Abstract

It was shown that, for vowel sounds, the spectrum relates to fundamental frequency (fo) and the spectral envelope is ambiguous, often representing two or three different vowel qualities if fo is varied substantially. Thus, from a speech perception perspective, vowel quality is indicated to relate to the pitch of a vowel sound. In this contribution, a concept is outlined addressing the experimental question of the relationship of vowel recognition and pitch. According to this concept, two or three harmonics are extracted near the spectral peaks of a natural vowel sound, representing assumed F1–F2 or F1–F2–F3, and the lowest harmonic(s) below the first spectral peak are added. Based on a single extracted harmonics pattern, a series of sounds are synthesised stepwise attenuating the levels of the lowest harmonics with the aim to effect a low-to-high transition of the highest common harmonics factor and pitch. Pitch level and vowel recognition of the sounds is then investigated by means of listening tests. Results of a first pilot study based on sounds of mid-closed vowels /e, ø, o/ confirmed the vowel–pitch relation hypothesis and also revealed cases of sounds with double-pitch and double-vowel recognition.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call