VoIP players are setting new standards in voice quality with HD codecs. Cellcos have had their own HD voice standards for years, but challenges abound, and the simpler solution could be to embrace VoIP for themselves
Television channels aren't the only things going "high-definition" (HD) these days. Even voice calls - the traditional cash cow of the telecom sector - are moving into HD territory, particularly in the mobile sector. HD voice promises to make callers sound like they're practically in the same room. And it's been generating considerable buzz in the past six months.
Several HD voice solutions were on show at this year's Mobile World Congress in Barcelona. And cellcos in the UK have been demonstrating HD voice, including 3 UK and Orange, the latter of which launched trials of HD voice in several cities with plans for a nationwide rollout. Orange also has HD voice trials ongoing in Eastern and Western Europe.
However, for the most part it's not the traditional cellcos that are turning users onto HD voice, but upstart VoIP players - which is ironic, not least because VoIP call quality is typically portrayed as a matter of low latency. Anyone who has used Skype or similar VoIP apps knows when the call is having latency issues. However, they also know that in terms of actual voice fidelity, VoIP codecs are way ahead of the game.
Skype executives know it too and have been going out of their way to tout the value of HD voice. At the Emerging Communication Conference America earlier this year, Skype chief technology strategist Jonathan Rosenberg included a slide showing that as his company's own SILK codec has improved in audio quality, the frequency and duration of voice calls go up.
Rosenberg said that HD voice quality can increase the length of a voice call 45% (see chart, page 20). Little wonder that Skype's latest iPhone client, which allows Skype calls over 3G networks rather than just Wi-Fi for the first time, boasts "CD quality" voice capabilities.
Meanwhile, Google clearly sees value in HD voice. In May this year, it paid $68.2 million in cash for Norwegian company Global IP Solutions (GIPS), which specializes in VoIP and video processing platforms. Industry observers have speculated that Google could put GIPS' HD voice engines to work to enable an HD version of Google Voice for its Android OS.
What's even more ironic about VoIP's lead in HD voice is that circuit-switched HD codec standards have been around for years. The mobile version, Adaptive Multi Rate Wideband (AMR-WB), has been an ITU standard (G.722.2) for close to a decade, as has the fixed-line version (G.722), and solutions have been commercially available since at least 2006. But deployment has been held up by a number of factors, from cost of deployment to sheer lack of incentive. After all, customers have been using standard voice codecs for ages - they're used to it, and as long as they can understand the person on the other end, why fix what ain't broke?
But that's becoming less and less the case as VoIP - which claims over 900 million users, according to ABI Research - increasingly sets expectations on what voice calls ought to sound like. The trick is that the challenges that have delayed HD voice for so long for operators remain in place - so much so that (and this may be the biggest irony of all) the answer may be to embrace VoIP for themselves.
Handsets, expense, coverage
The chief factors holding up HD voice for operators boils down to a few key themes: handsets, expense and coverage.
The problem with handsets, says Ericsson CTO Michael Lee, is simple: mobile phones have to support AMR-WB for HD voice to work, and at the moment, few do.
"We haven't seen a lot of support from the handset market to accommodate AMR-WB," Lee says. "But to be fair, it's also a chicken-and-egg problem. If there are few operators interested in AMR-WB, why should handset suppliers support it?"
Another issue is network investment, says Lee. "Today operators deploy transcoders in every switching site of the network, so that the 16-kbps codec I use between the mobile and the base station can convert into a 64-kbps connection to adjust for the traditional circuit-switched PCM-based voice codec," he explains. "If you want to do AMR-WB, you have to take those transcoders away and introduce new functionality at the gateway of the architecture that handles transcoder-free operation. All this means you have to make some investment in the network."
This factor has given VoIP players the edge over traditional operators in HD voice, says Alexander Kravchenko, marketing director for voice/video engine specialists Spirit DSP.
"Using HD voice with something like Skype doesn't bring any additional cost to the service provider because the software is free," he says. "For carriers, the situation is quite different because you have to install a lot of equipment."
