Abstract

In this paper we propose a new method to reduce phase vocoder artifacts during attack transients. In contrast to all existing algorithms the new approach is not based on fixing the time dilation parameter to one during transient segments and works locally in frequency such that stationary parts of the signal will not be affected. For transient detection we propose a new algorithm that is especially adapted for phase vocoder applications because its detection criterion has a direct connection to the phase spectrum and estimates the quality of the transformed signal. The evaluation of the transient detection shows superior performance compared to a previously published algorithm. Attack transients in sound signals transformed with the new algorithm provide very high quality even if strong dilation is applied to polyphonic signals.

Citations