Fig. 5From: A unit selection text-to-speech-and-singing synthesis framework from neutral speech: proof of conceptPitch-scale factors (αst) for different vocal ranges (S0, S4, S7) and unit selection configurations with T=100 bpm. Whiskers are set to 2nd and 98th percentile. Differences between all configurations are statistically significant (p < 0.01) except for the pair S0pdC-S0pdLCBack to article page