Space/time data-information in the A.R.I.A.L. project ear model

Abstract While reporting on work done to elucidate the structuring-up process of acoustic space vs. time data-information with in the cochlea, the present article stresses the paramount importance of the phase effect within the ear. This work is based on results secured with a peripheral-hearing model previously tested for reliability. Here, this model is broken up into mutually independent parts for the specific function each fulfils between the external ear and the nerve fiber endings. This model is described through a set of equations that apply to acoustic wave propagation, to the mechanicl vibration of the basilar membrane and to electro-mechanical transduction in hair cells and nerve fibers. Representing information that runs inside the cochlea becomes a crucial problem when trying to detect pertinent acoustic cues to be used in signal analysis. Moreover, for an improved perceptual quality, (e.g. in speech synthesis), phase is a definitely non-negligible factor. Therefore, while the sonagram should be retained as an indispensible instrument in monitoring intensity evolution, it is equally imperative to secure a mode of sample-by-sample visualization of other relevant phenomena. Given the above, a few examples suffice to demonstrate that: (1) characteristic sound patterns can be read out of neurogram, (2) phase is a parameter that should be included in the transmission of data-information as this structures up in space and time; indeed, as i may well turn out, structural parameters could contribute the most in speech recognition to the “sturdiness” of acoustic cues.