There is a sound specification.
It is here: http://0pointer.de/public/sound-theme-spec.html
There is also a sound naming specification.
It is here: http://0pointer.de/public/sound-naming-spec.html
To be honest, I don’t have any problems with the former document and really couldn’t care less about how sounds are stored, retrieved, played, or compressed (*). The latter however is complete and utter unmitigated bullshit.
I shall explain why.
The problem with the naming spec is two fold:
a) They define too many arbitrary sounds.
b) They define so many sounds that it is impossible for a sound designer to make sounds that differentiate between the actions
99% of computers have had the sounds turned off. Why is that? It is because users do not want a computer telling them every single time they do something. This is similar to the butler principle, but I call it the child principle. A computer should be seen, and not heard, unless absolutely necessary.
There are 125 sounds defined by the sound naming spec. One Hundred and Twenty Five different sounds for the user to learn and recognise. Have you ever played BopIt! where you try to react to 4 different sounds? Imagine it with 125 sounds. Wait, was that swishy sound New Email or was it Network lost, or was it computer made a move?
Here is the solution.
Firstly sounds need to be subtle.Think of it in terms of the butler principle. When a butler wants the master’s attention he doesn’t scream “OI LOOK AT ME! OVER HERE!!!!!”. He politely and subtley gives an *ahem* and then waits for the master to finish what he is doing before giving the information.
If a sound has to get in your face and wrench your attention away from what you are doing, it has failed. Think the Windows’ error dialog sound. DUN! One of the most annoying sounds ever. A subtle sound will happen but it will just register in the back of your mind. You cannot make 125 subtle yet different sounds that the user will be able to associate with specific things.
The Sound Spec (the good, mostly sane one) defines 5 categories, Alerts, Notifications, Actions, Input Feedback and Games.
We replace it with 3 sounds: Alert, Notification and Input Feedback. Yes, 3 sounds. One sound for all alerts, one sound for all notifications and one sound for all input feedback. Radical, I know.
“But what about Actions and Games?” I hear you cry. Well, in the sound naming spec both Actions and Games have a very arbitrary range of available sounds. Games is really only useful for card games, and actions is so arbitrary that it only covers 4 possible applications. Applications which need these sounds can provide their own, just like was the case when the icon naming specification was trimmed of all the application specific crap. (Ironically, one of the main proponents of the icon naming spec being trimmed is also an author of the sound naming spec)
You will note above that I said 99% of computers have sound effects turned off. What is that 1% that doesn’t? I think its the ipod. How many sounds does it have? 1
A soft subtle click. You pushed a button, click. You unlocked something, click. You scrolled the popup menu, click, click, click. It doesn’t try to convey information in the click, it is just a response to the action.
Sounds need to be < .5s long, otherwise they are too long for being repetitive (think the windows’ sample again being repeated DUN! DUN! DUN! DUN! How many times can that be repeated before you want to smash something?) or subtle. The sounds I create for sound effects are usually < .2s long. It is impossible to convey specific information in that length of time.
There we go. The “sound problem” solved in one easy step. I bet if we only had 3 sounds people would stop turning off sound effects within minutes of installing, and sound designers would be able to stop tearing their hair out trying to make sounds that differentiate between menu-popup, menu-replace and tooltip-popup. Maybe if some sound designers were involved in the making of the sound-naming-spec this wouldn’t have happened?
Of course, the beauty of this blog is, no-one who cares reads it 🙂
(*) Beyond my slowly dawning realisation that the concept of themes is really just a band aid over the fact that we can’t decide on one good style, and I think themes will hurt more in the long run. Ohhh controversial.
Agreed. I want one gentle ping which is simple and *well used*. None of this avalanche of effects when you toggle a check box as GTK+ used to do. I turned sounds off years ago and have yet to turn them back on, so the recent work may have fixed that.
Totally agreed – Interestingly, Vista seems to have taken this approach too…