How twang vocal timbre impacts intelligibility in noise

Twang vocal timbre is thought to be an efficient voice strategy that minimizes vocal effort while maximizing acoustic output. Therefore, the use of twang timbre may enhance communication success in noisy environments while preserving the voice. However, the intelligibility of twang speech has not been evaluated. This study investigated how vocal timbres (neutral and twang) influence intelligibility in common traffic (train- and plane-shaped) noises. Sixty-two participants identified words within a 50-word English sentence matrix. Sentences with twang and neutral timbres were synthesized from eight speakers (four female and four male) with deep-learning timbre conversion that controlled sound level, pitch, prosody, and accent. Stimuli were masked with white noise shaped with spectral contours from ambient train and plane noise at −18 and −24 dB SNRs. Intelligibility, determined by word identification accuracy, was significantly higher in twang speech than in neutral speech for both female and male speakers. The twang advantage might arise from higher perceived loudness, higher formant frequencies from F1 to F3, the clustering of F3 and F4, and/or the less breathy or hoarse voice quality in twang compared to neutral timbre. The result suggests that vocal timbre adjustment is a valid strategy for talkers to improve intelligibility in noise.