Blog Post

High Definition to Crash the Voice Party

The iPhone and the Apple (s aapl) App Store may have set a new standard for design and availability of mobile apps, but the iPhone operates within the same decades-old voice quality constraints as other handsets. Indeed, the rapid pace of handset innovation does not change the fact that AT&T (s t), BT (s bt), Telefonica (s tef) et al cannot improve voice quality. But a new generation of VoIP devices supporting G.722 and other wideband codecs threatens to disrupt a telcom industry built on mediocre voice quality.

On one hand, the analog infrastructure between the end user and the telco central office limits landline voice quality; on the other hand, the network equipment cell phone companies deploy sacrifices voice quality in order to make more efficient use of spectrum. While overcoming these limitations requires a complete overhaul of the telephone network, Internet-connected devices, represent the only practical means to build a High Definition Voice Network (HDVN). This is especially true when it comes to landlines; a G.722 wideband codec doubles landline frequency response and pushes fidelity toward an FM broadcast signal. (See a demo created by Siemens Gigaset.)

As with any technology rollout, there are identifiable obstacles ahead. Improving voice quality requires compatible devices at both ends of the connection, for example, and the specific communication contexts or content that require voice quality improvements remain uncertain. Notably, an emergent HDVN will not immediately enjoy the reliability and ubiquity of the traditional telephone network. And since it will require reliable broadband access, the early focus will be on fixed locations.

A similar list of issues did not, however, prevent the introduction FM, color TV, CDs, DVDs or HDTV. And people will embrace the HDVN as they experience the benefits of improved voice quality in contexts that involve emotional content. In particular, the HDVN ecosystem will operate more like the Internet and web rather than the vertical integration model of traditional telecom.

For a glimpse into the benefits of an HDVN, look no further than Skype (s ebay). If an HDVN can replicate Skype’s ease of provisioning and use while leveraging open rather than proprietary protocols, the telephone business will change forever. But the decision to compromise voice quality via a revenue model based on connecting calls to the telephone network also makes Skype a cautionary tale. An HDVN needs to survive as a platform for selling software and devices, as in the case of the web. Telephone companies will look for ways to undermine an HDVN as they’ll see it as a threat to their connectivity gatekeeper status. The success or failure of these efforts will determine whether voice quality remains unchanged for another 50 years.

Daniel Berninger
is the CEO of Free World Dialup.

20 Responses to “High Definition to Crash the Voice Party”

  1. I ran across an interesting article titled “Talk Is Cheap” reviewing top VoIP programs at This is the article link here to read a description and see videos about top webchat programs.

  2. Some of the comments about analog being good enough and people don’t want any better reminds me of another debate that I think has been pretty much settled. And that question was whether consumers will spend $1.50 for freshly brewed Ethiopian coffee when they can get the neighborhood mud at the diner. People did notice the difference and an enormous market was born. As far as analog audio being good enough technically, there’s important research that shows that certain aspects of the human voice signal, called formants, are blocked by the PSTN’s 3.4 Khz bandwidth. The HD experience is simply better, and here’s hoping that carriers and service providers will begin to support it

  3. Wideband audio is a solution to a problem most people don’t know they have. Point-to-point mobile-to-mobile phone calls work most of time, even though they sound like crap. I tell people: I talk to my brother over those connections all the time. The audio isn’t 100% intelligible, but we’ve been finishing each other’s sentences for almost 50 years so it doesn’t matter. We’d all like it to be better, and the technology community really should be ashamed of what they’re delivering here. But it’s functional and people are (obviously) willing to pay for it.

    Conferencing (my business) is a different animal. You’re often talking with people you haven’t met. You’re trying to recognize voices, understand foreign accents, and decipher unfamiliar speech. There’s background noise and double-talk. Now layer on narrowband sampling and encoding, which strips away half of the audio information, and these calls can be disasters.

    Ever wonder why you’re exhausted after a 45 minute conference call? It’s because a sizable fraction of your brainpower was devoted to “recovering” the missing information. Spend a couple hours on an HD conference call and you’ll feel a lot better.

    Someday, hopefully we’ll all enjoy wideband audio as part of the telecom routine — even on your mobile phone. Technology, as noted, is not the problem (there’s already a wideband codec defined for GSM, AMR-WB / G.722.2). But it’s going to take a lot of time and work to make the subconscious benefits more obvious to the general population, and even to the “specialists” in the field.

