Audio : Perception (Psychoacoustics)

The Ear

Frequency Range

• from 20Hz to 20 kHz
• pitch = perceived frequency

Dynamic Range (amplitude)

• from the threshold of audibility to the threshold of pain (from 0 dB to 120dB)
• loudness
  • perceived amplitude
  • increases logarithmically with amplitude
  • Fletcher-Munson curvesshow how perception is frequency dependent. Each curve in the graph shows the sound intensity level needed to produce a perception of constant loudness, as a function of frequency. The ear is most sensitive in middle frequencies, between 700Hz and 6KHz.
• decibel
  • a relative measure of loudness
  • dB = 20 log₁₀ (A/B),
    where A,B are peak amplitudes of two sounds
  • 6 dB = twice the (physical) loudness
  • for absolute loudness, set B = threshold of hearing.
    0 dB = quietest sound, positive dBs are louder
  • for relative loudness, set A = 0dB (loudest sound present, or "unity gain")
    negative dBs for quieter sounds
• percieved loudness varies as the cube root of intensity (dB).
  eg) need 8 times the loudness to get a doubling of perceived loudness (18 dB)
  eg) 800 watt amp sounds twice as loud as 100 watt amp

Spectral Characteristics

formants

• regions of spectrum that are emphasized/deemphasized
• due to physical resonator (eg, mouth cavity, violin body)
• usually at a fixed position in spectrum, even if pitch changes

timbre

• subjective quality of a sound (what it "sounds like")
• includes spectrum changes over time

pitch

harmonic spectra perceived as a single tones

Masking

critical band

• a region of spectrum related to perception
• frequencies within a critical band "fuse"
• frequencies separated by more than a critical band sound separate
• are closer together at low frequencies

Binaural hearing

2D Localization

IID - Interaural Intensity Difference

sounds louder in ear facing source (panning)

ITD - Interaural Time Difference

sound hits the ear closer to the source first

3D Localization

HRTF - Head-Related Transfer Function

frequency-dependent amplitude and time differences due to shaping of outer ears (put this microphone in your ear, like this, and sit here and we'll do the rest (measure your HRTF). trust me.)

Distance

intensity

closer is louder: inverse square law says 6dB per doubling of distance (but 9 dB per doubling sounds better!?)

reverberation: reflections (echoes) due to enclosure. More reverb ("wetter mix") sounds farther.