Anti Aliasing Filter Design for Multi Nonlinearity Systems
Aliasing is a fundamental problem in nonlinear signal processing, particularly when memory less nonlinearities are simulated in discrete time. A conventional remedy is to operate at an oversampled rate. In the digital signal processing signal undergoes a nonlinear operation; its bandwidth is expanded, leading to a spurious mirroring of components back to the baseband. A new aliasing reduction method is proposed here for discrete-time memory less nonlinearities, which is suitable for operation at reduced oversampling rates. Aliasing is particularly problematic in audio applications. The method employs higher order anti derivatives of the nonlinear function used. The first order form of the new method is equivalent to a technique proposed recently by Parker. Higher order extensions offer a considerable improvement over the first anti derivative method, in terms of the signal-to-noise ratio. A commonly used method to reduce aliasing in memory less nonlinearities is oversampling, in audio applications, the input signal is typically unsampled by a factor 8 or 16 using an appropriate interpolation factor. The main disadvantage of oversampling is the proportional increase in the operation count. Interpolation/Decimation filters add to the workload per sample. A major new consideration will be the determination of numerical stability conditions for such antialiasing methods and will form the basis for future investigations. Keywords - Aliasing, Harmonic Distortion, Nonlinear Systems, Signal Denoising, Signal Processing Algorithms.