There have been interesting strides made in the frontier of path tracing solutions for global illumination of volumes. Within global illumination of participating media, it is of- ten preferable to stochasticaly sample the volume and how a photon would interact within it in order to generate unbiased and performance efficient renderers. The sampling is done by transmittance estimators. This work examines the performance difference of several state of the art transmittance estimation algorithms. The new Poisson trackers are based on sampling the participating media directly according to a Poisson probability distribu- tion. This is afforded by a theoretic derivation that rephrases transmittance estimation as sampling the terms of the optical depth’s Taylor expansion representation. This derivation also left room to theorize about potential estimators using alternative distributions than Poisson. The Poisson class of estimators are performance tested along with the contem- porary methods to determine scenarios in which one method would be more performant over an other. To do this, the software Inviwo was extended to support compute shader. This extension allows for more general deployment of the GPU not tied to rendering. The potential for new transmittance methods built around the geometric distributions is also discussed. This work confirms that the new Poisson Estimators, especially the one based on Residual-Ratio Tracking, outperform all other estimators in the general case. The discussed geometric transmittance estimators is shown to require a tricky optimization equation to be solved.