EURASIP Journal on Audio, Speech, and Music Processing

Table 2 The performance of speech recognition for clean emotional speech using Kaldi ASR trained with neutral speech

From: Robust emotional speech recognition based on binaural model and emotional auditory mask in noisy environments

Emotional states	Anger	Disgust	Fear	Happiness	Sadness	Neutral
WER (%)	11.31	13.92	13	12.52	11.60	1

Back to article page