This isn't necessarily the case for everyone, says ABI Research principal analyst Fritz Jordan. "Newer 3G networks - those deployed since about 2005 and 2006 - can already use the new format and require only a software update and a changeover to HD handsets," Jordan said in a research note. "That's why HD voice, unlike most technologies, will first find traction in developing markets," while markets with older 3G networks will have to upgrade their networks.
However, that raises another key problem - HD voice has to be supported on both ends of the call, otherwise it drops to the default narrowband codec. And islands of HD voice support inevitably mean inconsistent service, says Lee.
"What that means for the end-user's point of view is that sometimes you get high-quality voice and sometimes you get standard-quality, depending on whether you call someone on another network that doesn't support HD voice, including a fixed-line phone, or someone whose handset isn't HD-enabled," Lee says. "It can even be an issue in cases where the operator has deployed HD voice for its 3G network but not its 2G network to cut costs."
That's tricky for operators, he adds, because once users try HD voice, the lower quality of narrowband becomes much more noticeable. "Once they try the high-quality, they notice when it's lower quality and they think it sounds bad, even though it was what they were used to in the past."
A related issue with connecting HD voice islands is interoperability between the various codecs in play, from G.722 and AMR-WB to VoIP codecs like Skype's SILK, says Jim Machi, senior VP of worldwide marketing for Dialogic.
"Interoperability is a major issue, whether it's interoperability from HD voice-enabled networks to non-HD voice-enabled networks, or from one HD voice-enabled network to another, as well as potentially from one HD voice codec to another type," Machi says. "There are infrastructure elements called gateways that would need to be installed to enable this. Or more appropriately said, there are gateways being installed every day in networks around the world, so we would need HD voice enabled gateways."
The interoperability problem also includes related value-added services like voicemail, Machi adds. "If you are on an HD voice-enabled network and you want to record a voicemail, you would want to record it in HD voice format, right? So we need media servers to be HD voice enabled."
Perhaps the biggest obstacle for enabling HD voice is the business case for it. Put simply, there isn't one - at least not if operators see HD voice as a potential new source of premium revenue.
HD voice is a hard sell, partially because of the inability to guarantee HD connections regardless of call destination, but also because not everyone can tell the difference on first listen, says Kravchenko of Spirit DSP.
"If you play narrowband and wideband voice samples for the customer, not everyone can perceive a big difference between them right away," he says. "It takes time for them to use it and get used to it, and only when they go back to narrowband do they hear the difference. So it's hard to sell HD instantly. That also means that you can't really charge more for HD voice, even for enterprise users."
The real value for HD voice, says Ericsson's Lee, is in two main areas: longer talk time (as championed by Skype above) and customer retention. "If the question is whether that translates to more ARPU, we see some doubt from the operators on that, and we haven't yet seen a successful case where the operator can generate higher ARPU directly because of HD voice."
That's not to say there won't be opportunities to put HD voice to use in creative value-added service bundles to differentiate themselves from the competition, says Lee.
"Take couples, for example - if you and your lover both subscribe to the same mobile operator, you can enjoy very intimate HD conversations with your partner over the phone," he says. "For the corporate segment, you can offer HD voice to everyone in the company, which is useful for certain businesses that really require good quality voice, like stock traders, for example. You could also use it for DTMF speech recognition services."
If you can't beat 'em ...
Interestingly, despite all the talk about HD voice and VoIP, there's some disagreement over just how much pressure operators are under to take voice high-def.
Lee, for one, says cellcos feel more pressured by VoIP players on things like IDD tariffs rather than voice quality.
"They don't see it as anything urgent, because the value for mobile networks still comes from mobility, and as long as customers still enjoy that mobility, mobile operators still have an advantage," he says. "Also, they still face the same challenges now as they have in the past. So when the competition comes up from other mobile operators, they'll move more quickly to deploy AMR-WB."
Machi of Dialogic, however, credits increased usage of VoIP for resetting customer expectations for voice call quality.
