HRTF

Head-related transfer function (HRTF) is a transfer function that describes how an ear receives sound from a point in space, accounting for the filtering effects of the head, torso, and outer ears. For a given direction, the HRTF specifies the frequency-dependent modifications that reach each eardrum. In practice, HRTFs are typically represented as head-related impulse responses (HRIRs) in the time domain or as magnitude and phase spectra in the frequency domain for the left and right ears.

HRTFs depend on the direction, usually described by azimuth and elevation, and vary with the listener’s anatomy. They encode localization cues such as interaural time differences, interaural level differences, and spectral shaping produced by the pinnae, head, and torso.

Measurement and data: HRTFs are measured from a subject or a dummy head with microphones placed in the ears in a quiet, anechoic environment, using sound sources from many directions. Datasets are then interpolated to render continuous space. Publicly available examples include the CIPIC HRTF Database and KEMAR/IRCAM-based measurements. Data are provided as HRIRs or corresponding directional filters.

Applications: In binaural rendering for headphones, a mono input is filtered by the left and right HRIRs corresponding to the perceived source direction to produce spatialized audio. HRTFs are widely used in virtual reality, gaming, and other 3D-audio systems, often with head tracking to update cues as the listener moves.

Variability and limitations: HRTFs are highly individualized, so generic sets can produce plausible but imperfect localization and front-back confusions for some listeners. Individual customization requires measured HRIRs or optimization, and real-time rendering may incorporate dynamic updates for head movements.