Figure 1From: Context-dependent acoustic modeling based on hidden maximum entropy model for statistical parametric speech synthesis HSMM structure and its output and duration probability distribution. Back to article page