In this study an interactive source model is proposed for voiced speech. The proposed model is compared with non-interactive model on glottal flow estimates obtained using closed-phase co-variance analysis. The results show that the interactive model is better than its non-interactive counterpart.