What are good reasons for and against using audio for feedback?

Question

If a device like a printer has a keyboard for users to interact with it (keyboard as well as navigation), does it help users to hear audio every time they use the keyboard for text input? Remember this is not a touch interface, so users can't touch the screen or have an onscreen keyboard.

What are good reasons for and against using audio?

Also it might be relevant to mention that the hard keyboard produces tactile feedback and normal sounds when users press down a key.

Answer 1

If the keyboard is giving tactile feedback and makes its own sounds, then I see no need for trying to better nature. Audio would be needed only if your application is not responsive, i.e. when nothing changes on screen within the 100ms of the causality barrier, but you still want to tell the user that you got the input. Remember, though, that's a only an ugly patch. You'd better spend energies to make the application responsive to keyboard events.

UPDATE: references for the causality barrier: The first one was Michotte, A.(1946) The Perception of Causality. New York Basic Books, 1963. Originally published in French, "La Perception de la causalite", Publications Universitaires de Louvain, 1946.

The next best is Card, Moran, and Newell's "The model human processor: An engineering model of human performance", available at Xerox PARC: http://www2.parc.com/istl/groups/uir/publications/items/UIR-1986-05-Card.pdf

For an university-level explanation of the model human processor, see http://jkwp.itsligo.ie/wp-content/uploads/hci/Human Abilities Cognition.ppt

Answer 2

I think that it largely depends on the type of buttons the printer is operated with. One way or another, it is better for each of the buttons, or group of similar buttons, to have as different a response (look, feel and sound) as possible.

On a printer, for example, it is advisable for the "print" button to have a distinct colour, shape, size and ideally also sound and texture from the other buttons. Arrows could then have their own set of characteristics and so on for each group of buttons.

That way one can operate the device purely by touch, as well as for example to describe someone easily what they should do ("Press the big green circle to print.") The sound of the button then reassures us that we did the right action.

That sound may come from the physical button itself, but if all buttons are identical, or product a similar enough sound (e.g. like buttons on the computer keyboard), an additional sound generated by the device may be a good idea, both to reassure yourself every time you use the button (I heard a beep = the text will be printed.) and to reassure those you explain it to ("You'll hear a beep if you did everything all right.")

Answer 3

It sounds like you're talking about "soft keys".

Feedback is necessary so the user will know they successfully manipulated a control. The main areas we deal with a sight, sound, and touch. In an ideal case your controls will give feedback in all three dimensions.

Pros for control activation sound:

• More feedback = greater re-assurance to user
• Critical feedback for blind user
• As @Erion said, may indicate higher level of machine response
• Different sounds to show scale of response (e.g. 1st, 2nd, 3rd press of fast-forward button on TiVo go up in pitch as it increases speed)
• as above but different sound for related buttons (e.g. touch tone)

Cons for control activation sound:

• Potential for annoyance to non-users (others in the room) = allow to mute, reduce volume
• Potential for annoyance to user (e.g. too loud, jarring, inappropriate)
• May not be heard (20% of americans have hearing loss)
• Interferes with an audio function (e.g. button on an MP3 player)