Conditional Spectral Analysis Research Articles

CSA-based acoustic beamforming for the contribution analysis of air-borne tyre noise

The contributions of air-borne tyre noise to the overall interior noise sources of a vehicle were investigated using an acoustic beamforming technique based on the conditioned spectral analysis (CSA). This was realised using a set of accelerometers located at the various sources to measure the reference signals of these sources and, these reference signals were then applied successively in the CSA to eliminate the linear effects of these reference signals from the array microphone signals being processed. CSA is a signal separation technique that can be used for multi-reference purposes. In the CSA technique, multi-reference signals can be used to modify the original cross-spectral matrix of array microphone signals, and the combination of the acoustic beamforming method with CSA results in a powerful technique that can be used to separate sound maps into components of different mechanisms. In this paper, different signal separation techniques and their applications in acoustic beamforming are introduced. CSA-based acoustic beamforming is then proposed. The efficacy of the CSA-based acoustic beamforming was examined experimentally. Finally, the CSA-based acoustic beamforming was applied to the interior noise source localization of a vehicle to estimate the contributions from the air-borne tyre noise and other noise sources to the overall interior sound maps.

Multi-channel spectral analysis of multi-pass acquisition measurements

Conditioned spectral analysis (CSA) and virtual source analysis (VSA) are two techniques currently used nowadays to post-treat measurements on multiple input multiple output (MIMO) systems. Those techniques are based on the estimation of cross-spectral quantities between outputs and the so-called “references”. In some measurement situations, all outputs cannot be acquired simultaneously, several measurement passes are necessary (using a scanning system, a robot or a laser vibrometer). An approach is proposed in this paper to adapt CSA and VSA to MIMO systems requiring a multi-pass acquisition. The approach is based on the average over all passes of a set of reference channels, and on the correction of cross-spectral quantities by using transfer functions and averaged reference spectra. An academic experiment is finally realized to illustrate the proposed approach.