The Phone Network, an 80-year-old dog

Why has it been so hard to teach that old dog, the phone network, the new trick of sound quality that's as good as what we hear?

In 1930, the goal of the phone system was stated like this:  “In the Bell system the general objective which has been set up for the transmitted frequency range for new designs of telephone message circuits is a range having a width of 2,500 cycles, extending from about 250 cycles to about 2,750 cycles.”

The performance of the public telephone network has not gotten much better - even a little worse, in some ways, over the years.  As recently as 1984, the higher end was still about 2.7 kHz for long-distance calls, and 280 Hz at the lower end.   It’s hard to take something like the global phone network, one with fundamental goals set eighty years ago, and change its underlying fidelity in any meaningful way. 

This why telephones are moving to VoIP, Voice over Internet Protocol (or internet telephony), to make better fidelity possible for most people.  The internet carries data, and since any kind of signal can be converted into data, this means the internet can carry almost any kind of signal.  Did you get that?

1.  The Internet Carries Data

2.  Any Signal can be turned into Data

Therefore:  The Internet Carries Any Signal

Signals can be HD Voice, desktop video, Immersive Videoconferencing, and most other kinds of information you may want to share.  Because the Internet can be made arbitrarily versatile and fast, it can keep up with the needs of live human communication for a long time to come. 

Reference:  “Transmitted Frequency Range for Telephone Message Circuits,”  W.H.Martin, Bell System Technical Journal July 1930    Ref. JR1/11 

What Is "HD Voice?"

There’s a spectrum for sound as there is for light, and it spans the range from low, booming bass like kettledrums and distant thunder, to high whistles and hisses like birdsong and squeaking hinges. 

The “color” of sound is described by its frequency.  A single note of music has a frequency, the number of times it vibrates in a second.  You see this when a guitar string is strummed – the lowest string, you can almost watch it throb, while the higher strings moves so quickly they’re just a blur.  The moving string makes the air move, and it’s those repeating cycles of the moving air, that’s what we hear because they move our eardrums. 

It makes sense, then, that the frequency of a tone is measured in cycles per second, or cps.  In 1960, this nice clear name was renamed the “Hertz” after Heinrich Hertz, thereby inconveniencing all posterity for the sake of a dead guy, but the term still means cycles per second.  One of these is a hertz (1 Hz), a thousand of these is a kilohertz (1,000 Hz or 1 kHz), a million a megahertz (1 MHz).  The human ear is usually described as hearing 20 Hz to 20 kHz, which is three orders of magnitude or ten octaves.

In between these extremes is the human voice.  The sounds we make when we talk or sing lie mostly between 80 Hz at the low end, and 14 kHz at the higher end.  Vowels and “smooth” sounds are mostly below 4 kHz, while a lot of the consonants that tell one word from another, like “fell” from “sell” are above 4 kHz.

The telephone, however, only carries the thin slice of frequencies from 300 Hz to 3300 Hz.  That’s right, your ears can hear with five times the fidelity of your phone, which is why phones sound so muffled.  This started accidentally in the twenties, because the metal wafers, carbon granule microphones, and cloth-insulated coils they used could do no better, but phones today haven’t gotten much better then they were back then. 

That’s where HD Voice comes in.  “HD Voice” means a phone that has extended its fidelity to at least 7 kHz – doubling the frequency response compared to conventional narrowphones, and dramatically improving the sound and the ease of use.

I’ll be writing more to give a fuller perspective on why that’s important, but in short: by restoring the missing four-fifths of speech that the telephone cuts out, HD Voice boosts accuracy, reduces fatigue, overcomes accents and background noise, and makes telling people apart easier and a lot more natural.