부산대학교 도서관

The effects of environmental noise were estimated from automatic computer-assisted analyses of voice samples. Signals consisted of a live voice sample and a synthesized triangular waveform. Noise was generated from a personal computer fan. Six different A-weighted signal-to-noise [S/N(A)] conditions were created for the live voice and synthetic signal— 25, 20, 15, 10, 5, and 0 dB. Results revealed that automatic estimates were systematically affected by different S/N levels. As the noise floor increased, baseline estimates of jitter and shimmer also increased in value. Results are discussed with reference to safeguards and standards in voice recording and analysis.