Constant Q Transform for Automotive NVH Signal Analysis 2010-36-0373

The constant Q transform consists of a geometrically spaced filter bank, which is close to the wavelet transform due to the feature of its increasing time resolution for high frequencies. On the other hand, it can be processed using the well-known FFT algorithm. In this sense, this tool is a middle term between Fourier and wavelet analyses, which can be used for stationary and non-stationary signals. Automotive NVH signals can be stationary (e.g., idle, cruise) or non-stationary, i.e., time-varying signals (e.g., door closing/opening, run-up, rundown). The objective of this work is to propose the use of the constant Q transform, developed originally for musical signal processing, for automotive NVH (run up, impact strip and door closing) time-frequency analyses. Also, similarities and differences of the proposed tool when compared with Fourier and wavelet analyses are addressed.