results but I find them pretty limiting.

ChromeStar posted at 8:55 PM Wed, 30 November 2022 - #4450679

adp001 posted at 7:29 PM Wed, 30 November 2022 - #4450677

Rendering with many bounces instead of HDRIs is probably only good for Poser 11 and 12 (we will see how it is in P13). It should not be taken as a general recommendation!

What if I want to render in a scene for which I don't have an HDRI?

What if I want to move through a space instead of remaining at the center point of the HDRI?

HDRI's give good results but I find them pretty limiting.

Try using HDRIs made just to lit a scene. Here is an Image showing "HDR Light Studio": https://www.lightmap.co.uk/dynamic_resources/webp_images/homepage/hdrls_ui.webp

Light up the scene with such a "grayscale image" and you save the heavy computing power for the bounces.

You can also take one of the typical HDRIs (e.g. from one showing a photo studio, available for free) and prevent the image from being displayed. Then only the lighting information remains to illuminate the scene.

I made some samples. For all images I made the standard Groundprop all black (so no indirect light from this one).

One Spotlight, 300%, no HDRI

Same Spot, HDRI intensity 0.05

Same Spot, HDRI intensity 0.10

HDRI Background

The "hsv2 value" is the intensity (mentioned above). Intensity depends on the image used. Saturation is reduced to zero, so the image has no colour information (b/w). For special effects one may let the colour information alone. Also worth checking out: False colours.

Remember that the ground prop should be all black. The background shouldn't bounce light (unless one wants to achieve special effects).

With this method you get fast renders with just one key light.

adp001 posted at 7:29 PM Wed, 30 November 2022 - #4450677

But isn't that exactly what HDRI is good for? Reflected light coming from all directions to all parts of the illuminated object. Also to places which are turned away from the light source. Exactly what a large number of bounces does. Only more effective, faster.

That depends - you'd have to find an HDRI with the exact lighting you want for the scene. And HDRI-lit scenes aren't exactly fast to render in Poser, they take forever to clear out the noise...

adp001 posted at 2:56 AM Thu, 1 December 2022 - #4450684

Try using HDRIs made just to lit a scene. Here is an Image showing "HDR Light Studio": https://www.lightmap.co.uk/dynamic_resources/webp_images/homepage/hdrls_ui.webp

Light up the scene with such a "grayscale image" and you save the heavy computing power for the bounces.

You can also take one of the typical HDRIs (e.g. from one showing a photo studio, available for free) and prevent the image from being displayed. Then only the lighting information remains to illuminate the scene.

Honestly? Amping up the bounces in my render a bit (and it wasn't even a lot, I think I had 8?) is much faster than rendering a scene as an HDRI and then using THAT as a light source.

"Complicating matters, however, is the problem that the Cyles implementation in Poser 11/12 doesn't handle HDRIs well. They quickly create fireflies depending on the image. Depending on the HDR material, you may not be able to get rid of them. No matter how many bounces you render with."

I have determined that low-resolution HDRis do create Firefly artifacts.  And I stress low-resolution images, 1 to 2k.  Replace the same HDRI scene at a higher resolution, 3 to 5k, and the Fireflies are resolved at the expense of memory.  Just using the luminance of an HDRI appears to also resolve Firefly artifacts, something to consider when using EZDome.  

This is what a Convoluted HDRI will look like when applied to the background:

This is just pure luminance from an HDRI.  Unfortunately, this needs to be created in a 3rd party application.  However, If you are only interested in providing lighting, and not an environmental background, this type of image will suffice.

Doesn't change the fact that you'd need a new hdri for every change in lighting you want to make. Increasing the render bounces didn't increase my render time too much, and allowed me to use Poser lights as I see fit.

That's a great render, Ohki. I'm also a fan of Poser lights and bounces; I hardly ever use HDRIs.

odf posted at 7:23 PM Thu, 1 December 2022 - #4450738

That's a great render, Ohki. I'm also a fan of Poser lights and bounces; I hardly ever use HDRIs.

Thanks!

I purchased Poser 12  on Nov.28 to take advantage of the Nvidia RTX 3060 in my new PC.  This 4k image was rendered using the Optix:RTX 3060 setting. The ladder truck is fully animatable. All parts move. Test animations of 600 frames took a little over five hours.  With the CPU ( i7 Generation 12-12700F ) it would take days and the cpu fan sounding like a Hoover vacuum cleaner would be too much to bear. lol          The cpu load was only 4% and the GPU card did all the heavy lifting like it was intended to do. Area lights were used for the side markers and headlights. The Roto-Ray ( the light on the front grille) had a combination of area lights and spotlights. It mimicked the real world fixture quite well especially for night scenes. Shown here it is in the off position since it's purpose is to "clear the road" when racing to a scene.

My dad had a '68 Charger when I was a kid. I loved that car.

ghostship2 posted at 8:08 PM Fri, 16 December 2022 - #4451925

My dad had a '68 Charger when I was a kid. I loved that car.

She looks great, indeed, especially when comparing to nowadays' ...... stuff

Superfly PBR + HDRI

Just to say I think its an interesting question how to represent the quality of "ivory" in either a texture or material way. Both have their challenges. I spend many hours trying to represent ivory as a texture but also how would it be done as a material, I think it is an interesting point of discussion - ie textures vs materials.

I'm assuming that by materials you are referring to procedural texturing as opposed to image-based.

Yes so materials are based on the physical properties of the object so it can work with any light set-up. After many years of working on that I've  not had much success so have moved onto pre baking the light into the textures. So I seperate materials as  physically accurate and textures as pre baked. Hope that makes sense. I'm not trying define the terms just interested what peoples thoughts are.

Where physical properties are concerned, prebaking light is a bad process especially if your environmental illumination varies in your render.  If there's a close match between texture and lighting then it's not noticeable, however, if you're like many users who are not satisfied with initial lighting conditions and change them often, you will see discrepancies in the renders.

you could try this