4 ms·
Can you not separate into two phases? Speech separation to get source per speaker, and then whisper on each in isolation (maybe interlacing prompts)?
by graderjs 4y ago
Can you not separate into two phases? Speech separation to get source per speaker, and then whisper on each in isolation (maybe interlacing prompts)?