Why do Voip phones sound so bland?

9 views
Skip to first unread message

Brian Gaff (Sofa)

unread,
Oct 23, 2021, 5:22:47 AM10/23/21
to
IE you seem to find a very compressed gritty sounding bandwidth limited
audio often with tiny glitches when companies are obviously using voip, even
over local connections. With today's technology we should be able to always
get something like facetime audio call quality. It would surely help those
with hearing loss, and also when you hear automated call centres that give a
message then say press 1 for this, all obviously digital, cant they get
their levels right. You find the message is quiet, and the menu loud.
This can be done in software quite easily.
Brian

--

This newsgroup posting comes to you directly from...
The Sofa of Brian Gaff...
bri...@blueyonder.co.uk
Blind user, so no pictures please
Note this Signature is meaningless.!


Andy Burns

unread,
Oct 23, 2021, 5:32:19 AM10/23/21
to
Brian Gaff wrote:

> IE you seem to find a very compressed gritty sounding bandwidth limited
> audio often with tiny glitches when companies are obviously using voi

what are you using at your end? If it's a POTS phone that probably limits the
call to G.711, but that should be no worse than a PSTN call. If you're using a
mobile you might be limited to G.729 codec?

If you have a voip phone, then subject to what can be agreed between the
endpoints, much better wideband codecs are available.

tony sayer

unread,
Oct 23, 2021, 6:38:09 AM10/23/21
to
In article <iti351...@mid.individual.net>, Andy Burns
<use...@andyburns.uk> scribeth thus
Indeed been using VoIP here for must be 10 years now! But i do know some
users where they do get "gritty" lack of available bandwidth usual
cause..
--
Tony Sayer


Man is least himself when he talks in his own person.

Give him a keyboard, and he will reveal himself.


David Woolley

unread,
Oct 23, 2021, 7:17:20 AM10/23/21
to
On 23/10/2021 10:32, Andy Burns wrote:
> If you're using a mobile you might be limited to G.729 codec?

Standard mobile phone codecs are from the GSM family, although I think
there is a wider band (G.723?) one now being used in some services, but
it is still a voice only one. I'm not aware of the UK landline network
using anything more than G.711, though, which is a 3.1kHz audio
(300-3,400 Hz) codec (sampling at 4kHz, which is basically the same
frequency response as mid-twentieth century analogue carrier systems for
inter-exchange calls.

Woody

unread,
Oct 23, 2021, 7:30:30 AM10/23/21
to
That last comment puzzles me. Single audio circuits are usually 64Kb
which is 8 bits at 8000 samples/second. Nyquist theory says that the
sampling rate should be at least twice the highest frequency in use and
as a line circuit is technically 4KHz hence the 64Kb.

Or do you know something different?

Andy Burns

unread,
Oct 23, 2021, 10:10:53 AM10/23/21
to
David Woolley wrote:

> Standard mobile phone codecs are from the GSM family, although I think there is
> a wider band (G.723?) one now being used in some services, but it is still a
> voice only one.  I'm not aware of the UK landline network using anything more
> than G.711, though, which is a 3.1kHz audio (300-3,400 Hz) codec (sampling at
> 4kHz, which is basically the same frequency response as mid-twentieth century
> analogue carrier systems for inter-exchange calls.

mobile phones are certainly capable of running better codecs, but whether mobile
networks do accept them, when connecting a call to a voip user, I don't know ...
it will of course cost them more bandwidth on the airside and depending how they
terminate the call, I have no idea if they coul link direct to internet, or they
would all be routed over PSTN at 64kbps

tony sayer

unread,
Oct 23, 2021, 6:08:35 PM10/23/21
to
In article <sl0rol$6f0$1...@dont-email.me>, Woody <harro...@ntlworld.com>
scribeth thus
Have a butchers for "Vocoders" Woody....

Woody

unread,
Oct 24, 2021, 3:50:26 AM10/24/21
to
On Sat 23/10/2021 23:04, tony sayer wrote:
> In article <sl0rol$6f0$1...@dont-email.me>, Woody <harro...@ntlworld.com>
> scribeth thus
>> On Sat 23/10/2021 12:17, David Woolley wrote:
>>> On 23/10/2021 10:32, Andy Burns wrote:
>>>> If you're using a mobile you might be limited to G.729 codec?
>>>
>>> Standard mobile phone codecs are from the GSM family, although I think
>>> there is a wider band (G.723?) one now being used in some services, but
>>> it is still a voice only one.  I'm not aware of the UK landline network
>>> using anything more than G.711, though, which is a 3.1kHz audio
>>> (300-3,400 Hz) codec (sampling at 4kHz, which is basically the same
>>> frequency response as mid-twentieth century analogue carrier systems for
>>> inter-exchange calls.
>>
>> That last comment puzzles me. Single audio circuits are usually 64Kb
>> which is 8 bits at 8000 samples/second. Nyquist theory says that the
>> sampling rate should be at least twice the highest frequency in use and
>> as a line circuit is technically 4KHz hence the 64Kb.
>>
>> Or do you know something different?
>
>
> Have a butchers for "Vocoders" Woody....
>
Ah......

Brian Gaff (Sofa)

unread,
Oct 24, 2021, 5:08:53 AM10/24/21
to
I can hear the difference even over my old school landline. It just sounds
flat bland and muffled to me, and all the other things, when their olds
systems used to sound more like telephones are supposed to sound.
Brian

--

This newsgroup posting comes to you directly from...
The Sofa of Brian Gaff...
bri...@blueyonder.co.uk
Blind user, so no pictures please
Note this Signature is meaningless.!
"Andy Burns" <use...@andyburns.uk> wrote in message
news:iti351...@mid.individual.net...

