Transcripts

The starting point of accessible multimedia, regardless of the media type, is having a transcript. As defined by the World Wide Web Consortium (W3C) Web Content Accessibility Guidelines 1.0 (WCAG):

Audio presentations, such as a spoken passage or song, need a transcript.
Video presentations, such as a movie or animation with sound, need to have an audio transcript that is synchronized with what is currently happening in the programming (a process called captioning).

(Source: W3C Core Techniques of the WCAG 1.0 , 13-16: Audio and Video (browse to website) ).

Getting the Transcript

There are various methods and considerations for getting the transcript which include:

Video Capture
A majority of commercial multimedia, such as television programs and movies, already have the transcript embedded in the video signal. You can usually request a transcript, but the recommended solution is to buy a video capture card that can capture transcripts as well. Usually when using one of these cards, while the video is being captured, the transcript is dumped into a separate text file. These cards are relatively inexpensive and regardless of budget or the amount of captioning, are also a worthwhile investment.
Speech-to-Text Technology
If you are doing a home video or a lecture without the luxury of already having it transcribed, you may choose to use some kind of speech-to-text software, which listens to and transcribes human speech into text. There are claims that speech-to-text software has voice recognition rates up to 98 percent but generally the average is about 85 percent because:
- The process of training the software to recognize your voice builds over time, requiring reading numerous paragraphs to it multiple times.
- A high percentage of voice recognition is only attained under ideal conditions, where the user sits in a background noise-free room with a headset on, speaking in relatively clear sentences. However, most videos have background noise, including tape hiss, music, side conversations, and other noises.
- There is fluctuating consistency and clarity in your voice throughout the day.
Note: Even if you get 80 percent recognition using speech-to-text technology, you still have to go back and correct the transcript.
Text Editor By Hand
Until speech-to-text technology improves, transcribing by hand using a text editor, such as NotePad or vi, works better for getting a transcript. With this approach, there is no time wasted in "researching" or training the speech-to-text software and correcting the errors. The transcriber can begin immediately, or in parallel with other transcribers working on different sections to transcribe a video segment in 1/n time, with n representing the number of transcribers. Depending on the amount of dialog and the typing speed of a transcriber, as much as 10 minutes of video or as little as 2 - 3 minutes of video could be transcribed in 1 hour.

Creating A Transcript

Things to Include in the Transcript

When creating transcript, it is important to transcribe spoken words as well as other things that convey information. For example, a phone ringing in the video should be indicated in a transcript such as [phone ringing]. Transcribing things like this helps to clarify the situation, because it tells the viewer the phone has rung rather than the person is making a call. Pick-up cues or off screen things that convey context should also be transcribed; to illustrate, consider a horror movie. Seeing Jamie Lee Curtis turn around in shock obviously shows she is afraid of something, but it helps to know if she heard a scream, a glass break somewhere in the house, or a creepy voice saying, "Jay Jay Jay Jay, Son Son Son Son".

Size of Sentences in Transcripts

The size of transcribed sentences can vary in length, and depends directly on the amount of room available for the captioning as well as the size of the font. A general rule is to keep things in proportion to one another and usually display only one or two sentences at a time. If you want larger fonts, create more space for the captioning. However, consider that larger fonts with a small captioning space mean less text on the screen; the larger the font size the more rapidly the text on the screen changes. Alternately, smaller fonts with a large captioning space mean more text on the screen; the smaller the font size, the more difficult it is to read and the greater the likelihood of getting lost in the transcript, not knowing which text currently relates to what is occurring in the video.

The next page will discuss the general aspects of captioning video.

Previous Page
Page 3:
Basic Concepts

Top of Page arrow up

Page 4

Page 5:
Captioning