Speech emotion recognition using deep learning: from basic to complex emotions in unimodal and multimodal frameworks | Synapse