**7. Signal filtering**

Filters can be used to remove frequency components that do not belong to the signal or components that are irrelevant for a given analysis.

The captured signal can be filtered by hardware or software. Signal-filtering hardware can be used in the amplification step, while signal filtering by means of software can be performed during processing.

When using surface electrodes to measure EMG signals, interference from various sources can be mixed with the EMG signal. Each type of interference has its own characteristics that must be understood in order to remove it during the measurement phase or during processing. The useful information in the sEMG signal, which is a sum of the waves of varying frequency, is located between 20 and 500 Hz [12]. The signal is reduced due to the filtering effect of tissue located between the muscle fibers and the active sensing surface. The band pass filter corresponds to the frequency between the low frequency (high pass) and high frequency (low pass) cut-offs. Specific frequencies can also be filtered out with what are called "notch filters" [5, 11, 3, 23].
