Listeners continuously re-normalize distorted HRTFs using visual and motion cues

Evidence

A well-known example where vision modifies the auditory perception is ventriloquism (Alais and Burr, 2004), where listeners perceive the sound to be coming from a direction other than the one it actually comes from, if the visual movements associated with the sound generation are synchronized with the sound. The synchronized movements then ‘capture’ the spatial perception of sound. The effect is most powerful if the separation between the sound and visual sources is less than about 30◦. However, the auditory cue may still be more salient than the visual cue in some cases. If the visual image is severely blurred, the perceived direction matches that of the sound source, and if both of them are blurred, an average direction is perceived. — Ville Pulkki, Matti Karjalainen, Communication Acoustics- An…, 1466 (#corroborates, via linked claim)

In one case at least, Stereophile magazine, useful measurements are made (full credit to John Atkinson), but then, by my observation, these data are rarely incorporated in the discussion by the subjective reviewers who practice the “take-it-home-and-listen-to-it” method. Adaptation sets in, and numerous non-auditory influences are at play. There are many quotable passages in these reviews, but one sticks in my memory. It pertained to a loudspeaker with significant faults, clearly revealed in Atkinson’s measurements (and ours), and one can only presume heard by the reviewer. The subjective reviewer, perhaps believing in “breaking in,” persisted and stated that after living with the speaker for about two months he experienced an epiphany, and he suddenly heard wonderful things. So, did the loudspeaker “break in”? No. Did the reviewer come to an accommodation with the faults in it? Yes. In the absence of comparison listening to other products, the reviewer will never know what really happened. Readers were encouraged to go spend a lot of money on a mediocre product. — , Toole, Floyd E - Sound reproduction-…, 158 (#expands, via linked claim)

Categorization

  • XREF: connects to: Immersive headphone audio remains compromised without personalized HRTFs; In-head localization stems from mismatched HRTFs and missing head tracking; Humans normalize individual HRTFs from birth to recognize neutral timbre (+6 more)

Broader Connections