Figure 10From: Context-dependent acoustic modeling based on hidden maximum entropy model for statistical parametric speech synthesis RMSE as objective measure to compare log F0 trajectories generated by decision tree-based HMEM and conventional HSMM. Back to article page