Live broadcast captioning is a highly specialized process designed to provide real-time transcription of live television programming, including sports events, news broadcasts, and award ceremonies. This type of captioning is executed by skilled professionals known as real-time captioners, who utilize specialized stenographic (steno) machines to swiftly and accurately transcribe spoken words into written text.

How It Works

• Stenographic Machines: Real-time captioners use steno machines, which allow them to type words phonetically rather than letter by letter. This enables them to transcribe at speeds of over 250 words per minute.
• Software Integration: The steno machine is connected to specialized software that translates the phonetic shorthand into readable English text. This software includes a personal dictionary built by the captioner to ensure accurate translation of frequently used terms.
• Transmission to Broadcast Stations: The transcribed text is transmitted to the broadcast station through various methods such as phone lines, modems, or IP connections.
• Encoding into Broadcast Signal: The text is encoded into the television broadcast signal. As the program airs, the captions are displayed on the screen, providing immediate access to the spoken content for viewers.

Challenges and Expertise of Human Captioning

Mastering steno machines requires 2-5 years of training. Our captioners achieve up to 99.5% accuracy at speeds over 250 words per minute, specializing in content-specific terminology for maximum accuracy.

Due to the live nature of this work, captioners do not have the luxury to pause or correct mistakes. They must maintain high accuracy in real-time, dealing with varying speech patterns, accents, and speeds. The complexity is heightened by the need to also transcribe background noises and other audio cues essential for context.

Is Automated Speech Recognition (ASR) Captioning a Solution for Broadcast Television?

The FCC generally does not permit automated captioning for broadcast television due to its accuracy limitations. The FCC requires that captions meet high accuracy standards, synchronicity, completeness, and proper placement. While automated captioning systems can achieve around 90% accuracy, this is typically insufficient for meeting the FCC's stringent requirements for live broadcast captioning.

The FCC's guidelines emphasize that captions should match the spoken words in the dialogue and convey background noises and other sounds to the fullest extent possible. They must also be displayed at a speed that can be read by viewers and should not lag far behind the spoken words. The current technology for automated captioning often falls short of these standards, leading to potential issues with timing, contextual errors, and missing content.

These regulations ensure that captions provide a reliable and accessible viewing experience for all audiences, particularly those who are deaf or hard of hearing. Despite advancements in automated captioning technology, human real-time captioners remain essential to meet these rigorous standards due to their ability to adapt and respond to live, dynamic content.

Therefore, while automated captioning can be useful for certain applications, it is generally not acceptable for broadcast television where high accuracy and reliability are mandated. Human captioners are preferred for live broadcasts to ensure compliance with FCC standards.