I am looking for a program or web service that can output the speaker distribution given an audio file containing a discussion between different interlocutors.

Example of speaker distribution:

  • speaker 1 talked 70% of the time
  • speaker 2 talked 20% of the time
  • speaker 3 talked 10% of the time

For example, uberconference.com gives the following summary (but it cannot take an audio file as input: it is simply based on the speaker source):

enter image description here

Any price, license or operating system is fine.

1 Answer 1


One possibility is to use BOB.

Bob is a signal-processing and machine learning toolbox originally developed by the Biometrics Security and Privacy Group, the Biosignal Processing Group, and the Research and Development Engineers at Idiap, in Switzerland. Bob is primarily developed through Gitlab.

It includes Speaker Recognition Alogorithms

  • Free, Gatis & Open Source (MIT licence)
  • Potentially just about any Platform/OS but may have some problems under Windows so may need to be installed as a Docker image (see here).
  • Written in C & Python
  • Can accept file inputs

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.