CONSUMER AND GOVERNMENTAL AFFAIRS BUREAU SEEKS COMMENT ON PETITION FOR DECLARATORY RULING AND PETITION FOR RULEMAKING FILED BY TELECOMMUNICATIONS FOR THE DEAF AND HARD OF HEARING ET AL.

ON LIVE CLOSED CAPTIONING QUALITY METRICS AND THE USE OF AUTOMATIC SPEECH RECOGNITION TECHNOLOGIES

CG Docket No. 05-231
RM-11848

Notice of Inquiry

Frequently, I've experienced poor-quality captions on live programming. Automatic Speech Recognition (ASR) is not reliable because ASR could not adapt anything such as background noises, accents, or other conditions, thus ASR could not produce contextual information. I've missed many important information. It's so easy to misunderstand or misinterpret because information was left out. Inaccurate captions also lead to miscomprehension of content and misinformation. Examples of misinformation include clues, wrong assumptions, anger, false rumors, insults and pranks, while examples of more deliberate disinformation include wrong content such as confusion, hoaxes and propaganda. It is important for the broadcasters to reach 100% accuracy rate for live captioning.