1995•The Journal of the Acoustical Society of AmericaRequires access

The identification of vowel-like harmonic complexes: Effects of component phase, level, and fundamental frequency

Joseph I. Alcántara, Brian C. J. Moore

Open publisher page 21 citations

Abstract

These experiments investigate how the identification of vowel-like harmonic complexes, similar to those used by Leek et al. [J. Acoust. Soc. Am. 81, 148-154 (1987)], is affected by spectral contrast, overall level, component phase, and fundamental frequency (F0). Four normally hearing subjects were required to identify which of six vowel-like harmonic complexes was presented on each trial. The test stimuli were complex tones containing the first 35 harmonics of a 100-Hz fundamental or the first 70 harmonics of a 50-Hz fundamental. All of the harmonics with frequencies below 3000 Hz were equal in amplitude except for three pairs of successive harmonics which were located at the first, second, and third formant frequency values, and incremented in level by 1, 2, 4, 8, and 16 dB relative to the other components. Three overall levels were used, 85, 65, and 45 dB SPL, and harmonics were added in either cosine or random phase. The results indicated that identification was better for cosine phase than for random phase, except for the 100-Hz fundamental at 45 dB SPL. The difference between the two phase conditions increased with increasing presentation level and with decreasing fundamental frequency. The results are explained in terms of the waveforms that would occur at the outputs of different auditory filters. It does not appear necessary to invoke nonlinear enhancement mechanisms to explain the results, although an influence of such mechanisms cannot be ruled out.

About this research paper

What this paper is about

These experiments investigate how the identification of vowel-like harmonic complexes, similar to those used by Leek et al. [J. Acoust. Soc. Am. 81, 148-154 (1987)], is affected by spectral contrast, overall level, component phase, and fundamental frequency (F0). Four normally hearing subjects were required to identify which of six vowel-like harmonic complexes was presented on each trial. The test stimuli were complex tones containing the first 35 harmonics of a 100-Hz fundamental or the first 70 harmonics of a 50-Hz fundamental. All of the harmonics with frequencies below 3000 Hz were equal in amplitude except for three pairs of successive harmonics which were located at the first, second, and third formant frequency values, and incremented in level by 1, 2, 4, 8, and 16 dB relative to the other components. Three overall levels were used, 85, 65, and 45 dB SPL, and harmonics were added in either cosine or random phase. The results indicated that identification was better for cosine phase than for random phase, except for the 100-Hz fundamental at 45 dB SPL. The difference between the two phase conditions increased with increasing presentation level and with decreasing fundamental frequency. The results are explained in terms of the waveforms that would occur at the outputs of different auditory filters. It does not appear necessary to invoke nonlinear enhancement mechanisms to explain the results, although an influence of such mechanisms cannot be ruled out.

Why it matters

OpenAlex reports 21 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

These experiments investigate how the identification of vowel-like harmonic complexes, similar to those used by Leek et al. [J. Acoust. Soc. Am. 81, 148-154 (1987)], is affected by spectral contrast, overall level, component phase, and fundamental frequency (F0). Four normally hearing subjects were required to identify which of six vowel-like harmonic complexes was presented on each trial. The test stimuli were complex tones containing the first 35 harmonics of a 100-Hz fundamental or the first 70 harmonics of a 50-Hz fundamental. All of the harmonics with frequencies below 3000 Hz were equal in amplitude except for three pairs of successive harmonics which were located at the first, second, and third formant frequency values, and incremented in level by 1, 2, 4, 8, and 16 dB relative to the other components. Three overall levels were used, 85, 65, and 45 dB SPL, and harmonics were added in either cosine or random phase. The results indicated that identification was better for cosine phase than for random phase, except for the 100-Hz fundamental at 45 dB SPL. The difference between the two phase conditions increased with increasing presentation level and with decreasing fundamental frequency. The results are explained in terms of the waveforms that would occur at the outputs of different auditory filters. It does not appear necessary to invoke nonlinear enhancement mechanisms to explain the results, although an influence of such mechanisms cannot be ruled out.

Key concepts: Harmonics, Fundamental frequency, Harmonic, Vowel, Formant, Acoustics, Mathematics, Phase (matter)

Related papers

Back to paper searchBrowse research topicsOriginal source
The identification of vowel-like harmonic complexes: Effects of component phase, level, and fundamental frequency — Research Paper | ScholarLens