A question we get asked a lot is regarding the sizing of voice mail messages in Exchange 2007. If you’re not familiar with the built-in voicemail capabilities, Exchange can function as a voice mail system (or Unified Messaging system, really – it’s a way of unifying voice and inbound fax messages with email).
What’s particularly nice about this is that as far as Exchange is concerned, email and voicemails are just messages. I can respond to a voicemail (such as the one pictured here) by hitting reply, and Outlook (or OWA, or Windows Mobile etc) will create a email response to the “sender” of the voice message, assuming it can work out who they are based on the caller ID that was identified when the message was left.
Lots of people get nervous when thinking about holding voice mail in Exchange, worrying that the message sizes will burden their already overloaded mailboxes. In reality, the size is rarely a big deal – we tend not to get too many voicemails (I probably get less than 10 a week), at least in comparison to the volume of emails received. Add to this the fact that most voicemails are relatively short (and you set a limit on how long the system will let a caller ramble before cutting them off anyway: generally if it’s more than 2 minutes long, then it’s more of a soliloquy).
There are a few ways of encoding the voice content that Exchange will record as voice mails, and which option you choose might depend on how the users are going to be collecting the voice mails. Outlook, OWA and Windows Mobile can all play Windows Media (WMA) format files, so that’s the default – and offers the highest quality for minimum size of message – typically a couple of Kb per second or so (a combination of some overhead for the message, and then the encoding rate of the sample).
The options are to stick with WMA, or if you’re looking to interoperate playback of voice content with other telecoms equipment, you may want to encode using GSM 06.10 (an 8-bit compressed format derived from the GSM mobile specifications), or G.711 (a 16-bit PCM non-compressed format, defined as an ITU standard). Both GSM 06.10 and G.711 use the WAV format for representing the sound, and will deliver larger sound files than WMA.
There’s a nice explanation of the options over on http://technet.microsoft.com/en-us/library/aa998670.aspx, including this comparative graph of the file sizes:
Basically, don’t use G.711 unless you want *really* big voicemails…
Finally, SP1 will add the option of using Microsoft’s RTAudio codec for playback to Office Communicator Phone Edition devices – part of the integration between OCS and Exchange 2007.