Detection of hypernasal speech in children with cleft palate

The pronunciation of a speaker with cleft lip or palate or subjects with defective velopharyngeal mechanism is marked by hypernasality and actually an operation is necessary to reduce this hypernasality. Then an assessment of hypernasality is imperative to quantify the effect of the surgery and design the speech therapy sessions. In this paper, a new quantitative method is proposed to estimate hypernasality. The proposed method used the fact that an Autoregressive (AR) model for vocal tract system of a patient with cleft palate is not accurate, because the zeros appear in the frequency response of vocal tract system. Therefore in our method hypernasality was estimated by comparing the distance between the sequences of cepstrum coefficients extracted from AR model and Autoregressive Moving Average (ARMA) model. Since the proposed method needs only digitized speech data, it is provides a straightforward and cost-effective evaluation of hypernasality.