Flow velocities measured in the near-wake of bluff bodies contain information of mean flow, turbulent fluctuations and vortex motions. Flow parameters like RMS values of velocity fluctuations, Reynolds shear stresses, and auto- / spatial correlation of velocity fluctuations, derived by direct averaging of measured data, will include the total contributions of the periodic vortex motions and the random turbulent fluctuations. In the case of near-wake flows behind a cylinder at sub- critical Reynolds numbers, a vortex-street is formed in the wake and the vortex-shedding frequency is well defined. This offers a possibility for simplifying the decomposition or regular vortex motions and the random turbulent fluctuations. Velocity profiles and cross-stream spatial correlation of streamwise velocity were measured by positioning a TSI's two-component LDA system in the region 2d to 4d downstream of a circular cylinder at Re equals 35,000. The randomly sampled data were first interpolated linearly, and then resampled with a pertinent sampling frequency. The optimal FIR filters, designed using the Remez exchange algorithm, were applied to reject the digital signals in a narrow band around the vortex-shedding frequency. Various flow parameters associated with random turbulence were computed. The integration of the auto- and spatial correlation, obtained based on the filtered data gave a physically adequate estimation of the integral time and length scales of the turbulent fluctuations. Errors due to linear interpolation and filter were discussed.