The detection of electromagnetic radiation (EMR) during the fracture of solid materials such as rocks and coal has been widely used in seismic exploration and mine dynamic disaster prediction. As described in this paper, we conduct uniaxial compression tests on coal, cement, and glass materials to determine the characteristic EMR differences among materials. A band-stop filter based on the Fourier transform and the wavelet packet transform method are used to conduct signal denoising and analysis. Basic analyses of the pulse-time characteristics, energy distribution, cumulative energy, and waveform characteristics of EMR are conducted. The research results show that there is a strong corresponding relation between the loading time, loading stress and EMR energy. A large number of EMR events are released in rapid succession during the main rupture of coal and cement, while the EMR events are evenly distributed throughout the whole loading process of glass. For the same material, the maximum EMR amplitude increases with an increasing peak value of the stress. The EMR pulse waveform of coal and cement agrees well with predictions based on the theoretical formula of the electromagnetic dipole oscillation EMR generation mechanism. The paper provides further theoretical basis for understanding the mechanism of EMR, with great significance for improving coal mining safety.