divid one voice from the other

Hi, I am an electrical engineering student and I have a project in signal prosessing course.
Lets say that I recorded 2 people talking. the recording is perfect and in high quality.
How do I divid one voice that I would like to differ from the other voice or the environmental noise?
I need guiding lines how to know what belong to who.
It can be solved mathematically, by logic, or whatever…
Tanks a lot

You can plan on that not working.

Audacity can’t separate a mixed performance into individual voices, instruments or sounds. It’s particularly a problem if you need to split two of the same thing, such as two voices. There is one arithmetic “trick” where we can sometimes manage a singer in the middle of a stereo song, but it fails more often that it succeeds.

Effect > Vocal Reduction and Isolation.

There is an experimental program called ISSE. It’s Alpha software, so it was never issued as a complete program.

http://isse.sourceforge.net/

https://forum.audacityteam.org/t/is-it-possible-to-remove-trumpet-from-organ-piece-of-music/47281/4

Koz

Hi, I am an electrical engineering student and I have a project in signal prosessing course.
Lets say that I recorded 2 people talking. the recording is perfect and in high quality.
How do I divid one voice that I would like to differ from the other voice or the environmental noise?
I need guiding lines how to know what belong to who.
It can be solved mathematically, by logic, or whatever…
Tanks a lot

It’s complicated. Try searching Google for “Source separation”.
Here’s the Wikipedia page: https://en.wikipedia.org/wiki/Source_separation
and a paper (pdf format) called: “A Very Short Introduction to Blind Source Separation - LISIC”
www-lisic.univ-littoral.fr/~puigt/LectureNotes/Intro_BSS/Lecture_BSS.pdf

Topics merged.
Please avoid double posting.