The minimum and maximum sample rate Turtle should use.
Turtle uses an adaptive sampling scheme that can perform both undersampling (less than one sample per pixel) and oversampling (multiple samples per pixel). A sampling rate from one sample per 256 pixels and up to 256 samples per pixel can be used.
The Min Sample Rate sets the initial resolution, where a negative value means a lower resolution than the original image (in powers of two). For example, -4 gives the original resolution divided by 16. The default rate is 1.
The Max Sample Rate sets the resolution of the final prepass, with 0 giving the same resolution as the original resolution. These controls should normally be left at default values -4/0, however you can lower them to render very fast GI preview renderings. The default rate is 2.
If the contrast differs less than this threshold Turtle will consider the sampling good enough.
Select from the filter kernels Box, Triangle, Cubic, Gauss (default), Catmull-Rom, Lanczos, Mitchell.
The width of the filter kernel in pixels. Range from 1.0 to 3.0. The default width is 3.
The height of the filter kernel in pixels. Range from 1.0 to 3.0. The default height is 3.
The maximum number of 'bounces' a ray can take before being truncated. A bounce can be a reflection or refraction.
The maximum number of reflections a ray can take before being truncated. This limits the value set in each material.
The maximum number of refractions a ray can take before being truncated. This limits the value set in each material.
Controls at what depth a ray can spawn shadow rays. If set to 1, only primary rays will spawn shadow rays. If set to 2, the first bounce will spawn shadow rays as well. This limits the value set in each material.
The maximum number of transparent surfaces a GI ray can go through, before being truncated. This is used by all gathering light integrators, like Final Gather and Monte Carlo.
An error threshold to avoid self intersections. For example, a shadow ray should not intersect the same triangle as the primary ray did, but because of limited numerical precision, this can happen. The bias value moves the intersection point to eliminate this problem. If set to zero this value is computed automatically depending on the scene size.