Concatenative Text‐to‐Speech Synthesis Based on Sinusoidal Modelling