Skin

Table Of Contents

• Introduction
• Scattering Layers
• Common
• Shallow / Mid / Deep Scatter Layers
• Single Scattering
• Reflections
• Primary / Secondary Reflection
• General
• Optimization
• Bump Mapping
• Skin Example

Introduction

This shader simulates the scattering effect of skin. In real life, skin is made up of three distinct layers:

• Epidermis - defines the skin tone and waterproof layer
• Dermis - contains tough connective tissue
• Subcutis (hypodermis) - made up of loose connective tissue, fat and blood vessels

To model this, the Redshift skin shader has three equivalent translucent, sub-surface scattering layers:

• Shallow Scatter - defines the pigment/skin tone of the skin
• Mid Scatter - defines the dermis layer
• Deep Scatter - defines the thick, subcutis layer

Redshift also simulates the 'waterproof' oily property of the top-most epidermis layer by having a 'Primary Reflection' control. A 'Secondary Reflection' layer is also available for added sheen from, for example, oily cosmetic products, or for clear-coat effects.

Free head model, courtesy of 3D Scan Store, with a textured skin shader material, illuminated by two area lights and a dome light

You can see how sub-surface scattering of light affects thin parts of the model (the ear) versus the thick part of the model (the head).

Scattering Layers

Each layer uses sub-surface scattering to compute a diffuse lighting result and the results are blended together based on each layer's individual weight.

The Shallow Scatter layer is the thinnest, so it should have the smallest Radius, yielding an almost diffuse scattering response. The Deep Scatter layer is the thickest layer, so it should have the largest radius, adding the blood tone under the skin.

Common

Radius Scale

Allows you to adjust the strength of the scattering effect for all the layers. Lower values means scattered light is absorbed quicker, yielding a more diffuse look. Conversely, higher values yield a more translucent look.

Overall Scale

Allows you to adjust the overall brightness of all the scattering layers.

Diffuse Amount

Controls how much of the SSS effect you get. A value of 0.0 means that you'll only get subsurface scattering. On the other hand, a value of 1.0 will make this surface behave as a perfectly diffuse surface without any SSS. This parameter is useful for materials that need a weaker SSS effect.

Disable Diffuse Bump Mapping

This option allows you to switch to using the default un-bumped surface normals for diffuse lighting calculations.

In the examples images below note how when this option is enabled the overall SSS effect is smoothed out since the geometry is much smoother than when the diffuse bump map is used when disabled. 

Normalize Diffuse Weights

In order to be physically correct, by default Redshift shaders conserve energy by not emitting more light than they receive. When this parameter is checked, we apply physical correctness by ensuring the sum of the weights of each layer do not exceed 1.0.

When this option is un-checked we leave the weights alone, which can result in brighter, but not physically correct results.

Mode

• Point-Based
• Point-Based (Legacy)
• Ray-Traced

You can use either point-based or ray-traced multiple sub-surface scattering in Redshift and both have their own advantages and disadvantages. In quite a few cases the techniques should look similar to each other but there are situations where the differences will be readily apparent, this is primarily due to the normalization of light that occurs when using point-based mode. Ray-traced mode does not normalize the light which can lead to differences that are most notable on thin objects or objects with more surface detail.

Point-Based

• Faster and smoother

• Less detailed / accurate

• Does not work in progressive mode

• Requires a “prepass” stage

• Higher chance of flickering in difficult lighting situations.

• Not possible to isolate SSS effect on a particular object which can result in unnecessary “light bleeding” artifacts.

Ray-Traced

• Slower and noisier

• More detailed / accurate

• Works in progressive mode

• The higher the scatter radius the more samples are needed for clean results.

• Possible to isolate SSS effect between objects or have it affect all objects. Please see here for more info.

Samples

This controls the number of samples cast for Multiple Sub-Surface Scattering rays. Higher numbers will reduce noise issues but will take longer to render.

In the images below the sample difference is most noticable in the ear. 

Include Mode

• All Objects : All other objects participate in the SSS effect.
• Only Self : Contain the SSS effect in the same object only.

Notice the difference between the different methods around the collar and neck in the images below. To make the difference more prominent a Redshift Skin shader is applied to the jacket part of the model and the Radius Scale for both the head and jacket shaders has been increased to 10. 

Shallow / Mid / Deep Scatter Layers

Color

This is the color of the scatter layer.

Weight

This is a multiplier of the scatter layer color and can be used to adjust the brightness / strength of the layer.

This value is multiplied by the 'Overall Scale' parameter.

Radius

This parameter is scene-scale related. It controls how deep light can travel within this surface. If the value is high, light can go deeper which creates a softer effect. Smaller values, on the other hand, mean that the light reaches extinction early which creates a harder more diffuse effect.

This value is multiplied by the 'Radius Scale' parameter.

Single Scattering

Transmittance

Describes the sub-surface attenuation color. A color of black means full light absorption, yielding a solid material, while a color of white means light rays will pass right through without being absorbed. Dark colors can be used to describe denser materials.

