We have developed and are exploiting cutting-edge technologies for the automated evaluation of the singing voice. Our technologies allow personalisation while simultaneously addressing music style-specific aspects. These technologies are already making their way to Riyaz, an android application to learn and improve singing. Riyaz users’ rave reviews testify the value of such technologies.

 

Audio content extraction

Our core audio analysis technology used in Riyaz extracts the relevant audio characteristics of a recording needed to study the quality of a singing voice. It is built on top of the Essentia audio library and it is based on years of research.

 

Music performance analysis

The music analysis technology used in Riyaz combines our core audio analysis methods with sophisticated machine learning techniques that take advantage of the huge number of recordings that we are gathering from our Riyaz users. Using the recordings and annotations given by our expert musicians we have been able to develop an algorithm that can automatically compare the user recordings with the data from the experts, giving a feed back to the user that emulates a human guide/teacher in all the aspects of learning.

 

Singing evaluation

We are on a mission to remove all the barriers for facilitating music education on scale and quality. With our core audio and music technologies we have developed the complete singing evaluation system being used in Riyaz. We have developed a friendly and intuitive system to interact with the users and give them relevant musical feed back including pitch accuracy, timing, stability, singing voice quality besides many others.