18 Comments

Summary:

In a few hours, the annual Spring VON is going to kick off in San Jose, California. Hundreds of companies, big and small will hawk their wares, pundits will pontificate and a lot of people will talk about convergence. But no one will bring up the […]

In a few hours, the annual Spring VON is going to kick off in San Jose, California. Hundreds of companies, big and small will hawk their wares, pundits will pontificate and a lot of people will talk about convergence. But no one will bring up the dreaded question: Is VoIP an excuse for bad voice quality?

Earlier this morning, with my broadband on the blink, instead of iChatting (free) with my parents, I called them from my old fashioned telephone line – you know the kind the incumbents have been selling for over 100 years.

The conversation involved a lot of yelling into the phone – like we used to back in the day when the Internet wasn’t around, and long distance phone calls cost $3 a minute. While the price of the calls has declined to a few pennies, so has the quality of voice.

One can clearly recall a time when Sprint made a big deal about its voice quality, touting it in a “hear the pin drop” advertisement. AT&T spent hundreds of millions in coming up with a better voice experience, a business that didn’t clearly help save the company, though it made Joe Nacchio (oh yeah, the very same one) quite famous. All that is part of the history phone companies seem to want to bury.

The long distance call between San Francisco and New Delhi might as well have been a call between the International Space Station and my landline – choppy, static filled and barely audible. It is a pattern you observe time and again, because more and more incumbents are using VoIP technologies to carry their international traffic, trying to squeeze whatever little profits there are from the ever-declining business.

Lowering their operational costs is an understandable business move, but for companies whose primary reason for existence has been voice, it is just not cricket. On a testier day, WTF would have been my choice of words, but today, the slight nip in the air, bright sunshine and backache in remission, I am in a more generous mood.

VoIP has been a protocol of choice for a while now, and that is why it is hard to understand the quality problems. While consumer facing services such as SunRocket and Vonage, the shoddy voice quality can be blamed on the broadband bandwidth constraints, the long distance carriers (owned by incumbents now) should not have these problems, given that most of them own their backbones, and the gear seems to have matured enough to provide better voice quality. (Read: PSTN vs. VoIP)

Call me old fashioned, shouldn’t incumbents and the upstarts make Voice their core competency, a deluxe experience (like BT), before offering television and high-speed connections or some dumb Wi-Fi phones? Or is it just that we as consumers have been desensitized, our expectations lowered by the poor quality of mobile phone connections that we will put up with anything as long as it is cheap? I don’t think that is the case – and I pray to god, I am not in minority. If it is poor quality one needs to put up with, then free iChat makes more sense to me. Even Skype – which does a relatively good job for a free service!

  1. Hi Om,

    Small mistake in the first paragraph, instead of “But mo one will bring up the dreaded”

    should be “no one”

    Should not be a comment, you can remove it from the comment list.

    Keep up good job,
    Moti.

    Share
  2. I have Vonage at home here in New York and I can report that the it is nearly impossible to call Europe, especially towards mobile phone. The quality is so horrible (most of the time I can’t hear what is being said), I have given hope on them and will soon switch to ATT.

    Vonage should not be allowed to sell overseas call, this is simply unusable and a total rip off.

    Om I understand your frustration !

    Share
  3. Unless consumers are given a standardised grading of voice communication quality, i.e. Kbps, compression level etc – but wrapped up into an easy to understand rating system (out of five stars maybe?), then they won’t care about the quality when signing up for a service – it’ll be the price that counts.

    And until the average consumers start to care about the quality of VoIP or other voice comm protocols, and start turning sound quality into a deciding factor when choosing one service over another, we won’t see the knock on affects of the providers embracing hi-fidelity services in order to obtain and keep customers.

    Share
  4. I have a varied and disjointed comments with at least might be considered a rib, but all in good nature.

    1. Joe Nachio’s response to “hear the pin drop” was a failure. Very few could tell the “true” voice. The only thing it produce was David Isenberg’s now famous paper.
    2. You failed to point out that AT&T had used “packet” technology to carry multiple calls over the same number of trunks. This increased the call completion rate and brought down the rate considerably. The so called “wide band packet technology” developed by Bell Labs was later usurped by Stratacom (?, memory eludes me) that Cisco acquired later on.
    3. Here is the ribbing comment: When the incumbents place interconnection requirements demanding QoS, many (guilty parties can raise their own hands) claim foul and say that this anti-competitive. When the prodigal son returns and complains it is not clear what should be the reaction.
    4. I use Reliance for every call to India for the past few years. Never had any problem. The rates are cheaper than Skype. Skype and iChat are not available choice in my case anyway.
    Share
  5. As a telecom guy I can tell you that the mobile experience from the early 90s dramatically lowered the expectations for pstn quality – and opened the door for the every voip company today. The first one that I remember was Selsius later bought by Cisco.

    There was a time (in the US anyway) that every call was expected to work flawlessly and with zero quality problems. Now, most folks just hang up and try again. The expectation for cheaper and faster trumps the expectation for better – ala the Walmart effect.

    –Jon

    Share
  6. This may be an unimportant distintion, but VoIP isn’t a protocol, but there are a good number of VoIP protocols.

    And Cauldwell, there is a standardized and simple grading of voice communication quality called a Mean Opinion Score (MOS). The regular telephone is supposed to get a MOS of 3, and VoIP can get rather higher depending on factors you mentioned such as bandwidth, codec, and others.

    Share
  7. Thanks William – obviously I was completely unaware of MOS – but once again this is just another acronym that will probably inhibit a widespread understanding of a simple concept, just as the RSS acronym has overcomplicated syndication.

