The choice of pinhole geometry is a critical factor in the performance of pinhole-collimator-based source tracking systems for brachytherapy QA. In this work, an analytical model describing the penetrative sensitivity of a single-cone pinhole collimator to photons emitted from a point source is derived. Using existing models for single-cone resolution and double-cone sensitivity and resolution, the theoretical sensitivity and resolution of the single-cone collimator are quantitatively compared with those of a double-cone collimator with an equivalent field of view. Monte Carlo simulations of the single and double-cone pinhole collimators using an accurate 3D model of a commercial high dose rate brachytherapy source are performed to evaluate the relative performance of each geometry for a novel real-time HDR brachytherapy QA system, HDR BrachyView. The theoretical penetrative sensitivity of the single-cone pinhole is shown to be higher than the double-cone pinhole, which is in agreement with the results from the Monte Carlo simulations. The wider pinhole response function of the single-cone collimator results in a larger total error between the projected center of the source and the estimated center of mass of the source projection for the single-cone collimator, with the greatest error (at the maximum FoV angle) being 0.54 mm for the double-cone pinhole and 1.37 mm for the single-cone at θ = 60°. The double-cone pinhole geometry is determined to be the most appropriate choice for the pinhole collimator in the HDR BrachyView probe.