The trade-off between temporal and spectral resolution in conventional pulsed wave (PW) Doppler may limit duplex/triplex quality and the depiction of rapid flow events. It is therefore desirable to reduce the required observation window (OW) of the Doppler signal while preserving the frequency resolution. This work investigates how the required observation time can be reduced by adaptive spectral estimation utilizing 2-D spatial information obtained by parallel receive beamforming. Four adaptive estimation techniques were investigated, the power spectral Capon (PSC) method, the amplitude and phase estimation (APES) technique, multiple signal classification (MUSIC), and a projection-based version of the Capon technique. By averaging radially and laterally, the required covariance matrix could successfully be estimated without temporal averaging. Useful PW spectra of high resolution and contrast could be generated from ensembles corresponding to those used in color flow imaging (CFI; OW = 10). For a given OW, the frequency resolution could be increased compared with the Welch approach, in cases in which the transit time was higher or comparable to the observation time. In such cases, using short or long pulses with unfocused or focused transmit, an increase in temporal resolution of up to 4 to 6 times could be obtained in in vivo examples. It was further shown that by using adaptive signal processing, velocity spectra may be generated without high-pass filtering the Doppler signal. With the proposed approach, spectra retrospectively calculated from CFI may become useful for unfocused as well as focused imaging. This application may provide new clinical information by inspection of velocity spectra simultaneously from several spatial locations.