Surveying disease vectors is currently excessively laborious for continuous and widespread monitoring. Wing beat modulation spectroscopy gives opportunity for species and sex recognition in electronic traps or mosquito target classification in lidar. We used a polarimetric dual-wavelength-band laboratory system to record kHz modulated backscattered light from insects. The system operates in the near and short-wave infrared at 808 nm and 1550 nm and retrieves both co- and depolarized light. Here we give clues on the harmonic content and covariance of four mosquito species and fruit flies. Further, we interpret the interdependence of harmonic strengths when insects transit the probe volume with random heading direction and provide correlation matrices for coherent and incoherent light. Using the obtained parameters, we demonstrate that species that are difficult to distinguish with microscope can be classified with high accuracy. The results are valuable for understanding wingbeat harmonics in relation to heading and valuable for optimal sensor design for disease vector surveillance.