separate voices of different people

[RICARDO23969030tkks]

hello

does anyone knows the best automatically way of separating 2 voices in conversation so that i can get a final track in which i only hear one of the two voices

thnak you very much

[SteveG_AudioMasters_]

Are they on one track, and about at the same level? If they are, then there's no automatic way to do this at all. In order to answer your question properly, we'd need to know what the basic circumstances of your recording are...

[RICARDO23969030tkks]

hello, thank you very much, the question is if in an audio you can listen to the voice of one person and eliminate that of another so that the audio is shorter (the two voices are separated in time but it is only one track), even if it is me It would be worth it to put automatic marks for me to delete manually or to identify the frequencies

[SteveG_AudioMasters_]

You haven't answered the questions.

[RICARDO23969030tkks]

sorry what is the data needed?

[SteveG_AudioMasters_]

Is it a stereo recording, with the speakers partly separated, or mono? Are the levels of both speakers similar? We need to know basically what you've got to work with, otherwise we can't tell you definitively what it's possible to do. Whilst there is no automatic way to separate two voices, sometimes you can get a lot closer if there are other audible differences you can make use of.

[RICARDO23969030tkks]

hello thanks again

the audio is mono 44100 hz and 32bits floating compressed with quicktime

[RICARDO23969030tkks]

the level of both speakers one is a little louder

[SteveG_AudioMasters_]

If it's mono, then I'm afraid that the only way you can separate the voices is by manual editing. There is no automated process available on any DAW I'm aware of - primarily because none of them have ears, which are a primary requirement for this type of operation.

[RICARDO23969030tkks]

thank you, in that case what i can do is doing in premiere speech to text and put ccs and try to see the video with the text to go faster, the problem was that i onlyt need to extract the info of one of the 2 voices and i wanted to go faster, thank you very much for all the answers

[Artfundi_Multimedia]

Spectral Layers Pro by Steinberg does that with its Unmix Vocals module. You select the reference, and it detects where the rest is and creates a separate track from that. You could probably reverse engineer this with Audition. 

[Unmixing30325228z2ho]

@RICARDO23969030tkks 

You mean something like THIS!(Tap Here/Click Here) 

[Laurence33709984xkje]

Yes like that!!

[Unmixing30325228z2ho]

Well! If it's something like THIS then this has to be done manually. If anyone wants something like THIS done or something more complex (like unmixing voices in a podcast) unmixed then reach out to me.  "Unmixing @outlook . C o m"

[Dora356367934bdf]

Is there a way to do it on a home video recorded via phone?