Integrated tone evaluation in Mandarin CALL systems using competing model based approach

Mandarin is known as a tonal language, mainly because its tone is a distinctive discriminative feature. This paper describes an attempt to automate the tone evaluation for continuous Mandarin speech in CALL systems. A competing model based integrated evaluation approach is introduced to obtain objective tone evaluation results. The evaluation includes two parts: tonal syllable evaluation and tone evaluation. In tonal syllable evaluation, we generate the syllable-based competing models by using the simplified linguistic knowledge based initial/final net. Then the competing model based approach is introduced to get tonal syllable score. In tone evaluation, an HMM forced alignment based tone model is used to get tone score for Mandarin syllables. Finally, we integrate tonal syllable score and tone score together to acquire the overall tone scoring results. The experimental results demonstrate that this proposed method, using forced alignment based tone model and competing models based tone evaluation, gives an accurate tone assessment.