Effect of frequency modulation of F0 on the accuracy of formant frequency matches
Pascal Dissard, Chris J. Darwin
Abstract
Pascal Dissard, Chris J. Darwin
Abstract
Listeners adjusted the formant frequency of a one-formant complex sound to match the timbre of a similar fixed sound (formant frequency =1100 or 1200 Hz). On each trial the two sounds’ fundamental frequencies (F0) were in the ratio 1 : 1.13. In half the blocks the sounds were on low fundamentals (80 and 90.4 Hz), giving unresolved harmonics in the region of the formant peak; in the other half of the blocks the F0 was high (221.2 and 250 Hz), giving resolved harmonics near the formant peak. Orthogonal to this manipulation the F0 was either unmodulated for both sounds, or had a frequency modulation (FM) of 6 Hz and 3% depth. Data from seven listeners showed that, with no FM, matches are more variable within each listener for sounds with resolved than for sounds with unresolved harmonics, but that this difference disappears with FM. The result provides evidence that frequency modulating F0 improves listeners’ ability to estimate a formant frequency that is cued by resolved harmonics. [Work supported by EPSRC Grant No. GR/L03422.]
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Listeners adjusted the formant frequency of a one-formant complex sound to match the timbre of a similar fixed sound (formant frequency =1100 or 1200 Hz). On each trial the two sounds’ fundamental frequencies (F0) were in the ratio 1 : 1.13. In half the blocks the sounds were on low fundamentals (80 and 90.4 Hz), giving unresolved harmonics in the region of the formant peak; in the other half of the blocks the F0 was high (221.2 and 250 Hz), giving resolved harmonics near the formant peak. Orthogonal to this manipulation the F0 was either unmodulated for both sounds, or had a frequency modulation (FM) of 6 Hz and 3% depth. Data from seven listeners showed that, with no FM, matches are more variable within each listener for sounds with resolved than for sounds with unresolved harmonics, but that this difference disappears with FM. The result provides evidence that frequency modulating F0 improves listeners’ ability to estimate a formant frequency that is cued by resolved harmonics. [Work supported by EPSRC Grant No. GR/L03422.]
Key concepts: Formant, Harmonics, Acoustics, Fundamental frequency, Timbre, Frequency modulation, Modulation (music), Harmonic