discuss-gnuradio
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Discuss-gnuradio] question to measure SNR


From: Johnathan Corgan
Subject: Re: [Discuss-gnuradio] question to measure SNR
Date: Wed, 25 Mar 2009 12:55:33 -0700

On Wed, Mar 25, 2009 at 5:20 AM, feldmaus <address@hidden> wrote:

> i saw there is a function called <gr_probe_mpsk_snr_c>,
> but i doesn't know how to get this to work.

This a sink you can attach to a raw MPSK constellation complex sample
stream, and it will estimate the signal-to-noise ratio using the
"squared signal over noise variance" (SNV) technique for SNR
estimation.  It's simple and fast, but not all that accurate below
about 8 or 9 dB.

> Is there a example which is using this function ?

It is used in the bit error rate tester (digital-bert) example
receiver.  The block is designed to be queried for its current SNR
estimate by an external thread (created by the user outside the
flowgraph).  In the BERT receiver, a background thread once per second
queries the flowgraph for the current SNR, bit error rate, frequency
offset, and timing error, then displays these to the screen.

> Can i use this to measure my SNR ?
> I have a test sin signal.

No, it is explicitly designed for BPSK or QPSK signals, not carriers.
Importantly, it needs to follow a frequency/phase recovery loop such
that there is no residual carrier in the signal.

> I also found the possibility by measuring the SNR by
> examine the bit-error-rate.
>
> Are there any example which do this ?

It the digital BERT example, you can see the BER and SNR once per
second, and in live testing, they do correlate.  But you can't assume
a fixed relationship between BER and SNR except in a the ideal ("not
real world") case of pure additive white gaussian noise.

Johnathan




reply via email to

[Prev in Thread] Current Thread [Next in Thread]