Async transcription accuracy on hard audio: noisy call centers, overlapping speakers, and filler words

Analysis of async transcription performance on challenging audio scenarios.