Harmonic pitch: dependence on resolved partials, spectral edges, and combination tones

Area of Science:

• Auditory perception
• Psychoacoustics
• Signal processing

Background:

• Pitch perception is crucial for understanding complex sounds.
• Harmonic sounds contain multiple partial tones that influence perceived pitch.
• The relative contribution of individual partials to pitch remains an area of investigation.

Purpose of the Study:

• To quantify the perceptual weights of individual partial tones in harmonic sounds.
• To determine how the removal of low-numbered partials affects pitch judgments.
• To investigate the role of resolved and unresolved partials in pitch perception.

Main Methods:

• A pitch-comparison experiment was conducted using harmonic sounds (F0=200 Hz) with partials up to the 12th.
• Partial frequencies were randomly perturbed, and perceptual weights were estimated using correlation coefficients.
• Low-numbered partials were systematically removed to isolate the effects of higher partials.

Main Results:

• Listeners predominantly relied on low-numbered, resolved partials within the dominance region when all partials were present.
• When low-numbered partials were removed, listeners shifted their attention to spectral edges (lowest and highest partials).
• No influence on pitch was observed from unresolved partials, suggesting no contribution from temporal cues.

Conclusions:

• Perceptual weights for harmonic sounds are primarily determined by resolved partials with high frequency precision.
• The findings align with Goldstein's optimal-processor model, where precise frequency estimation leads to greater influence on pitch.
• Listeners adapt their listening strategy based on the available spectral information, prioritizing resolved components and spectral edges.