We have recently proposed and demonstrated an approach that enables the acquisition of 2D nuclear magnetic resonance (NMR) spectra within a single scan. The approach is based on spatially encoding the spins' evolution along the indirect domain with the aid of a magnetic field gradient, and subsequently decoding this information numerous times over the course of the signal acquisition while spins are subject to a train of gradient echoes. The present paper discusses further considerations pertaining the 2D line shapes arising from this new way of collecting NMR data. Specific issues that are hereby addressed include (i) the effects introduced by fast relaxation onto the spatial encoding process, particularly the line widths and line shapes that will then arise in the frequency domain; (ii) approaches capable of correcting for the mixed-phase kernels resulting in these fast-relaxation cases, corresponding in essence to spatially encoded analogs of the TPPI and hypercomplex time-domain acquisition procedures; (iii) the enveloping characteristics imposed by the use of discrete excitation pulses on the attainable spectral widths along the indirect domain; and (iv) an analysis of the signal-to-noise characteristics of the methodology, with experimental corroborations of theoretical predictions and an illustration of the method's capabilities to analyze protein solutions in the mM-range concentration.