This paper introduces a practical precoding method for the downlink of Filter Bank Multicarrier-based (FBMC-based) massive multiple-input multiple-output (MIMO) systems. The proposed method comprises a two-stage precoder, consisting of a fractionally spaced prefilter (FSP) per subcarrier to equalize the channel across each subcarrier band. This is followed by a conventional precoder that concentrates the signals of different users at their spatial locations, ensuring each user receives only the intended information. In practical scenarios, a perfect channel reciprocity may not hold due to radio chain mismatches in the uplink and downlink. Moreover, the channel state information (CSI) may not be perfectly known at the base station. To address these issues, we theoretically analyze the performance of the proposed precoder in presence of imperfect CSI and channel reciprocity calibration errors. Our investigation covers both co-located (cell-based) and cell-free massive MIMO cases. In the cell-free massive MIMO setup, we propose an access point selection method based on the received SINRs of different users in the uplink. Finally, we conduct numerical evaluations to assess the performance of the proposed precoder. Our results demonstrate the excellent performance of the proposed precoder when compared with the orthogonal frequency division multiplexing (OFDM) method as a benchmark.