Tensile tests of miniature spruce wood specimens have been performed to investigate the damage evolution in wood at the microscopic scale. For this purpose, the samples were stepwise tensile loaded in the longitudinal (L) and radial (R) directions and the damage evolution was monitored in real-time by acoustic emission (AE) and synchrotron radiation micro-computed tomography (SRµCT). This combination is of outstanding benefit as SRµCT monitoring provides an insight on the crack evolution and the final fracture at microscopic scale, whereas AE permits the detection of the associated accumulation and interaction of single damage events on all length scales with high time resolution. A significant drawback of the AE testing of wood has been overcome by means of calibrating the AE amplitudes with the underlying crack length development. Thus, a setup-dependent and wood species-dependent calibration value was estimated, which associates 1 µm2 crack area generating of 0.0038 mV in the detected AE amplitude. Furthermore, for both L and R specimens, AE signals were classified into two clusters by using a frequency-based approach of unsupervised pattern recognition. The shares of AE signals of both clusters correlate with the ratio of the relative crack area of the interwall and transwall cracks gained from the fractographic analysis of SRµCT scans.