• The Max Depth is the maximum number of bounces allowed for a path. A very small value speeds up the rendering but will introduce a lot of bias.

Default value: 1024. 

Accepted values: [1; 1024].  

• The Sampler defines the algorithm generating the random numbers that are used to sample the image plane and the scattering direction at each step of the path. As solving the light transport equation via  Path Tracing is a stochastic process, generating random numbers that are low-discrepancy (i.e. well-distributed in space and not clumped too closely together) can be very useful to reduce noise but can also produce patterns and aliasing. 

You can choose among three Samplers :

• Halton : Based on low-discrepancy sequence of the same name, the Halton sampler is the best compromise between rate of convergence and noise patterns. 

• Sobol : Using Sobol matrices, the Sobol sampler generates the low-discrepancy sequence with the faster rate of convergence but is prone to aliasing and noise patterns.

Both Halton and Sobol samplers are deterministic, which means that outputs are reproductible.

• Random : The Random sampler draws from a uniform distribution and offers the slowest convergence rate of the three but is immune to noise patterns and aliasing.