David Woolley

unread,
Oct 24, 2021, 7:05:35 AM10/24/21
to
On 23/10/2021 12:30, Woody wrote:
> That last comment puzzles me. Single audio circuits are usually 64Kb
> which is 8 bits at 8000 samples/second. Nyquist theory says that the
> sampling rate should be at least twice the highest frequency in use and
> as a line circuit is technically 4KHz hence the 64Kb.
>
> Or do you know something different?

It's impracticable to make brick wall anti-aliasing filters, and if you
could, I think they would probably ring badly. The nominal 3.4kHz cut
off means that anything that would alias down from above 4kHz is well
attenuated.

You can only get the full Nyquist range if your sampled signal doesn't
contain any frequencies outside a range which spans half the sampling
frequency.

The old carrier systems were 4kHz spaced channels on a single sideband
suppressed carrier system, and, again, you needed realisable channel
filters to avoid adjacent channels interfering with each other.

David Woolley

unread,
Oct 24, 2021, 7:08:10 AM10/24/21
to
On 23/10/2021 23:04, tony sayer wrote:
> Have a butchers for "Vocoders" Woody....

Vocoders relate to mobile phone codecs, and ones like G.729. They are
the ones that provide only a speech service.

Woody

unread,
Oct 24, 2021, 7:56:06 AM10/24/21
to
The audio passband of landlines may nominally be 4KHz with an actual
speech band of 3400Hz, but in practice and my experience the speech band
is -3dB around 2KHz and passes very little much above 2K5Hz.

[Interestingly NASA actually break the speech audio bands of astronauts
into three ranges although the actual frequency ranges escape me at this
minute. It was something like 300-500Hz and 1400-1700Hz for both sexes,
but the mid-band range was higher for women than for men to match their
inherent speech pitch. I am of the understanding that it was done to
save data usage. Whether they still use it I know not.]

I suspect the main reason for a 4KHz bandwidth - apart from its radio
heritage - was the use of a 3825Hz tone to indicate an unused telephone
circuit, known in the US as 'whistlers.' That was how, in eons gone by,
people were able to make international calls (IME largely initiated in
the US) free of charge.

David Woolley

unread,
Oct 24, 2021, 10:31:53 AM10/24/21
to
On 24/10/2021 12:56, Woody wrote:
> a 3825Hz tone to indicate an unused telephone circuit,

The idle tone used to be 2600 Hz, which is where the name of the
alt.2600 newsgroup is derived from (I don't know if that still exists).
There may have been other tones, but 2600 Hz was the main one in the
USA, where blue boxing would have been most common.

Woody

unread,
Oct 24, 2021, 11:11:48 AM10/24/21
to
You are of course correct. 3825Hz is used by mux cards for sync. My
mixed up thinking.....

notya...@gmail.com

unread,
Oct 24, 2021, 2:34:21 PM10/24/21
to
You probably have the wrong codec selected. Voip to Voip or HQ mobile usually better than POTS.

Brian Gregory

unread,
Oct 24, 2021, 7:09:03 PM10/24/21
to
Most likely the ones where you can hear the difference are doing it
wrong. Perhaps they're not giving VOIP priority so that congestion due
to other data carried over the same internet connection messes it up.
Perhaps they're using really cheap headsets or really cheap ATAs.

On 24/10/2021 10:08, Brian Gaff (Sofa) wrote:
> I can hear the difference even over my old school landline. It just sounds
> flat bland and muffled to me, and all the other things, when their olds
> systems used to sound more like telephones are supposed to sound.
> Brian

--
Brian Gregory (in England).

David Woolley

unread,
Oct 25, 2021, 7:54:33 AM10/25/21
to
On 24/10/2021 19:34, notya...@gmail.com wrote:
> You probably have the wrong codec selected. Voip to Voip or HQ mobile usually better than POTS.
Very commonly used codecs are A-law, and µ-law (often mis-transliterated
as ulaw), which are the rest of the world and US PSTN codecs, and are
variants of G.711. G.729 and GSM are commonly used. These are voice
only and the same audio bandwidth, so they are worse than PSTN. (I've
observed that some people select G.729 because they think it is higher
quality, when that is only relative to simpler voice only codecs, and
the real benefit is lower bit rates.)

HD voice on mobiles has a higher audio bandwidth, but is voice only, so
it is difficult to classify, although you could look up its MOS (mean
opinion score), a subjective measurement, averaged over a group of users.

There are wide band codecs, but they are probably mainly used within
executive suites.

The most commonly used codec that has a better MOS than G.711 is G.722.
That is basically because it is wider band, and that appears to more
than compensate for its having a lower noise margin.

Typically, when you are calling a business, you are calling a call
centre, which may be in a distant country. The reason for using VoIP in
that case is cost reduction, rather than the wider choice of codecs, so
they may well be using G.729 over a congested internet connection.

tony sayer

unread,
Oct 29, 2021, 9:55:30 AM10/29/21
to
In article <sl3eqp$9jh$2...@dont-email.me>, David Woolley <da...@ex.djwhome
.demon.invalid> scribeth thus
We had, a while ago, a few MVP-410 Multivoip devices around that
provided speech bandwidth over IP circuits in 2 ands 4 wire formats plus
E&M signalling. I remember that most every codec around was selectable
in them!

Since the demise of Two way radio no longer used!....
Reply all
Reply to author
Forward
0 new messages