Decoding the Human Intonation

When people speak, their voice generates a collection of sound waves of various frequencies and intensities.
The choice of component frequencies and intensities is both instinctive and involuntarily.
Drs. Levanon and Lossos hypothesized that the choice of frequencies and intensities used in speech are reflective of a speaker’s personality and his emotional state.
To test their theory, they conducted personality assessments of many individuals and compared the resulting profiles with voice patterns.
And indeed, after analyzing thousands of digitized voice samples, during eleven years of research, they were able to identify physical speech variables that correlate with personality types.
Additionally, they also discovered physical voice parameters that are reflective of discrete emotional attitudes. Building on their work, eXaudios Technologies has developed computer programs that can analyze voice and determine both the instinctive-emotional personality of the speaker and his emotional attitude.
The computational process powering these programs is based on established links between emotional attitudes and physically measurable parameters in voice. The system is so accurate, in fact, that a mere 30-second sound bite is sufficient for conducting the analysis.
In addition, the two inventors also made the serendipitous discovery that some physical conditions, e.g. reading dyslexia, Parkinson’s disease, heart ailments and childhood autism, all left unique “fingerprints” in a speaker’s voice.
This discovery led to a number of joint research projects, with leading medical research institutes in Israel, and the potential for future diagnostic devices that will be based on voice analysis.
As an example, initial results from a study on autistic children (presented at an international conference in Seattle WA in the spring of 2007) suggest that voice analysis could be an effective tool in the early diagnosis of autism in infants.