Live Closed Captioning

For both live video captioning and text streaming, you need a way to input the text in real time. It is not possible to type text on the computer keyboard as fast as people speak. There are two common practices - Steno typing  and using a speech recognition software. For more details, click  on Steno Machine vs. Speech Recognition.

Speech Recognition Based Realtime Captioning

• Simple to learn and use
• Cost-effective
• Helps comply with the FCC captioning mandate which took effect 1/1/06

The CaptionMaker-Live and MacCaption-Live are easy to use. First train speech recognition software like Naturally Speaking (Win) or MacSpeech (Mac) to recognize your speech. Next, connect your computer via a serial cable to a closed caption encoder which also receives a feed of your video. Then simply speak and your words are automatically turned into captions. To have the speech of multiple speakers appear as captions use a “shadow speaker”.  

High Accuracy	Shadow Speaker
A new generation of speech recognition software, such as Naturally Speaking (win) and MacSpeech (Mac) provide accuracy rates in the 90-95% range. The CaptionMaker-Live and MacCaption-Live software interface to this off-the-shelf speech recognition software to produce realtime captions.	Single and multiple speakers can be captioned using a shadow speaker A single speaker simply speaks into a traditional microphone. For multiple speakers a “shadow speaker” is used. A shadow speaker is a person who first trains the speech recognition software to recognize their speech. The shadow speaker then repeats everything being said into a microphone.

Television Use 

Sports and Weather Captioning
If your newsroom system automatically captions the teleprompter text, you are probably are not captioning sports and weather, since these segments usually are not on teleprompter. CaptionMaker-Live and MacCaption-Live can automatically caption these segments if you have your sports and weather talent train the speech recognition software to their speech pattern. 

Emergency Announcement Captioning
If you have to go on the air quickly with an emergency announcement, CaptionMaker-Live and MacCaption-Live can help fulfill your captioning needs. Have two or three employees on each shift train the speech recognition software to their speech pattern. When the talent (who has not trained the speech recognition software to their speech pattern) goes on air on short notice, have one of these employees act as a shadow speaker to caption the talent’s speech.

Classroom Use
Currently, many schools use speech-to-text note taking systems which display text on a screen in front of the student. When reading the text, the student is unable to watch the instructor without taking their eyes off the screen. This setup is far from ideal, and clearly puts the student at a disadvantage. Studies have shown that visual perception plays a big part in a student's ability to learn new material. CaptionMaker-Live and MacCaption-Live allow the student to see the instructor and any demonstration they may be conducting via a video monitor while simultaneously reading the text of what is being spoken. The text produced from the instructor's speech is automatically displayed as traditional captions overlaying the video.

Meeting Use
CaptionMaker-Live and MacCaption-Live make it practical to caption almost any meeting because it is so simple to use. If there is only one speaker, that person trains the software to their speech prior to the meeting. During the meeting they speak into a microphone connected to the captioning system. If there are multiple speakers, a shadow speaker is used (see sidebar below). It takes practice to get used to being a shadow speaker, but most people master it with little difficulty. This contrasts with traditional realtime captioning systems which have high labor costs (e.g. for stenographers) that often makes them prohibitively expensive to use.

Helps you comply with FCC, ADA and Section 508 regulations

---

System Diagrams
.

Television Use	Classroom Use

Note: There are many setup options. For example, the video can be open captioned, displayed on a TV monitor on a student's desk and/or go to a record deck. How it works The talent or shadow speaker (the person that repeats all that is said into a microphone) first trains the ViaVoice speech recognition software to recognize their speech. This takes a few hours, and only needs to be done once. In the studio, either the talent wears a microphone, or if there is rapid dialogue between the talent, the shadow speaker repeats everything said into a microphone. In any case, the microphone is attached to the computer running  ViaVoice speech recognition software, which converts their speech into text. The text is then automatically fed into to CaptionMaker-Live and MacCaption-Live software which transform the text into caption format. This caption formatted text is then sent to the closed caption encoder via a standard serial cable. Simultaneously, the broadcast video is sent to the closed caption encoder. The encoder then physically places the captions onto the video in real time. The video out of the encoder (which contains the captioned video) is then broadcast out.