We're working on a project using hundreds of prerecorded phone calls in separate audio files. Each call has two people: a customer and a customer service rep (CSR). We need to eliminate the CSR's voice so we have only the caller's voice in the audio waveform. These are mono audio files with no L/R channel separation, and the CSR will change in each file.
Is there a way to isolate the voice of one speaker and eliminate that from the waveform? I know this can done with noise reduction, but not sure if can go so far as to identify vocal differences or how to go about that.
Other than manually marking in/outs of the CSR and dropping volume to zero or cutting out, any ideas of a better way to handle this?
Any ideas or leads for other software solutions is most appreciated.
- Eric Klassen
I don't know if I understand you completely. its just simple editing. audition has markers and can auto detect blank audio spots, but it wouldn't know who's talking. Do they ever talk at the same time? If they do, all I can think of is try rx-6 advanced's dialogue isolate effect. In that case, removing the noise will be the last of your problems.