Fractionated intracardiac electrogram (EGM) record- ings contain information about the underlying pathophys- iology and are promising to guide clinical ablation pro- cedures. Lowpass filters (LPF) are commonly used to de- noise EGM recordings. An alternative is wavelet denois- ing which does not rely on a constant separation of sig- nal and noise in frequency space. Unipolar and bipolar EGM recordings of 2.5 s registered during a contact map- ping guided ablation procedure were exported from the Carto3® (Biosense Webster) mapping system. Different wavelet denoising algorithms are evaluated. The success in reducing noise while retaining the physiologically rel- evant signal morphology of each algorithm is assessed. For this purpose recordings with a prominent signal are selected such that a signal dominated interval and an iso- electric interval with dominant noise can be identified. The standard deviation of the deviation from baseline in the isoelectric interval are calculated to quantify noise. The change due to denoising in peak counts of bipolar voltages and the negative derivative of unipolar voltages during sig- nal dominated intervals is calculated to quantify the stabil- ity of physiologically relevant morphology. The han5.5 wavelet with BlockJS thresholding and decomposition level 8 is identified as a successful choice for a denoising algorithm.