Accessible Electronic Content > Guidelines > Audio Content


Link to Accessible Electronic Content Homepage    Audio Content

If you have questions about these guidelines or the accessibility of any audio content, contact ITS Web Services.

According to the W3C's Web Content Accessibility Guidelines 2.0 (WCAG), there must be a text equivalent for all non-text content. A transcript of the audio, as a text document, must be produced and then incorporated into the web page or another electronic resource.

Transcription Options

There are several different options for transcribing. Which ever one you choose, remember that a content specialist should still proofread the transcription once completed.

  1. Outsourcing
  2. Transcribing with voice recognition software
  3. Start-Stop Transcription System
  4. Synching audio and text using MAGpie 1.0 and MAGpie 2.0

Integrating the Transcription

  1. If the audio clips are short, then the transcription can be added to the page. For example, if you have slides that accompany an audio presentation, incorporating the transcript into that page will make it easier to view the graphic and transcription at the same time. Example Slide Show.

  2. If the audio is a long continuous presentation or story, with no accompanying visuals, then it would be more suitable to add a link to the transcript. This allows the user the option of reading the content at his or her own pace, as well as converting it to Braille or displaying it in a larger font or different colors.

  3. Under special circumstances, you may want the text to synchronize with the sound. If you do, MAGpie 1.0 or MAGpie 2.0 would be the best tool to use in order to achieve this. If you decide to use this method, the media player must be incorporated into your HTML page as a standalone player, as opposed to an embedded player. Screen readers cannot recognize embedded players. See Synching audio with text using MAGpie 1.0 and MAGpie 2.0.

Recommendations

The transcription solutions that best suit you will depend on your budget and the amount of time you can devote to the project. The easiest and simplest way to integrate that transcript is to incorporate it into your content as a link or as part of the HTML page or other electronic resource.

Samples: Outsourcing

Type-Thing Services http://www.type-thing.com/web.html

We Type 4 U http://www.wetype4u.com/

Recommendations: Transcribing with voice recognition software

Voice recognition software enables people to transcribe by simply using their voice. Building a large vocal library requires training the software. More than one person can train the software, but each person will have to train and build his or her own vocal library.

Required Hardware, Software, and/or Other Tools:

Dragon Naturally Speaking
Windows 98,2000, NT, XP, ME: 500 MHz, 128 MB

Pro ViaVoice
Windows 2000, ME: 300 MHz, 96 MB
Windows XP: patch CD

Advanced ViaVoice
Windows 98: 300 MHz 64 MB
Windows XP: 300 MHz 192 MB, patch CD
Windows ME: 600 MHz 64 MB

Required Skills:

Precise, well-enunciated speaking voice. Good listening skills are also needed if you are transcribing someone else's audio.

Required Time Needed for Development:

These programs take time to train in order for them to become familiar with your voice. Mistakes, as in interpreting what you say incorrectly, become fewer and fewer the more you train. Training could take up to 3 months to achieve 99% accuracy.

Degree Compliance Can Be Part of Initial Production:

This process could be a part of the initial production, but will be time consuming in the beginning. The editing process may take longer in the initial stages of training the software.

Special Problems & Issues:

  1. When using voice recognition software, everything has to be spoken. For example periods, quotes, commas, have to be said.
  2. Inconsistencies in the transcription will require editing. For example, saying "divided by" may be written using the symbol, /, or the program may write out "divided by". When it comes across a word which is not recognized in its vocabulary that is spoken frequently, it will show a pop up box guiding you through steps to add that word to its vocabulary and your voice profile. Editing in this manner, as you go along, increases the speed and accuracy through the rest of the transcription and for transcriptions to follow.

Recommendations: Start-Stop Transcription System

This is a foot pedal transcription system. It works with PC's and makes it possible to play/pause, fast forward and rewind in a codec player with a foot pedal. This will cut down considerably on transcription time if you decide to do it "in house".

Required Hardware, Software, and/or Other Tools:

  1. Start-Stop UNIVERSAL Transcription System
  2. Start-Stop PEDAL POWER for RealPlayer™
  3. Windows 95, 98, ME, 2000, NT 4.0 and XP™ compatible

Required Skills:

Typing, listening, and familiarity of vocabulary used in the presentation.

Required Time Needed for Development:

Depends on how long the audio track/s are and how many words the speaker speaks per minute along with your typing speed. See test.

Degree Compliance Can Be Part of Initial Production:

This does not have to be a part of the initial production. If the audio files are available, that is all that is required to transcribe.

Special Problems & Issues:

The pedal system can decrease transcription time by about 60% as shown in the example below.

Examples:

Test using Start/Stop Transcription System

Recommended Resources:

Official Start-Stop Site: You can purchase the software here.
http://www.startstop.com/

Recommendations: Synching audio and text using MAGpie (Media Access Generator)

MAGpie is a freeware captioning tool for video intended for the web, but is not exclusive to video. You can caption your audio files as well. It outputs the necessary files to play in RealPlayer (SMIL, .rt), QuickTime (SMIL, .txt), or Windows Player (SAMI). You don't have to download both versions of the program. Which one you will need depends on the type of audio files you have. MAGpie 1.0 has a Windows Media internal player, and MAGpie 2.0 has a QuickTime and GRiNS internal player (GRiNS uses the RealPlayer engine). The file formats for each player are listed below.

File Formats Supported by Windows Media Player :

File formats that play in QuickTime:

MOV, AVI, DV, MPEG, AIFF, WAV, MP3, SWF (Flash)

File formats that play in GRiNS:

GRiNS uses the RealPlayer engine, so it will play the same file types as RealPlayer does.
RM, RA, MP3, SWF (Flash), SMIL/SMI, WAV, AIFF**, MPG, MPEG, AVI

Required Hardware, Software, and/or Other Tools:

For more information see the MAGpie website.

Costs:

Free

Required Skills:

Good typing and organization skills are useful. The skills required to learn the program are comparable to the skills required for Microsoft Word.

Required Time Needed for Development:

Learning MAGpie might take a couple of hours. The captioning part is the most time consuming area. If you outsource your transcriptions, the time it takes to caption obviously will be cut down. Without a transcript, a 5-minute video takes approximately 2 hours to caption, transcribing as you go along within MAGpie. With a transcript, the time was cut almost in half at approximately 1.25 hours. The whole captioning process will speed up once you become more familiar with how it works.

Degree Compliance Can Be Part of Initial Production:

This process can be a part of the initial production.

Special Problems & Issues:

  1. Your media cannot be in use by any other application in order to output a SMIL or SAMI file.
  2. Windows Media Player uses SAMI instead of SMIL. It works a little differently. Unlike SMIL, you must open up the base media in the Windows Player in order to view the captions and not the SAMI file. Windows Media Player has an internal mechanism that automatically looks for a SAMI file that is associated with that media.

Link to UNC-Chapel Hill Homepage

© 2002, The University of North Carolina at Chapel Hill ITS.
Last revised: January 30, 20075.