We explore the use of a processing procedure based on restricted least square minimization as a tool for reducing the time versus resolution dilemma often encountered for biomolecular multidimensional spectra. Using a 2D spectrum as a reference, we obtain the necessary input of frequency components and linewidths. Combined even with a limited time evolution in the indirect dimension, the amplitudes of the correlation peaks in all planes of the 3D spectra can be extracted, and can be used to reconstruct the interferograms in the third dimension. Parameters such as number of lines, threshold choice, resolution, lineshape, number of experimental data points and finally signal to noise ratio of the spectrum are examined starting from a triple-resonance HNCA spectrum of ubiquitin.