In recent years, confocal Laser-Scanning microscopy became the most sophisticated microscopic technique for 3D-imaging in biomedical microstructure research. Experimental evidence, however, showed that under optical conditions relevant e.g. for nuclear genome analysis, the resolution of such an instrument as given by the Full Width at Half Maximum of the confocal point-spread function (PSF) is limited to about >= 0.3 micrometers laterally and to about >= 0.7 micrometers axially. A recently introduced light- microscopic approach, termed Spectral Precision Distance Microscopy (SPDM), allows the precise measurement of distances and angles between specifically labeled target sites far below the above mentioned resolution limit. SPDM is based on the use of 'point-like' objects labeled with different spectral signatures. Since in most cases, spectral signature differences have been realized by variation of excitation/fluorescence emission spectra, the calibration of chromatic aberrations is of utmost importance. Here, an improved procedure for the correction of chromatic shifts is presented. Statistical errors introduced by the localization accuracy can be minimized by the multiple measurement of the 3D-distances between the same specifically labeled target sites in a number of cases and subsequent averaging. However, the thus obtained mean distance, the estimate for the 'true' distance may be biased and therefore limited by the localization accuracy. Virtual microscopy simulations of test objects using an experimentally obtained PSF, showed that a few thousand detected fluorescence photons are sufficient to measure reliably distances down to about 20 nm, if other sources of error apart from voxelization, digitization and photon noise are negligible.