"In the enterprise, using the latest equipment from Avaya or Microsoft or Cisco, you can get HD voice, and Skype also uses an HD voice codec," he points out. "So people have been exposed to HD voice and know how much better it sounds."
Either way, it's going to take time for HD voice to catch on in the cellco world - but when it does, it's going to ramp up fast, according to ABI Research. An April report says serious growth for mobile HD voice won't kick in until at least 2013, but usage of HD-enabled handsets will skyrocket to 487 million subscribers by 2015.
That growth could be even faster for cellcos that are prepared to embrace VoIP for themselves, says Kravchenko.
"In the mobile space, it's probably easier to move to HD voice by enabling VoIP," he says. "Instead of enabling wideband voice on the traditional circuit-switched network, you can use a VoIP software app on the data network and offer it that way. It's a way to get people to try HD voice without the expense of changing the terminals or the base stations."
Spirit's VoIP expert for Asia Slava Borilin adds that using VoIP as an HD enabler also makes the interoperability problem simpler to address.
"There are about five or six software codecs that you'll typically come across in VoIP, and it's not as difficult to get them to interoperate," he says.
The trick there, of course, is convincing cellcos to embrace VoIP in the first place. Most operators have resisted VoIP out fears of cannibalism of existing voice or data capacity issues, and many still block usage of VoIP over their 3G networks (although Wi-Fi usage is usually still allowed). But migration to all-IP LTE could mitigate the capacity concerns, says Kravchenko.
Meanwhile, Skype has been maintaining its charm offensive with cellcos, pitching a November 2009 case study by CCS Insight showing that 3 UK's partnership with Skype has not only lowered churn, but also boosted traditional voice and SMS usage rather than cannibalize it.
To date, only Verizon has taken Skype up on its offer to follow in 3 UK's foodsteps. But the idea does seem to be catching on elsewhere. In July, Korean operator SK Telecom announced that it would include mobile VoIP as part of a broader updated mobile broadband strategy that includes unlimited data plans and an accelerated LTE rollout timetable.
Granted, there are caveats in place - while mobile VoIP subscribers are free to use any VoIP client they wish on SK Telecom's data network, actual mobile VoIP usage will be metered and capped to avoid data congestion.
What's striking is SK Telecom's public admission that the operator's previous objections to VoIP - that it "might act as a disincentive to carriers to make investments and hinder industrial development," as president and CEO Jung Man-Won put it during the press conference - have been overridden by the realization that the pros of VoIP outweigh the cons.
"The introduction of m-VoIP is projected to have a negative effect on our revenues in the short term. However, we expect bigger positive effects in the mid- to long term," Jaeeun Namgung, manager of SK Telecom's pricing strategy team, told Telecom Asia. "It will deliver greater customer satisfaction and help us better retain our customers as we'll be providing them with a wider range of experience and options. Also, we expect growth in our ARPU and revenues as we attract new high paying customers and as our existing users switch to higher priced plans."
Interestingly, Namgung didn't specifically mention HD voice as a particular benefit of mobile VoIP, focusing instead on VoIP's inherent flexibility in terms of creating new voice apps and business models that would be too difficult to do with traditional voice.
"For instance, m-VoIP functions can be added to various applications including real-time remote lecture system, remote health-care service and real-time games," she said.
That said, however, the HD capabilities of voice certainly wouldn't hurt in any of those apps or services.
Either way, analysts are tipping mobile VoIP as a major growth opportunity. Frost & Sullivan says mobile VoIP will generate $29.57 billion by 2015, despite ongoing resistance from cellcos. And a May report from Ovum warned cellcos that those that choose to block or avoid VoIP do so at their own risk.
"Blocking VoIP is like trying to control the tides. Most mobile operators today have attempted different means of hindering the use of VoIP, or are cautiously monitoring usage," said Steven Hartley, principal analyst at Ovum and report co-author. "However, these approaches merely garner negative publicity from vocal early adopters demanding access."