ASR (Automatic Speech Recognition)

Automatic Speech Recognition (ASR) is the technology that transcribes spoken language into text. It powers captioning, voice assistants, search, and analytics. ASR accuracy drops sharply when background music or noise is present, which is why isolating a clean speech or dialogue stem before transcription significantly improves results. Accuracy is commonly measured with Word Error Rate (WER).