Sharp spectral peaks, not narrow gaps, drive strong pitch perception

Evidence

If pitch is not determined by periodicity, perhaps then by autocorrelation? Now we are getting somewhere. The first major peak in the autocorrelation after the ever-present peak at time 0 in the preceding example occurs at 0.00956 second, which corresponds to 104.6 Hz. If sounds resolve into several distinct pitches, these will have corresponding peaks in the autocorrelation. Pitch is associated with something more important than the presence or absence of a single partial: its tendency to repeat itself at given intervals. The generalization of this concept to sounds that are not strictly periodic is the autocorrelation function. Periodic signals will always have peaks at multiples of the period that mirror the peak at time zero. The analogs of such periodicity are peaks in the autocorrelation of nonperiodic signals. — Eric J. Heller, Why You Hear What You Hear- An…, 2033 (#corroborates, via linked claim)

Sharpness is a sensation which we can consider separately, and it is possible, for example, to compare the sharpness of one sound with the sharpness of another. Sharpness can also be doubled or halved if variables are available that really change sharpness. The variability of sharpness judgements is comparable to that of loudness judgements. One of the important variables influencing the sensation of sharpness is the spectral envelope of the sound. Many comparisons have indicated that the spectral fine structure is relatively unimportant in sharpness. A noise producing a continuous spectrum, for example, has the same sharpness as a sound composed of many lines if the spectral envelopes measured in critical-band levels are the same. — Professor Dr.-Ing. Hugo Fastl, Professor Dr.-Ing. Eberhard, Psychoacoustics- Facts and Models, 1340 (#contradicts, via linked claim)

Categorization

• XREF: connects to: Complex-tone pitch tracks the spectral pitch of lower components — Professor Dr.-Ing. Hugo Fastl, Professor Dr.-Ing. Eberhard, Psychoacoustics- Facts and Models, 706; Spectral line structure predicts pitch strength magnitude.; Pitch perception corresponds to peaks in the autocorrelation function — Eric J. Heller, Why You Hear What You Hear- An…, 2033 (+8 more)

Related

• Spectral envelope, not fine structure, determines perceived sharpness (similarity 0.90)
• Pitch resolution exceeds smeared bandwidth via peak-locating algorithms (similarity 0.88)
• Pitch perception corresponds to peaks in the autocorrelation function (similarity 0.88)
• Continuous-spectrum noises can carry pitch perceived from spectral peaks (similarity 0.86)
• Even nonperiodic noise sounds convey a weak pitch sensation based on spectral energy distribution (similarity 0.86)
• Sharpness depends primarily on spectral content, not on level or bandwidth. (similarity 0.86)
• Higher partials dominate pitch because their sharper autocorrelation peaks win (similarity 0.85)
• Pitch is a human synthesis keyed to autocorrelation, not pure physics (similarity 0.85)
• Perceived pitch is set primarily by fundamental frequency. (similarity 0.85)
• Noise with a sharp spectral peak is perceived as pitched. (similarity 0.85)
• Sounds with steep spectral peaks evoke stronger pitch than sounds with narrow spectral notches. (similarity 0.84)
• Narrow spectral bands with steep slopes produce stronger pitch than narrow spectral gaps do (similarity 0.84)
• Listeners reconstruct a perceived pitch that is not physically present (similarity 0.84)
• Sharpness depends on the spectral envelope, not fine structure (similarity 0.83)
• The auditory system performs pitch analysis as autocorrelation per sub-band, summed into a correlogram. (similarity 0.83)

Broader Connections

• Pitch perception corresponds to peaks in the autocorrelation function #corroborates (0.85) — Eric J. Heller, Why You Hear What You Hear- An…, 2033
• Spectral envelope, not fine structure, determines perceived sharpness #contradicts (0.80) — Professor Dr.-Ing. Hugo Fastl, Professor Dr.-Ing. Eberhard, Psychoacoustics- Facts and Models, 1340