Theory behind MAVIS

It has been appreciated for centuries that most peoples’ voices sound different and the ability of the human ear to distinguish the differences of one voice from another (irrespective of the words spoken) is substantial.

When we sub consciously form speech the process starts with our lungs, forcing air initially through our Vocal Chords, which resonate and generate a series of Fundamental and Harmonic frequencies. These are then further influenced by the Articulation and Nasal tracts the latter, in very simplistic form, being responsible for additional changes to the original frequencies generated by the Vocal Chords (e.g.: changes such as colds and coughs). This then becomes the sound that people know as you!

MAVIS embraces the above natural process and reliably processes a variable but “unique” identifier that will not only verify an individual but also discriminate between other speakers who may want to try and impersonate someone or use the same password.

Just think if we could develop an “electronic ear” the security benefits in today’s uncertain and technological world would be substantial.

Although speech is the most natural and efficient way of communicating it has proven to be the most difficult to process reliably. This is because of the many cultural, physiological, environmental, and physical factors that have to be taken into account. Things like:

In addition to all this, such a system needs to have the ability to be seamlessly integrated into existing systems. Most important of all, it needs to be a system that people and organisations can use without having to re learn or change their behaviour. We call this Continuous Innovation.

Uniquely, MAVIS is such a system.