The HD bandwagon

September 23, 2009 at 5:59 pm 2 comments

I was on a panel at the  HDcomms summit in NYC last week. I was supposed to be talking about innovation, but since the High Definition voice has been around for 20 years I felt a bit of a fraud!

HD comms is about raising the quality of the sound we hear when we use a phone. The limits of the current phone system were set a hundred years ago and have never really changed since.

HD more than doubles the frequency range carried by the system. To experience this, try this demo I put together (press the SD/HD button to hear the difference).

Jeff Pulver –  a VoIP pioneer – started us off, basically saying that HD is going to make several people rich, just like VoIP did.

Looking around the audience and talking with them in the breaks, it is clear that many of them did well out of VoIP and are keen for a second shot.

I noticed a very pragmatic attitude, everyone needed to know immediately what the business case was – no Web 2.0 “lets build it, scoop up a million users, then worry about income” here, these guys are not interested in free or freemium !

It was generally agreed that HD voice will be adopted on a 5 year timescale (but faster if the FCC weighs in).

France Telecom are taking a lead by rolling out HD voice on their ‘triple play’ offering in France and on cell phones in Moldova. Unfortunately they use incompatible systems. It  looks they will have 2 independent HD voice systems in France next year using the AMR-WB codec on mobile phones and G.722 on HD VoIP. Calls between the two will be “not the full HD experience” as the France Telecom man delicately put it!

The HD voice experience has a significant wow factor but it’s a one-to-one wow. A single HD TV can convince a whole crowd at a time, whereas each of them would have to make an individual phonecall, ideally to a friend, to be convinced by HD voice.

HD voice will spread best virally. We are hoping that the gadget can help with that viral usage, allowing people to try HD before they buy.

Astonishingly, Verizon is already fully HD enabled – for their own corporate communications – but not to the outside world. This typifies the ‘island’ problem, where there are many large networks of HD that don’t interoperate – rather like the early days of SMS.

There was some discussion about how such ‘islands’ can be brought together.This discussion seemed to come down to 3 things

  1. Recognition that both ends are HD capable (peering and registries) Xconnect and Voxbone offered solutions here.
  2. Selecting a common HD codec – either by specifying a common default codec ( but which one ?) or by transcoding (audiocodes are building boxes for this) or having endpoints dynamically load codecs as needed – which everyone felt was impossible – except us since we do exactly that in our web-based HD audio device.
  3. Incentivizing the islands to join up – currently the PSTN (the default) provides a financial incentive for companies to peer. No solutions were offered to this.

At this point in the discussion I asked why we couldn’t use the PSTN to bridge these islands. This was foolish, it wasn’t what the VoIP crowd wanted to hear. I was told it was impossible because the PSTN was incapable of carrying the wideband data.

I’ve done a little digging since – and it looks to me as if Q 931 (which is the protocol underlying all of Europe’s ISDN infrastructure) is capable of carrying g722. Indeed Asterisk’s libpri has been able to since Sept 2007 . This means big players with large ISDN investments may choose to continue to use circuit switch channels to bridge their wideband islands.

The other subject that was ‘verboten’ but everyone talked about was codecs. There are 3 main contenders:

  1. G.722 (and polycom’s smarter derivitives) used in desktop phones.
  2. AMR-WB (and close relations) used in 3g cellphones
  3. Silk (and/or whatever comes out of the IETF’s efforts) used in Skype

My view is that it isn’t a coincidence that all the handset makers want to use G.722 now it is out of patent life time. This illustrates the stultifying effect of patents on an eco-system. Fortunately the players seem to be learning from this and trying to open up the market.

  • Voiceage, who handle AMR-WB, are investigating new models for licensing it ‘more flexibly’.
  • Skype have offered Silk as an IETF standard as have some other contenders. (Silk takes the view that the network is a variable quantity and encourages the endpoints to adjust the codec parameters dynamically. In effect getting the best out of what is available.)
  • G.722 is already free and the source code is simple to implement on just about any hardware. It requires the same bandwidth as the existing G.711 standard making it simpler to ‘swap in’ at least in theory.

My main contribution to the panel debate was to point out that whilst the majority of delegates had traveled without a HD capable phone, almost all of them had brought a laptop with a built in microphone. Laptops make a perfectly acceptable HD device, given the right software and cheap headphones.

I’ll be watching this space, Doug Mahoney’s HDConnectNow site will be my first port of call.


Entry filed under: VoIP. Tags: , , , , .

The. best. job. ever. – build Skype 2.0 in a year. My Astricon Googlewave ibook skype demo

2 Comments Add your own

  • 1. Michael Graves  |  September 23, 2009 at 7:39 pm

    Your short list of leading codecs is right on the money. You miss on little factor. The 3G mobile space is divided into two camps: GSM and CDMA related. The 3GPP group behind GSM has selected AMR_WB, but Qualcomm and their group have selected EVRC-WB.

    This one factor, a split in the mobile space, dictates that there will be transcoding. The may never be just one codec of choice.

    • 2. babyis60  |  September 24, 2009 at 8:22 am

      Yes, you are right about Qualcomm and EVRC-WB. There was some discussion about this and I got the impression that EVRC-WB and AMR-WB aren’t actually that different, but I may have misunderstood what was said.

      I think it is inevitable that there will be multiple codecs. I think there is a nice market for transcoders in the short term, but loadable codecs in the endpoints is probably where we will end up.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Trackback this post  |  Subscribe to the comments via RSS Feed

Follow me on Twitter

%d bloggers like this: