In this report we demonstrate that strategically choosing sampling points with an intelligent use of adaptive blue noise sampling methods can drastically reduce the computation time required in the rendering process. We explore the state of the art in blue noise sample generation and explore new ways it can be used within the rendering process. Monte Carlo ray tracing is a vastly adopted image syntheses technique used ubiquitously across commercial and academic applications. It is capable of creating very high fidelity physically plausible images with computation that it unrestricted by dimensionality unlike most analytical approaches. Although capable of producing a high quality images Monte Carlo ray tracing often still needs to compute millions, if not billions of samples to produce a fully converged, noise free image. Tracing these samples comes at a cost and can lead to large computation times. Although we can reduce the cost of tracing rays with more optimized acceleration structures or naively throwing more hardware at the problem these are overshadowed by the improved quality gained via improved strategic sampling. Strategic sample placement has been proved to improve convergence rate of Monte Carlo ray tracing requiring fewer samples, and therefore decrease computation required to produce comparable results in quality. We explore the current literature on sampling methodologies and compare their implementation, performance and limitations and show that their sampling quality is inferior to adaptive blue noise. We will focus on applying the use of adaptive blue noise sampling within four dimensions of the rendering pipeline specifically. Firstly, we present a technique for generating primary ray samples that adaptively samples the image plane. We use a blue noise algorithm that adapts based on pre-existing information about the scene to increase the sampling frequency within areas of interest. Secondly we look at filter importance sampling, a technique that seems to be becoming ever more popular in rendering, and how we can use adaptive blue noise to generate higher quality sampling distributions than what is possible with the currently used methods of importance sampling. Next we will explore importance sampling BxDFs by generating samples over the hemi- sphere. Finally we will conclude with a brief discussion on some future ideas about direct light sampling of arbitrarily defined mesh lights. A challenge that is faced by all professional rendering software as efficient sampling methodologies are not well defined. Although not all our results in this report are entirely conclusive we believe that we have brought attention and provided promising results that build the foundations to an under researched area of knowledge that could help solve practical rendering problems faced by professional graphics.