    Share
  8. You’re assuming the problem is on the U.S. “big telco” side. Not necessarily the case. I don’t often see quality problems using POTS for inter-country U.S. calls. Perhaps the problem is with New-Dehli-Bell. Nonetheless, call quality will go down with cost-cutting, and cost-cutting is JOB ONE, as the old monopoly reconsilidates and tries desperately to maintain their business model long enough to transition to a shiny new one. Too bad video isn’t going to be the saviour for them.

    Share
  9. “routerguy” has put on the table the real problem. As the Chief Flattening Officer at the Flat Planet Phone Co. I am constantly asked how can a business depend a VoIP service?? The truth is that all the carriers from at&t down use VoIP for terminating their international calls. The typical carrier works with a few terminating partners in each country and sends traffic to the cheapest bidder. Quality is sacrificed to save a tenth of a cent. In fact in many cases our customers of the little VoIP company get better quality than the customers of the BIG Carrier…

    Share
  10. SIP + speex Wideband = Better sound than a phone. Problem : if you call on PSTN, most of the proxues does not support anything else than PCMU …

    Share
  11. We consumers have clearly demonstrated how little value we place on QoS, at least as it relates to voice communication. As more and more people flocked to ‘free’ internet services like Skype, and began thinking of the PSTN as either irrelevant, or a premium service (as you obviously did), the price of that ‘premium’ service would have had to go up to compensate for reduced usage and revenues and reflect that value premium, or the Service Providers could race for the bottom, and try and compete on cost. In their judgement, the latter must have appeared more viable.

    Share
  12. i agree with everyone that voice communication companies should focus on providing good voice quality before jumping on the next hot thing. But then companies fear competitive disadvantages. So the fact of life is, with the communication worlds converging, VoIP (or some other form of Internet based voice communication) is here to stay.

    Let me provide you with the other side of the story. In the last two years (working with Keynote Systems, Inc.), I have been involved with the customer experience measurment of the voice service quality in the USA. Our overall observation is that the VoIP quality is gradually improving and in some cases it performs better than the PSTN.

    Take the case of cable based VoIP providers. They have the bandwidth advantage as well as they are providing a new lastmile connection. Also as compared to the analog communication, the digital communication is more immune to the external imparement factors. Therefore, because of the last mile factors we have frequently observed that the Voice Over Internet services perform better than the PSTN especially in the old houses or in the houses away from the centeral office.

    As far as the international communication is concerned, I have not analyzed the data, but my reliance phone card experience has been extreamly pleasant (may be they are not using public Internet at all). I am sure once Vonage and likes can figure out how to reduce their extreamly high customer acquisition cost, they will start focussing on improving their call quality.

    Meanwhile, some of the VoIP service providers are becoming confident of their voice service and publishing their voice service quality numbers (monitored by an independent third party) on their website (example http://www.vonics.com). This trend suggests that just low cost is not going to be a differentiator for the VoIP companies, they will have to start cometpting on the overall service quality.

    • Arun
    Share
  13. So what is the best one now? Skype? My main interest is that it be free (of course), and at least intelligible voice quality. I’m not going to fault them too much now as long they intend to improve it. And when is Cisco going to silence all these wannabes? Let’s see free worldwide video conferencing!

    Share
  14. You are so right! Why does voice sound so bad on the net? We have great codecs, great technology, and still, regular calls and VOIP sounds terrible.

    The tech is there. Someone should make it work!

    Share
  15. VoIP Quality or Lack of it…

    Om asks the question, Is VoIP an Excuse for Bad Voice Quality?, the answer is no, it is not.The truth is that VoIP is capable of delivering better quality than most are used to but the technology is not given……

    Share
  16. Jesse Kopelman Monday, March 19, 2007

    william, a ‘regular’ telephone should actually get a MOS score of at least 4 (that is toll quality). An MOS of 3 is barely acceptable. Most humans have trouble differentiating between MOS=4 and MOS>4, but the differences below 4 are quiet dramatic. Also, when it comes to MOS your individual mileage will vary as the actual scores are based on a table of averaged opinions done way back when. Also, MOS is a measure of how good the reproduction is not how nice it sounds. Just like you may not care for a wine that got a 90+ score from 3 different reviewers, a call that scores a MOS=4 may still not sound that great to you. If you are very used to listening to CDs and especially MP3s and using digital phones, you may actually prefer a digital conversation with a lower MOS to an analog conversation with a higher MOS. Few digital phones break the MOS=4 barrier, due to the nature of vocoders used. A decent mobile phone call is usually around MOS=3.5. A final thing to consider is that a call may be of excellent quality over the actual communications network but still sound bad to you due to a poor quality components in your handset. A lot of cordless landline phones just sound terrible.

    Share
  17. Todd Holsombeck Tuesday, March 20, 2007

    Om…this deserves a bit more looking into. MOS scores can be rigged. LAN’s are often misconfigured. Poor quality endpoints effect call quality. There is also beginning to be a debate about the fact that consolidation plays a role here. All the dark fiber laid in the late 90’s has not been lit. As we move toward less and less CLEC’s how many routes are out there to move around. Combine this with “hot potato” routing schemes most larger backbones employ, and you may be beginning to see congestion. Not the kind of congestion to disrupt email or web surfing, but the kind that makes your real time call to India bad. Put in the Ted Steven’s vernacular, people are treating the Internet like a dump truck…not a series of tubes!

    Share
  18. I actually do care about call quality, since I call people in Asia a great deal. At the start of 2006, I was juggling Vonage and Skype. At the end, I threw in Yahoo! Voice in the mix. I dropped Vonage because the price and call quality were poor. I try to stick to Yahoo! since the rates are great, but the quality is a crapshoot. Skype is my old reliable, but unfortunately its rates are double that of Yahoo!’s (for the areas I call).

    Share

Comments have been disabled for this post