FM modulation over telephone question

Do you have a question? Post it now! No Registration Necessary

Translate This Thread From English to

Threaded View
I have an analog signal that I am FM modulating between 1600Hz and 2300Hz
that goes over a normal land-line telephone nicely.  But if I try to
transmit over a cordless phone or a cell phone then sometimes I get some
weird artifacts in the received signal.  Anyone know what causes this and
possibly what I can do about it?

Thanks,
--Keith Brafford



Re: FM modulation over telephone question

Quoted text here. Click to load it
cordless phone and cell phone usually sample your signal and then send it to
a analog to digital converter, and compress it. The compression algorithm
usually is tuned to human voice. It might distort your signal.

The first thing I would do is to check the spectrum of your FM signal. The
spectrum is not necessarily bounded in the range from 1600Hz to 2300Hz.

vax, 9000

Quoted text here. Click to load it


Re: FM modulation over telephone question
Quoted text here. Click to load it

Well, there is nothing you can do about it. A cell phone or a cordless DECT
phone does not simply digitize and compress the voice signal. Instead, it
tries to analyze it and break it down into small synthesizable pieces, like
"make a hiss for 2 ms". On the other end, the codes that describe these
"pieces" of voice are used to synthesize the complete voice again. An FM
modulated signal is treated the same way and comes out horribly on the other
end.

Meindert



Re: FM modulation over telephone question
Quoted text here. Click to load it
At the very least, GSM is digital not analogue. It goes through it's own
codec. This assumes you can also turn off the DSP in the phone which is
tuned to voice (it looks to remove "background" noise, echo
cancellation, comfort noise injection, etc.)
Quoted text here. Click to load it

Re: FM modulation over telephone question
Quoted text here. Click to load it

Can you adapt the signal to the characteristics of the codec? AFAIK it uses
a sort of physical model of the speech track, a sawtooth shape base
frequency with some filters for the vowels, plus shaped noise for the
consonants.

Wim



Site Timeline