Next:
Sound Pressure Level and
Up:
Lectures
Previous:
Resonance
The Hearing Mechanism
Critical Bands
Binaural Hearing and Localization
Hearing
The Hearing Mechanism
Figure 14:
The hearing mechanism (from Rossing, 1990).
The Outer Ear:
Pinna and auditory canal.
Auditory canal can be analyzed (to first approximation) as a open-closed pipe.
Both the pinna and auditory canal impose filtering characteristics based on unique shapes.
The Middle Ear:
Eardrum and ossicles.
Ossicles consist of three bones shaped like a hammer, anvil, and stirrup (not really, but that is the convention everyone uses).
Ossicles act as a mechanical transformer, converting pressure on the eardrum to pressure on the oval window with a step-up factor of up to 30.
Most of the step-up factor is due to area differences (eardrum vs. oval window) ... about 20 times.
Lever action of the ossicles accounts for another 1.5 factor.
The Inner Ear:
Semicircular canals control balance.
Cochlea transforms pressure variations to neural impulses.
Approximately 30,000 hair cells along basilar membrane, which itself is about 37 millimeters long.
Each hair cell has many cilia or hairs which bend with the vibrations of the basilar membrane.
High-frequency detection occurs along rigid section of the basilar membrane closest to the oval window.
Low-frequency detection (greatest low-frequency amplitude ripples) occurs near far end of the basilar membrane.
Bone conduction to inner ear is also significant.
Auditory nerve fibers are ``tuned'' to different center frequencies.
Critical Bands
Regions of amplitude envelope ``distinction'' along basilar membrane.
Determined by psychoacoustic experiments, their values typically vary depending on the type of response being measured.
About 24 critical bands along basilar membrane.
Each critical band is about 1.3 mm long and embraces about 1300 neurons.
Binaural Hearing and Localization
IID: Interaural intensity difference is dominant for frequencies greater than about 4000 Hz.
ITD: Interaural time difference is dominant for frequencies less than about 1000 Hz.
HRTFs: Head-related Tranfer Functions are commonly used in 3D reproduction systems.
Need a separate HRTF for each ear and for each position. Active research to find ways to interpolate and compress the data.
``The Precedence Effect''.
©1999 CCRMA, Stanford University. All Rights Reserved.
Maintained by
Gary P. Scavone
,
gary@ccrma.stanford.edu
.