Psychoacoustics

Psychoacoustics: Volume

  • Solid Extron article

  • Robinson-Dadson curve (AKA Fletcher-Munson curve)

  • Three frequency bands

    • Below 100 Hz: whatever
    • 100—400 Hz: bass
    • 400Hz—2 Khz: midrange
    • 2—10 KHz: treble
    • 10KHz and up: whatever
  • Three volume bands in phon: perceived dB relative to 1KHz

    • 40 phon: low (A-weighting, midrange)
    • 50 phon: normal (B-weighting, moderate midrange)
    • 70 phon: loud (C-weighting, flat)
    • 100+ phon: aircraft (D-weighting, treble)

Volume, Loudness, Presence

  • Volume knob is log: ideal midpoint around 50 dB

  • Voltage levels are a mess, with multiple standards: usually 1—2 Vpp maximum.

  • A "loudness" control typically provides a big bass boost and a smaller treble boost

  • A "presence" control gives a treble boost, but with some feedback and distortion at high volume (by reducing power amp high-frequency feedback [or simulating that])

Psychoacoustics: Harmonics, Stretch Tuning, Masking

  • Recall: harmonics are multiples of fundamental frequency produced by distortion

  • Because the ear is not so sensitive at low and high frequencies (at normal volumes), it selectively hears midrange harmonics of bass notes

  • This means that a piano, for example, needs to be "stretch tuned" so that the midrange harmonics sound in tune

  • The low frequencies are partially "masked"

Psychoacoustics: Time Scales

  • Let's assume a 50Ksps sample rate

  • Smallest useful sample chunk for most things: 100 samples, 50Hz, 2ms

  • Fused sound: 500-2500 samples, 10-50 ms

  • By 20ms (1000s) latencies will be perceptible

  • By 100ms (5000s) latencies will be annoying: larger latencies are perceived as intolerable

Last modified: Monday, 20 April 2020, 12:30 PM