Usage for TSE for a particular speaker

[mukherjeesougata-eros]

I would like to extract the voice of actor Shah Rukh Khan from an entire movie dataset. I can provide 2–3 clean speech audio samples of his voice to serve as the target cue. Could you please suggest the appropriate command line, code snippet, or recipe to use for extracting Shah Rukh Khan's speech from the full movie dataset?

[mukherjeesougata-eros]

@KyleZhang1118 can you kindly please reply to this issue?

[mukherjeesougata-eros]

@wsstriving can you kindly please reply to this issue?

[KyleZhang1118]

If you want to use TSE for speech extraction of a specific speaker, I recommend first slicing the speech clips. Segments of ten seconds or less are ideal. Then, refer to wesep.cli to call the provided pre-trained models. Additionally, I recommend using the speaker diarization tool to identify the target speaker's speech segments before performing extraction.

[mukherjeesougata-eros]

May I know if we can use the licence for commercial purpose?