When collecting time series data of biological transport processes, it is necessary to observe the system at discrete time points, for example via an imaging experiment. This can introduce errors when the motion is approximated with discrete steps. We study the impact of collecting data at different temporal resolutions on parameter inference for biological transport models. In this work, we have performed exact inference for velocity jump process models in a Bayesian framework. This allows us to obtain estimates of the turning rate and noise amplitude for noisy observations of this transport process. We show sensitivity of these estimates to changes in time discretisation and noise amplitude. For a fixed photon budget, our results suggest that better estimates of parameters can be obtained when imaging more frequently with more noise than imaging sparsely with low noise.