Aural training method and apparatus to improve a listener's ability to recognize and identify similar sounds

摘要

A listener's ability to recognize and identify similar sounds is improved by forming exaggerated sounds which differ in spectral/temporal energy distribution from one another more than the similar sounds differ from one another and in generally the same way in which the similar sounds differ and testing the listener's ability to recognize and identify the exaggerated sounds. The exaggerated sounds are formed as follows: The similar sounds are spectrally transformed into a coordinate space and a linear function is projected between a pair of points representing the similar sounds in the coordinate space. Between the pair of points, the linear function represents a continuum of sounds between the similar sounds. To exaggerate the sounds, points are extrapolated outward from the pair of points along the linear function. An inverse transformation is applied to the extrapolated points to produce exaggerated sounds which differ from one another more than the similar sounds differ and in generally the same way that the similar sounds differ. Filtering extrapolation coefficients before extrapolating and filtering inverse transformation coefficients before applying the inverse transform provide pleasant, naturally sounding exaggerated sound.