This paper presents some results of a larger study about vision geometry, carried out between 2014 and 2016 at the Lisbon School of Architecture, considering eye-tracking technology as an effective way of studying human vision. The methodology relied on portable eye-tracking equipment to analyze 3D immersive visual perception. 30 observers conducted 120 analyses of four different three-dimensional architectural spaces. The 120 samples allowed the analysis of 60,000 video frames to understand different kinds of elements that can be used to describe and study visual information. Quantitative and qualitative results are presented in the form of graphics to better understand eye movements in the perception of three-dimensional visual spaces, with a particular focus on macro-saccadic movements. The purpose of this work is to examine how the technology works, the possibilities it offers, and its limitations. The paper presents results in a way that seems trustworthy to work with at present, in order to obtain insights of scientific value that can point towards future possible resolutions of current methodological and technological issues.