    A friend of mine suggested a while back that the FCC should mandate wideband audio for all telephony by some date certain. Maybe we should have made that part of the stimulus package!

  4. If the only issues were technical then the US would have broadly adopted ISDN as a superior solution, including end-to-end digital voice with wideband potential.

    Incumbents will generally wring every last dollar out of the status quo, resisting any investment regardless of the potential. A such, I think that it will take a major change, revolution, not evolution, possibly by a new player in the game, to drive widespread migration off the PSTN.

  5. For those with perfect hearing, “good enough” quality may be just fine. For those of us with substandard hearing, any improvement will be welcomed.

    That said, I do not think that incumbent telcos will try to ignore this technology. Part of the reason that voice call quality has stayed stagnant the last 20 years is the increase bandwith on the line being dedicated to other uses. Currently I get traditional PSTN voice, 3 Mbps internet, and HDTV over a single copper pair (not a standard install, but all that was available in my particular case). Additionally, I could add a dedicated VOIP connection to the pair if desired.

    The incumbents have the ability to offer HDVN if they desire, and I am sure that they soon will offer it, although the price may be higher than some are willing to pay for many years still. It may just get implemented without upgrading the PSTN network. And it will likely not be exclusive in my lifetime (hopefully another 50 years). :)


  6. The voice-quality-is-good-enough crowd miss two things. They seem to forget conveying voice represents the core value proposition of a telephone call. It is unthinkable that the core value proposition of a technology based service should remain static for 50 years. Embracing good enough in the case of voice quality makes no more sense than embracing good enough in PC processing power. The existing applications may work adequately with the status quo. The point of continuous improvement is expanding the range of addressable applications. The fact voice quality does not change owes to the structural obstacles to improvements inherent with the telephone network. It is not because no one can imagine benefits coming from improved voice quality.

  7. I frequently observe people asking a caller to dial their landline to resume the conversation over higher quality. You’re right that there is a point where it’s “good enough” that it won’t matter, but I’d hypothesis that wireless voice is largely below this threshold.

    That said, I think some sort of high-quality voice solution will ultimately be delivered by incumbents, not a newbie startup. No carrier will permit the double-whammy of greater spectrum usage AND over-the-top voice cannibalization, and will likely pursue their own, possibly premium priced offering once the demand is validated.

  8. Through the transition to cellular we’ve been trained to accept substandard call quality. As someone who uses G.722 literally every day (there’s a Polycom IP650 on my desk) I speak from experience when I say that quality makes a difference.

    Since I first evaluated the Polycom HDVoice products about a year ago many people around me have started adopting these devices. Starting a month ago the week VoIP Users Conference call ( has been hosted on both wideband ( and traditional ( conference bridges. Thus our podcasts serve as examples of the difference between old skool PSTN and better.

    Happilt, ounterpath is about to release a new retail version of their Eyebeam soft phone for Windows that will also be G.722 capable. I’ve been using a pre-release vresion for the past week and it’s excellent.

    We need to not only demand better, but also do something about it. Use wideband IP telephony where possible, and push for better where it’s not offered.

  9. Jesse & Derek:

    There exists no end a debate about which factors have the greatest impact on voice quality or the voice quality required for a given context. The column makes a very simple point…we need to do better! A world where telephone network voice quality remains static for 50 years represents a crazy anomaly in our supposed technologically advanced society. A tremendous amount of information available in face to face conversations gets dropped above and below the 3.4khz pass band of “toll quality voice”. This has really productivity consequences for society. Yes, deal with jitter. Yes, deal with noise and stop making excuses.

  10. Derek Fields

    I would raise Jesse Kopelman’s comment up a level of abstraction. Most people don’t really care about voice quality as long as it is clear enough so that they can communicate effectively and efficiently. Audio quality is valuable for applications like music streaming, but that is an entirely different application of bandwidth.

  11. Jesse Kopelman

    But it’s already well established that the real driver in perceived call quality is consistency, not full-range audio. It is also well established that full-range does not equally good sounding for most people — case in point: the success of Bose. Full-range codecs are just a gimmick. The most important thing for VoIP is to control jitter and that is something completely in the domain of the network operator.