In this paper we consider the alternative way of development of spatial beamforming processing algorithm in time-domain under low-rank Massive-MIMO systems. It is well-known, that the telecommunication systems, which based on spatial multiplexing technology achieve following advantages: increase the spectral efficiency in many times due to spatial channel multiplexing gain, the beamforming technology implementation (effective energy distribution in the space). The organization of spatial filtering (the beamforming vector processing) – one of the most significant tools in MIMO-OFDM multiplexing. In this paper, we consider a new spatial filtering processing approach (precoding vectors calculation), which is a significant tool in large-scale antenna systems, due to the fact of smart processing of the radio narrow-beam implementation increases the efficiency of the communication system many times over. Nowadays, the spatial filter design algorithms with low computational complexity have attracted considerable attention from researchers seeking to improve existing MIMO-OFDM wireless communication systems. In the paper we developed basic algorithms for calculating precoding vectors in the frequency domain and proposed the time-domain precoder approach with aim to achieve the complexity reduction gain. The proposed method was tested on the geometric statistical simulation channel model QuAdRiGa according to the multipath channel scenario "3GPP-3D Uma-NLoS". We provide the assessment of whole presented methods and the evaluation range of discrepancies was revealed in the form of pictures of the correlation and integral metrics. The analysis of integral cumulative distribution function of simulation results shows the relation gap between accuracy of correlation metric and losses of beamforming processing recalculation approximately -0.5 dB for 1st layer. In conclusion of paper was revealed the fact that proposed SVD factorization for each impulse response did not take into account the beams and eigenvectors distribution over the whole frequency bandwidth