508 Accessibility and ELearning

Top 10 Accessibility design rules for e-learning

1. Images: Use alt text to describe all static elements, such as graphics, icons, graphic text.
2. Multimedia: Provide captioning and transcripts of audio, and descriptions of video.
3. Hypertext links: Use text that makes sense out of context. For example, avoid ‘click here’.
4. Page organization: Use headings, lists, and consistent structure. Use CSS for layout and style.
5. Graphs & charts: Summarize or use the longdesc attribute.
6. Frames: Use the noframes element and meaningful titles.
7. Tables: Make line-by-line reading sensible – summarize.
8. Questions: Avoid ‘difficult’ screen types, such as drag and drop.
9. Punctuation: Meet requirements of screen readers, e.g. full stops at the end of all bullet points.
10. Keyboard: Ensure clickable elements can be controlled via keyboard, e.g. Tab cycles and Enter.

For more detail, visit www.section508.gov.

WBT Audio Strategies and TextAloud Voice Synthesis

One of my clients wants to use audio in their Web-based training. Audio can improve the accessibility of the self-study, whether WBT or CDROM, and spices up a program. The program should have a visual element allowing learners to set their own pace. Also, a learner should be able to complete and continue through a screen before the audio is completed. It's about learner control.

I've used different audio strategies that include professional narrators, directly capturing subject matter experts (SMEs) talking about their subjects, and self-recording. The pros are expensive, but can actually save you money with their experience and efficiency--not to mention the slicker product. Sometimes we can get too slick. Recording SMEs in the field, on the job site, or over a conference set-up can be cheaper, and offers content validity with working audiences. This is one of our "Blair Witch" rapid course development techniques.

Another technique is voice synthesis. I recently tested a popular voice-to-text program called TextAloud from
Next-Up.com.

I was interested in using it to speed up development for an audio component in the WBT portions of a course. The program can be used to read email, stocks, news, documents, or in our case teaching scripts.

One very nice feature of the program is that it can reside in the background. If you come across some text, say in a PowerPoint or Word document and try to copy it, TextAloud pops up and asks if you want to include this for voicing. Once you have the text you want for audio, you save the piece as an "article." This basically creates a Wave file to use in your programming.

The default voice is named Mary. She sounds very human, but is too robot-like for anything longer than a sentence. I reviewed, some higher quality voices that are compatible with the program. These are very credible synthesized voices from other companies such as AT&T, NeoSpeech, and Acapela. The NeoSpeech voices really stood out, quality-wise, to me. They're quite affordable at $35 each and I recommend buying and alternating a female and male voice. The Next-Up.com voice engine (to run these voices) is only $24.95. This approximately $100 is considerably cheaper than a professional narrator or consultant.

My next step was to bring the audio to my WBT. I used Breeze Presenter and Meeting, but the same approach would apply to similar platforms such as Articulate. These programs convert standard PowerPoint to Flash files (SWFs). These work very well on the Web. By the way, these PowerPoints don't have to be the classic four-line bullet points that you see in most meetings. Think of them presenting material as an interactive book on a person's screen.

The downside is that adding audio to a moderately lengthy WBT text section requires bandwidth and disc space. One page of synthesized audio at 16 kHz produced a 4.3 Meg Wave file. Converting the file to an MP3 took it down to 495kb. I created a 10 slide PowerPoint using this small file on every page. Audio compression was good, but the published Breeze folder was 3.5 Meg. Without audio, the folder size was 467kb. Quite a difference! The decision to use this approach on our project is now in the hands of my client. I let you know in a future blog how it goes. By the way, sometime when I have time, I might do a real product showdown, but this quick exploration answered out questions. Good luck in your own audio WBT work.