GGoingGlobalingotoglobal.hashnode.dev·5d ago · 6 min readBuilding a Speaker-Aware Audio Pipeline: Diarization, Separation, and Overlapping SpeechWhen I started building Seply, I assumed that identifying speakers and creating separate speaker tracks were nearly the same problem. If an API knows that Speaker A talks from 00:05 to 00:12 and Speak00