A Hybrid MFWT Technique for Denoising Audio Signals

Any unwanted background sound is considered noise. The perception of a sound at the same decibel level is determined by whether it is considered pleasant music or irritating noise. The removal of such noise is necessary to increase the quality of the audio signal as it travels through the communication system.A microphone is used in this study to collect real-time audio signals. A real audio signal is collected and combined with salt, pepper and white Gaussian noise for this experimental study. To remove disturbances from an audio signal, this study integrates linear and nonlinear spatial filtering mechanisms such as the median low pass filter, finite impulse response low pass filter, and wavelet transform approach. The suggested hybrid filter technique is compared to individual filtering methods, and its performance is measured using fundamental metrics such as mean square error and signal-to-noise ratio. When compared to other individual nonlinear filtering approaches, the hybrid method has a high signal to noise ratio (SNR - 88.04%) and a lower mean square error (MSE - 0.003).