DMO-DB[24]是由柏林工业大学录制的德语情感语音库,由10位演员(5男5女)对10个语句(5长5短)进行7种情感(中性/nertral、生气/anger、害怕/fear、高兴/joy、悲伤/sadness、厌恶/disgust、无聊/boredom)的模拟得到,共包含800句语料,采样率48kHz(后压缩到16kHz),16bit量化.语料文本的选取遵从语义中性、无情感倾向的原则,且为日常口语化风格,无过多的书面语修饰.语音的录制在专业录音室中完成,要求演员在演绎某个特定情感前通过回忆自身真实经历或体验进行情绪的酝酿,来增强情绪的真实感.经过20个参与者(10男10女)的听辨实验,得到84.3%的听辨识别率.(DMO-DB[24] is a German emotional speech database recorded by the Technical University of Berlin, by the 10 actors (5 male and 5 female) of 10 statements (5 long 5 short) of 7 emotions (/nertral, /anger, neutral angry afraid /fear, /joy, /sadness, happy sad hate /disgust, boring /boredom) simulation, contains a total of 800 sentence corpus, sampling rate of 48kHz (16kHz, 16bit after compression). Quantitative corpus selection with semantic emotional tendency, neutral principle, and daily oral style, without too much modification of the written language. The speech recorded in a professional studio, requirements in the interpretation of a particular emotional actor before through the memories of their true experience or experience of mood brewing, to enhance the sense of reality of emotions. After 20 participants (10 men and 10 women) in the perception experiment, get some recognition rate of 84.3%.)