Visualize differences between voices across demographics

This dataset combines voices from a range of open-source libraries, featuring speakers from around the world recorded in both controlled and uncontrolled environments. Speakers may be reading a script or speaking spontaneously. The majority of demographic information is self-reported by the individual speaker.

These are not separate research metrics. Each excerpt is analyzed using the same system applied to your recordings in the app. The metrics are designed to help you understand your own voice and follow a structured plan to change specific aspects of it toward your desired outcome. Each metric is explained in detail on the page linked below.

This dataset shows that the average female voice drops by over 20 Hz over the course of their lifespan, while the average male voice remains relatively stable. This is an accurate reading of the data, but it does not mean that every female voice will inevitably fall with age. The data shows an association between age and vocal pitch; it does not, by itself, explain what causes that change.

The same is true of other associations in the data. Averages obscure individual variation and can lead to false assumptions in data. Most female voices fall within a certain range; that does not, however, indicate that a voice outside of those typical parameters will sound any less feminine to a listener.