Objective to develop a new strategy combining near-infrared (NIR) and dielectric spectroscopies for real-time monitoring and in-depth characterizing populations of Chinese hamster ovary cells throughout cultures performed in bioreactors. Results Spectral data processing was based on off-line analyses of the cells, including trypan blue exclusion method, and lactate dehydrogenase activity (LDH). Viable cell density showed a linear correlation with permittivity up to 6 x 10(6) cells ml(-1), while a logarithmic correlation was found between non-lysed dead cell density and conductivity up to 10(7) cells ml(-1). Additionally, partial least square technique was used to develop a calibration model of the supernatant LDH activity based on online NIR spectra with a RMSEC of 55 U l(-1). Considering the LDH content of viable cells measured to be 110 U per 10(9) cells, the lysed dead cell density could be then estimated. These calibration models provided real-time prediction accuracy (R-2 >= 0.95) for the three types of cell populations. Conclusion The high potential of a dual spectroscopy strategy to enhance the online bioprocesses characterization is demonstrated since it allows the simultaneous determination of viable, dead and lysed cell populations in real time.