Wideband Audio Anyone?

Glad I’m not the only curmudgeon who’s troubled by the state of 21st century audio! The lower quality of digitally compressed MP3 (compared to CD ) made the front page of yesterday’s New York Times. Economics and convenience are to blame for the lossy, lower-sampled recording formats that are used to cram more tunes into our portable devices.

So why are our cell phone conversations still stuck with a slice of audio spectrum that dates to the 1930s? I’ve written about a newer wide band codec (G.722 standard) that could deliver a far broader 7kHz of sound. Unfortunately, you’re more likely to experience that on station-to-station calls in a large corporate environment (courtesy of Avaya, Cisco, and other enterprise players).

Outside of the confines of an office park, we’re all struggling to make ourselves heard over a skinny, tinny sounding 3.3kHz swath. I found some of the answers as to why this is the case from a presentation given at this year’s eComm event.

To be fair, as consumers it is possible to experience 12kHz of sound with a Skype-to-Skype connection using their proprietary but royalty-free SILK codec. But for everyone else, the situation is less promising.

You would think since it built its network almost from scratch,  the cellular industry would at least improve on the legacy PSTN’s sound quality. While searching for videos from eComm 2010, I came across Dialogic’s Martyn Davies talking about HD voice and the TDM dependencies that are still with us.

At about the 5:20 mark, Davies has a slide showing why wideband is difficult to implement in a real cell network. The reason is that every part has to be upgraded—transmitting station, back haul, core network. The easiest part is upgrading the firmware in the cell phone itself!

One way out of this, at least in UMTS-type networks (too many acronyms, but UMTS is 3G for European and Japanese GSM systems), is something called TrFO (or transcoder free operation). The magic that TrFO performs is to negotiate a codec end-to-end between cell phones,  thereby avoiding triggering transcoders within the network.

Because UMTS is packet based, it’s possible for the two cell phone to signal each other in this fashion and communicate in wideband (forward to 8:13 in the video for the explanation).

What about getting wide-band audio to work between UMTS and everything else? The answer is a GSM peering standard called IPX.

Unfortunately, as Davies points, it’s not all sweetness and light here.

In short: we will eventually get wide band audio in the mobile world, but it will take some more hammering out of standards and agreements between carriers.