A socio friendly approach to the analysis of emotive speech

Proceedings of …, 2008

We propose a highly effective method of analysis of emotiveness in utterances, which clearly outperforms present ones. The method is based on analysis of emotive features of the lexical layer of user's utterances and is supported by analysis of non-lexical emotive features conveyed in text. The system based on this method acquired 93% of accuracy in recognizing emotiveness of user's utterances and was able to propose emotive value and extract specific emotion on a level comparable to human evaluators.

This paper presents an analysis of affective sentences spoken by a single speaker. The corpus was analyzed in terms of differ- ent acoustic and prosodic features, including features der ived from the decomposition of pitch contours into phrase and ac- cent curves. It was found that sentences spoken with a sad affect were most easily distinguishable from other affects as they were characterized by a lower F0, lower phrase and accent curves, lower overall energy and a higher spectral tilt. Fearful was also relatively easy to distinguish from angry and happy as it exhib- ited flatter phrase curves and lower accent curves. Angry and happy were more difficult to distinguish from each other, but angry was shown to exhibit a higher spectral tilt and a lower speaking rate. The analysis results provide informative cl ues for synthesizing affective speech using our proposed recombi- nant synthesis method.

Journal of Speech Sciences, 2025

This pilot study reports on acoustic and perceptual profiles of two American female speakers' productions of six American English social affective expressions: Authority, Declaration, Irritation, Sincerity, Uncertainty and walking on eggs (WOEG) as spoken in the linguistic sentence frame, Mary was dancing. The acoustic profile describes the prosodic characteristics (F0, intensity, duration, voice quality and tonal targets) of the utterances as a whole, as well as the voice quality characteristics of the nuclear stress syllable in the utterances. The perceptual profiles describe (1) listeners' 3-dimensional VAD emotional ratings, i.e., Valence, Arousal, and Dominance, of the utterances and (2) listeners' auditory impressions of the nuclear stress syllable. Multifactorial Analyses (MFA) were applied to examine the relation between the prosodic characteristics and the VAD scales, and also the relationship between voice quality measurements on the nuclear stress vowel and auditory perceptions. The prosodic MFA results indicate that for these two American English speakers, a soft / noisy voice, with weak harmonics and irregular rhythm with pauses and hesitations, as in the expressions of Uncertainty and WOEG, is perceived by listeners as accommodating (not Dominant) and not positive (negative Valence). Loud, tense voices with energy in the upper frequencies, as in the expression of Irritation, are perceived as Aroused. Expressions of Authority, Declaration, and Sincerity tend to have comparatively regular rhythm and relatively flat intonation. The MFA analysis of voice quality measurements and auditory perceptions suggests that Normalized Amplitude Quotient (NAQ) may indeed be a good estimate for tense voice due to glottal closing behavior, Cepstral Peak Prominence (CPP), a good estimation for strong non-noisy harmonics, Peak Slope, a good estimate of spectral related tense voice, and Hammarberg Index, for distribution of spectral energy, i.e., strong or weak energy in the upper frequencies. But these measures do not completely account for the auditory judgments of voice quality. An interpretation of the acoustic and perceptual profiles of the social affective expressions is discussed in terms of theoretical codes (Frequency Code and Effort Code) proposed for explaining symbolic relations between vocal characteristics and pragmatic affective processing in spoken human communication.

2013

Emotion based speaker Identification System is the process of automatically identifying speaker’s emotion based on features extracted from speech waves. This paper presents experiment with the building and testing of a Speaker’s emotion identification for Hindi speech using Mel Frequency Cepestral Coefficients and Vector Quantization techniques. We collected voice samples of Hindi speech sentences in four basic emotions to study speaker’s emotion identification and it was found that with proposed emo-voice model we are able to achieve accuracy of 73% of speaker’s emotion identification in speech out of 93% of the total speech samples provided to the system.

Speech is the most important bio signal that human being can produce and perceive. Speaking is the process of converting discrete phonemes into continuous acoustic signal. Speech is the most natural and desirable method of human communication as a source of information to communicate with intentions and emotions. it is a time continuous signal containing information about message, speaker attitude, language accent , dialect and emotion many psychological studies suggests only 10% of human life is completely unemotional the rest involves emotional. This paper aims to various speech features used for the analysis of speaker utterance of emotional speech of human being which are analyzed by using speech analysing software's(PRATT) and MATLAB. The emotional analysis helps in the study of what emotions are and how the speech features properties change for different emotional state of a human being for different languages as subject.

Proceedings of the 18th International Congress of Phonetic Sciences, 2015

This paper deals with the issue of the influence of verbal content on listeners who have to identify or evaluate speech emotions, and whether or not the emotional aspect of verbal content should be eliminated. We compare the acoustic parameters of sentences expressing joy, anger, sadness and neutrality of two groups: (1) where the verbal content aids the listener in identifying emotions; and (2), where the verbal content does not aid the listener in identifying emotions. The results reveal few significant differences in the acoustic parameters of emotions in the two groups of sentences, and indicate that the elimination of emotional verbal content in speech presented for emotion identification or evaluation is, in most cases, not necessary.

International Journal Of …, 2010

The exploration of how we as human beings react to the world and interact with it and each other remains one of the greatest challenges. The ability to recognize emotional states of a person perhaps the most important for successful inter personal social interaction. Automatic emotional speech recognition system can be characterized by the used features, the investigated emotional categories, the methods to collect speech utterances, the languages and the type of the classifier used in the experiment. Since a well defined database is the necessary precondition for improving the performance Automatic emotional speech recognition systems. This paper explores the theories that explain the social and cognitive roles of emotions and mental states and their expression in human behaviors and communication. The paper describes the planning and accomplishment of a native language emotional speech database of acted emotional speech by number of speakers, recording strategies, conversion etc as well as the alternative approach is briefly addressed. Such database would also contribute to research in intonation and emotion.

6th International Conference on Spoken Language Processing (ICSLP 2000)

Investigation was conducted on how prosodic features of emotional speech changed depending on emotion levels. The analysis results on fundamental frequency (F 0) contours and speech rates implied that humans have several ways to express emotions and use them rather randomly. Investigation was also conducted on what acoustic features were important to express emotions. Perceptual experiments using synthetic speech with copied acoustic features of target speech indicated importance of the segmental features other than the prosodic features. Especially, a high importance was observed in the case of happiness.

The Buckingham Journal of Language and …, 2010

In spontaneous speech, emotion information is embedded at several levels: acoustic, linguistic, gestural (non-verbal), etc. For emotion recognition in speech, there is much attention to acoustic level and some attention at the linguistic level. In this study, we identify paralinguistic markers for emotion in the language. We study two Indian languages belonging to two distinct language families. We consider Marathi from Indo-Aryan and Kannada from Dravidian family. We show that there exist large numbers of specific paralinguistic emotion markers in these languages, referred to as emotiphons. They are intertwined with prosody and semantics. Preprocessing of speech signal with respect to emotiphons would facilitate emotion recognition in speech for Indian languages. Some of them are common between the two languages, indicating cultural influence in language usage.