Thickness

This acts as a scale of the absorption effect, with a value of 0.0 meaning no absorption and larger values meaning stronger absorption and thus denser-looking materials.

Scatter Amount

Acts as a scale of the scattering effect, with a value of 0.0 meaning no scattering and larger values meaning more scattering and thus denser-looking materials.

Samples

This is the maximum number of samples shot for single scattering rays. More samples will mean less noisy looking single scattering, but can affect rendering performance.

Phase

This describes the anisotropy of the scattering effect, for forward and back scattering. Forward scattering requires values between 0 and 1, which means rays will scatter away from the light source. Back scattering requires values between -1 and 0 which means rays will scatter back towards the light source.

Scattering Options in Detail

The 'Overall Scale' allows you to adjust the brightness of all the layers together, so you can quickly fine tune the intensity your diffuse lighting for the overall material.

The 'Radius Scale' allows you to adjust the strength of the scattering effect for all the layers. Since sub-surface scattering is heavily dependent on the scale of the object, you can use this option to quickly adjust the scattering effect until it is visually pleasing. Smaller values mean quicker absorption of light, resulting in less scattering effect and a more diffuse look. With larger values you get a more translucent look.

For physical correctness, the 'Normalize Diffuse Weights' should be left enabled. This will ensure the material does not reflect more light than it receives. However, if you want better control over the overall layers brightness or want your material to be more emissive than physically correct, you can disable this option.

Below is a side-by-side example showing the difference in lighting with the 'Normalize' option disabled, using the default layer weights of the material.

The 'Diffuse Amount' lets you control how much of the overall sub-surface scatter effect you want, versus pure diffuse lighting. When the 'Diffuse Amount' is set to 1.0, there will be no sub-surface scattering. Equally, when set to 0.0, there will be only sub-surface scattering.

Now let's take a look at how each layer can contribute to the scattering effect. While the differences can be subtle at times, they are noticeable.

Reflections

Primary / Secondary Reflection

Color

This defines the color of the reflection.

Weight

This scales the overall amount of the reflection.

BRDF

Allows you to choose the technique for simulating physically correct rough reflections. Beckmann is generally considered the standard for a good range of materials, but GGX has a wide specular tail, which is good for chrome materials. 

• Beckmann (Cook-Torrance)
• GGX
• Ashikhmin-Shirley

Glossiness

This controls the glossiness, or roughness, of the reflection. A glossiness of 1 means almost mirror reflection, while a glossiness of 0 means very rough, almost diffuse reflection.

Samples

This defines the number of samples to use for the reflection. More samples means less noise.

IOR

The index of refraction which is used to compute the view-dependent Fresnel reflectivity effect. Reflections with an IOR value closer to 1.0 will have less facing reflectivity. The recommended default value is 1.44.

General

Disable Reflection Bump Mapping

This option allows you to switch to using the default un-bumped surface normals for direct and indirect reflection rays.

Optimization

Specular Highlights Only

This optimization option will disable indirect reflection ray tracing of the environment and just capture specular reflections of lights.

Enable Trace Depth Overrides

This enables the Max Reflection Trace Depth parameter. When not enabled, the global trace depth will be used.

Max Trace Depth

This parameter controls the trace depth for reflection rays shot from this material. If the reflections of a material are not very defined (meaning: they are blurry because of the glossiness parameter or have a low weight), then multiple reflection bounces can be a waste of rendering time. This parameter allows the user to reduce the trace depth and speed up rendering.

Cut-off Override Enable

This parameter allows you to override the global cut-off setting for reflections.

Cut-off Override Threshold

When the reflections of a material are very dark (because of low "Weight" or "Color" values) they contribute very little to the final image. This parameter defines what is considered "very dark" at which point no more reflection rays will be shot - which will speed up rendering. Scenes containing very strong lights might need this parameter set to very low values such as 0.0001 in order to avoid early termination of tracing which can produce a grain-like effect.

Reflection Settings in Detail

Below are some examples of how different reflection settings can affect the look of the material.

Bump Mapping

Add Bump Input To Material Bump

Skin Example

There are various tutorials on the web for different renderers that describe how to use texture maps for the sub-surface scattering layers. The tutorials that describe how to map three sub-surface layers also apply reasonably well to the Redshift skin material.

Although the best results can be gotten using dedicated textures for the layers and reflections, here we show a simplified example, using a single diffuse texture that can be manipulated to add color detail to the shader, while still maintaining the sub-surface effect.

The first step is to de-saturate the texture, then adjust its contrast to darken the dark details (such as facial hair). Once you have that texture you apply a red tint for the deep scatter color, a green-ish tint for the mid scatter color (so the sum of which becomes a deep orange) and simply use the diffuse texture for the shallow scatter layer. You can also use the de-saturated texture for the reflection color.

Below is an example showing the different layers, with its appropriate weight, and the other layers having a weight of 0.0.

Below now shows the layers successively summed together, by setting the layer weights to their appropriate values.