Carlos Ishi

Prosody and voice quality: - Analysis of laryngeal voice qualities: automatic detection of creaky voice; automatic detection of aspiration noise in breathy and whispery voices. - Mapping between prosodic + voice quality features and linguistic and paralinguistic functions (intentions, emotions, and attitudes) in Japanese. - Transcription of automatic prosodic events: extraction of perceptually meaningful prosody events for automatic prosody labeling: focus on phrase final prosody and voice quality. - Pitch perception: Correspondence between acoustically observed F0 and perceived pitch movements. - Robust F0 extraction. Speech and gestures: - Analysis of head motion and speech in spoken dialogue: automatic generation of head motions from speech. - Multi-modal dialogue processing. - Lip motion generation/synchronization for humanoid robots (including androids) based on speech acoustics. - Head motion generation from speech acoustics and linguistic information. - Facial expression and motion generation in humanoid robots (including androids) based on speech acoustics. Robot Audition and Sound Environment Intelligence: - Microphone array for audio source localization and separation. - Improvement of speech recognition and understanding in noisy environments. - Utterance interval detection based on sound directivity. - Utterance interval detection based on audio-visual information. -Sound environment map generation Speech Perception and Recognition - Auditory representation of speech signals: acoustic parameters related to auditory perception; masking functions. - Prosodic modeling applied to recognition of linguistic and paralinguistic information Speech Production and Synthesis - Mapping between physiological and acoustic features for laryngeal voice quality control - Prosodic control and voice quality control for Speech Synthesis