Audio Icons

Getting the recorded prompts sounding right is critical, but there are other sounds that can be used in an application besides the spoken word. Audio icons are short sounds ”for example, a blip, a swoosh, or a set of tones ”that may or may not have words spoken over them. Some well-known examples are the three tones associated with NBC radio and TV, the opening chord that sounds when a Macintosh computer is started up, or the four-note phrase that accompanies "Intel Inside" in all Intel radio and TV spots. These usually short sounds can add much to the feel and user experience of an application.

What Audio Icons Can Do

Audio icons serve a number of purposes. When callers hear a welcoming phrase and an audio icon (for example, "Welcome to FedEx RateFinder <audio icon>"), they are immediately informed that (a) they are not speaking to a live person (nobody I know picks up the phone, says "Hello," and plays a music chord); and (b) this is a well-produced, professional-sounding system.

But that's not all. If the speech system is a replacement for an older (perhaps touchtone) system, the audio icon alerts the caller to that fact ”without subjecting them to annoyingly long-winded statements like "Please listen to all the menu selections before choosing one, as our options have changed." (I've been known to call back into those systems to check if their options have changed. Usually I find that they keep that phrase in there for months or longer.)

Perhaps most importantly, audio icons can also aid in the branding of a service, product, or company. By repeating a unique sound used by the company in its commercials (for example, the Sprint "pin drop"), the audio icon conjures up a vision of the brand in the mind of the caller, reinforcing and strengthening that brand identification. Audio icons can even help create the feel of an application. For example, if we were to create a system for a manufacturer of large machines, we might choose to use a set of industrial sounds to underscore the identity of the company.

Audio icons can also act as anchors, helping users locate where they are in an application. If we created an application with a main menu, we could program the system to play an audio icon whenever a caller was at that place. Eventually, as callers learn to associate the audio icon with the main menu, the system could even stop saying "Main menu," using only the audio icon to orient the user.

And it doesn't have to stop there. For applications that are going to be used repeatedly by the same caller, designers can employ a series of audio icons throughout an application to remind callers where they are. When callers hear the same sound repeated in the same place in an application, they can quickly learn to associate the two ”a helpful method to let callers know where they are in an application without having to listen to lengthy spoken text.

Some Limitations of Audio Icons

One caveat. We shouldn't go overboard using audio icons. Audio icons can enhance the brand and improve the user experience ”but add too many and the entire experience becomes unpalatable. If an audio icon is used at the beginning of a prompt, taking up a couple of seconds before the spoken prompt starts, and if that audio icon doesn't serve as an anchor point to orient the caller, then eventually the caller will become frustrated with the system because the system is noticeably taking up the caller's time but not providing value to the caller. Also, if there are a lot of audio icons sprinkled throughout the application, the interruptions can erode the conversational flow of the system.

The only real limit to the use of audio icons is a technological one. The conventional telephone transmission system is unable to carry bright, high-frequency and deep, low-frequency tones. In fact, the telephone frequency range is only 300 Hz to 3,400 Hz, while the range most people can hear is approximately 30 Hz to 20,000 Hz. I once had an audio-icon producer make a set of audio icons. One that I particularly liked was a cartoon sound effect to accompany the wave of a magic wand. I thought it would be a great sound to be played underneath a "transferring to representative" message. However, when I tried it out on a real system, I couldn't hear a thing. The sound's frequencies exceeded the range of the telephone.

You can prevent this embarrassment by avoiding very high or very low tones. But just to be sure that any audio icon will sound right in an application, you should always listen to the audio icon as it will sound on the phone. It's easy to create that listening experience by using a simple sound-editing program to emulate how the audio would sound over a phone line, since the tiny speaker in the phone can't reproduce sounds like a good stereo system can.

Most audio files are made and delivered to the system designer in high fidelity or near CD quality sound (either 44.1 Hz or 22.05 Hz). Using a sound-editing program (several of which are available via free Web downloads), the designer can down sample the sound to 8,000 Hz and play it. If the audio quality of the down-sampled icon sounds awful , don't be alarmed (yet.) Often, audio that sounds pretty bad via speakers or headphones sounds surprisingly "OK" over the phone. This time the reason isn't technical; it's because our ears are so accustomed to hearing low-fidelity sound over the phone that our brains actually fill in the missing information (for example, the crisp s sound in "Christmas"). Result? It appears to sound better than it actually is.