Audio-vocal responses of vocal fundamental frequency and formant during sustained vowel vocalizations in different noises

Sustained vocalizations of vowels [a], [i], and syllable [mə] were collected in twenty normal-hearing
individuals. On vocalizations, five conditions of different audioevocal feedback were introduced separately
to the speakers including no masking, wearing supra-aural headphones only, speech-noise
masking, high-pass noise masking, and broad-band-noise masking. Power spectral analysis of vocal
fundamental frequency (F0) was used to evaluate the modulations of F0 and linear-predictive-coding was
used to acquire first two formants. The results showed that while the formant frequencies were not
significantly shifted, low-frequency modulations (<3 Hz) of F0 significantly increased with reduced audio
evocal feedback across speech sounds and were significantly correlated with auditory awareness of
speakers' own voices. For sustained speech production, the motor speech controls on F0 may depend on
a feedback mechanism while articulation should rely more on a feedforward mechanism. Power spectral
analysis of F0 might be