Optical coherence tomography (OCT) is a depth-resolved imaging modality that uses low coherence interferometry to gate the time-of-flight of backreflected sample light. The heterodyne frequency of the interferometric signal can be up- or downshifted by reflection off of a moving particle (e.g. red blood cells in an artery). This Doppler shift can be resolved with a time-frequency decomposition (TFD) and displayed as a color Doppler-OCT image. Here we compare the in vitro and in vivo performance of three TFDs: the short-time Fourier transform, the Morlet wavelet transform, and the short-time MUSIC transform (STMT). The STMT is a new TFD that incorporates the MUSIC eigenfrequency estimator in a generalized short-time framework. The Morlet transform excels at identifying blood vessels, while the STMT is the most accurate predictor of Doppler shift frequency.