Tonal Metrics in the Presence of Masking Noise and Correlation to Subjective Assessment

As the demand for Sound Quality improvements in vehicles continues to grow, robust analysis methods must be established to clearly represent end-user perception. For vehicle sounds which are tonal by nature, such as transmission or axle whine, the common practice of many vehicle manufacturers and suppliers is to subjectively rate the performance of a given part for acceptance on a scale of one to ten. The polar opposite of this is to measure data and use the peak of the fundamental or harmonic orders as an objective assessment. Both of these quantifications are problematic in that the former is purely subjective and the latter does not account for the presence of masking noise which has a profound impact on a driver's assessment of such noises.

This paper presents the methodology and results of a study in which tonal noises in the presence of various level of masking noise were presented to a group of jurors in a controlled environment. Their subjective ratings were collected and correlated to noise and vibration metrics. These findings were incorporated within a custom software application which imports a test data file, determines the present masking level, and assigns a rating for the tonal noises above masking noise which is robustly correlated to the subjective impressions of a large test sample of end users. This data collection and analysis procedure has been adopted as a standard procedure within a major American vehicle manufacturer.