The classification of speech according to emotional content employs acoustic measures in addition to pitch as classification input. In one embodiment, two different kinds of features in a speech signal are analyzed for classification purposes. One set of features is based on pitch information that is obtained from a speech signal, and the other set of features is based on changes in the spectral shape of the speech signal over time. This latter feature is used to distinguish long, smoothly varying sounds from quickly changing sound, which may indicate the emotional state of the speaker. These changes are determined by means of a low-dimensional representation of the speech signal, such as MFCC or LPC. Additional features of the speech signal, such as energy, can also be employed for classification purposes. Different variations of pitch and spectral shape features can be measured and analyzed, to assist in the classification of individual utterances. In one implementation, the features are measured individually for each of the first, middle and last thirds of an utterance, as well as for the utterance as a whole, to generate multiple sets of data for each utterance.

A method for electronically detecting human suicidal predisposition by analysis of an elicited series of vocal utterances from an emotionally disturbed or distraught person independently of linguistic content of the elicited vocal utterance includes converting the utterance into an electrical signal having time varying amplitude and frequency which are representative of the utterance, filtering frequency components of the signal above and below pre-selected frequencies to obtain a signal within the pre-selected frequencies, filtering non-repetitive components having amplitude above about 90 percent of average amplitude of the signal out of the signal, filtering repetitive signal components having frequency outside about 90 percent of frequency band width of the signal out of the signal and identifying as suicidally predisposed a person from whom the vocal utterance emanated if signal amplitude exhibits substantially non-instantaneous decays upon conclusion of each utterance or if amplitude of signal amplitude modulation is low or if frequency of signal amplitude modulation is low, relative to the decays of utterance signals or amplitude of signal amplitude modulation or frequency of signal amplitude modulation of depressed persons known otherwise to be in good mental health.

A method for detecting suicidal predisposition in a person by securing an utterance from the person, identifying the person as being suicidally predisposed if the utterance decays substantially non-instantaneously upon conclusion and identifying the person as being suicidally predisposed if signal amplitude modulation during the utterance is low. Low value of amplitude modulaation (of speech envelope waveform), as well as slow decay at the end of each utterance, are indicators of emotional disturbance, special filtering of repetitives and non-repetitive components enhanced the waveform for consideration.

A method for detecting suicidal predisposition in a person by securing an utterance from the person, identifying the person as being suicidally predisposed if the utterance decays substantially non-instantaneously upon conclusion and identifying the person as being suicidally predisposed if signal amplitude modulation during the utterance is low and identifying the person as being suicidally predisposed if variation in fundamental frequency during the utterance is low and identifying the person as being suicidally predisposed if frequency of amplitude modulation during the utterance is low.

The present invention relates to an apparatus for monitoring emotional states of an individual by using a voice analysis of said individual. The apparatus comprises a voice analyzer operative to input speech specimens, comprising an analog to digital converter operative to perform a digitization process of analog audio vocal segments, and a general emotion reporter, operative to produce an indication of any kind for the monitored general emotions. According to preferred embodiment of the present invention, the speech specimens are provided over the telephone to the voice analyzer and the report of the subject's emotional state includes a "love detection" report based on the individual's sub-conscious emotional state.