Sound field reconstruction methods enable the prediction of acoustic field quantities based on a limited number of measurements, including locations where direct measurements are unavailable. Their predictive capabilities over large spatial domains make them a promising tool for navigable reproduction and ‘6-degrees-of-freedom’ applications. Despite their potential, the methods are bandlimited due to spatial sampling, and the predictions can deteriorate far away from the measurement positions. In this study, we quantitatively examine the frequency and spatial limitations of sound field reconstruction methods using psychoacoustic metrics, including binaural auditory models to assess perceived source localisation and spectral colouration. The evaluated reconstruction methods include different wave-based approaches, such as plane waves and point sources. The analysis is conducted using an experimental dataset of spatial room impulse responses, measured sequentially along a line of high spatial resolution in a variable-acoustic room. The findings provide insights into sound field reconstruction techniques for navigable reproduction and the extent to which physical models contribute to practical, navigable auditory experiences.