Why, with all our advancements in telecommunications and phone technology, has phone call audio quality stayed virtually the same as ten or twenty years ago?


Why, with all our advancements in telecommunications and phone technology, has phone call audio quality stayed virtually the same as ten or twenty years ago?

In: Engineering

If you talk about landlines or anything that is talking to landlines: That is a dead technology, don’t expect anything there.

If you talk about VoIP: That is much better than years ago, with a better sampling rate and wider spectrum they capture.

However, that is part of the problem, the biggest problem is the transfer of the voice of the speaker to the microphone: Some microphones are crap (cheap headset), or not at the right place (laptop, or not close by the mouth) or interference with the environment (wind, background noise).

There is only so much technology can fix :-/

You won’t get movie quality because that’s rerecorded in the studio.

Of all the things that cell phones do better than land lines, the sound quality was so much better.

In simplest terms; because the telephone system is an internetwork of individual links, it is limited by the least capable link in the system. Think of having a road between two cities that starts out as a five-lane Interstate highway, drops to a dirt road, then goes back to a five-lane highway and then arrives at the destination – the traffic is limited to the capacity of the stretch of dirt road in the middle.

In standard telecoms that limitation is the enduring use of [G.711 encoding](https://en.wikipedia.org/wiki/G.711) to convert analog sounds from the two ends of the call into digital information that will be carried over digital [trunk circuits](https://en.wikipedia.org/wiki/Trunking) in the middle. Those circuits, at least in North America, are built around the [T-carrier architecture](https://en.wikipedia.org/wiki/T-carrier) (with the ulaw variant of G.711), which has a hard limit of 64kbps of information per circuit which, skipping over a whole bunch of technical stuff, translates into only carrying sounds between [300 and 3400Hz](https://en.wikipedia.org/wiki/Voice_frequency) – everything else within the typical range of hearing of 20–20,000Hz is simply discarded.

Various methods of delivering [wideband audio](https://en.wikipedia.org/wiki/Wideband_audio) over conventional digital telephone networks do exist (eg. [G.722](https://en.wikipedia.org/wiki/G.722)), but if the call will eventually terminate on a conventional analog telephone (or cellphone) then the extra effort to capture and transmit that extra audio information just goes to waste as the signal must be [transcoded](https://en.wikipedia.org/wiki/Transcoding) to meet the capability of the lowest common denominator in the end-to-end circuit.

On the other hand with pure VOIP calling, for example with Skype-to-Skype calls, most of the internetworking issues are eliminated as the sound is transmitted entirely as a stream of data which is controlled exclusively by the two ends of the call, and so the system can transmit as much audio spectrum as the designers wish.

Cell phone audio quality used to be remarkably clear. As good as landline communication. But that was when it was analog and a full three watts.

Modern cell phones (the last 15 years or so) use digital, low power signals. And the quality is pathetic.

Probably one of the really big reasons people don’t like to talk on the phone in modern times is because the quality is so horrible.

The audio quality of a mobile phones is based on following aspects: note this applies to both devices, originating and terminate end.
– bandwidth assigned to the voice channel
– vocoder technology used
– microphone quality of mobile phone
– speaker of device

Now, the best technology to measure Voice quality is based on a set of standards, which for VoIP technology, is PESQ (google it). PESQ measures the variance of signal and creates a score using various large database that has relative voice metrics to what it listens to. It scores it based on a 1-5 score. 1 means not understandable and 5 being it sounds perfect (original audio from originating device)

Since the 1990s, the quality has improved by 1 point. In 2G and 3G technologies, it was 3-3.5. In 4G, it has been improved towards 4.2-4.5. Note this is log scoring, so as you get closer to 5 it’s a huge incremental improvement.

One commenter brought up a good technical specs on the G.711.