Abstract:
This paper introduces the Octavector Model, a novel theoretical framework positing that the central auditory system represents sound not as a collection of disparate features, but as a unified eight-dimensional vector. We define the canonical axes of this auditory state-space as: Frequency (Pitch), Intensity (Loudness), Timbre (Spectral Shape), Duration (Temporal Boundaries), Azimuth, Elevation, Temporal Envelope, and Source Velocity. This work provides a comprehensive neurophysiological justification for the selection of these eight dimensions, reviewing the distinct neural pathways and coding mechanisms responsible for their extraction from the acoustic waveform. Furthermore, we propose a geometric interpretation of higher-order auditory processing. Within this framework, Auditory Scene Analysis (ASA) is conceptualized as the decomposition of the auditory state-space into distinct, low-dimensional neural manifolds, each corresponding to a perceived auditory object. Finally, we frame the entire perceptual process within the language of tensor network theory, modeling the auditory pathway as a metric tensor that transforms a covariant sensory Octavector into a contravariant perceptual or motor output. This model offers a mathematically rigorous and unifying perspective on auditory computation.
Yıldırım, E. (2025). A Geometric Framework for Auditory Perception: The Octavector Model of Neural Sound Encoding. Zenodo. https://doi.org/10.5281/zenodo.17049839
Leave a Reply