#### lines

##### Newbie level 6

**(anybody Help me )Questions about Sigma-delta in dsp .**

Hi,all

As we know, the output of the Sigma-delta modulator are bitstreams, such as ±1. In order to evaluate the performance of the modulator, we should use dsp technology.

I download a program from the net, which can calculate the SNR and PSD of the bitstreams. (the program is below)

My questions are:

1. why we should extracts sinusoidal signal?

2. If the bitsteams are not ±1, for example, they are 2bits Quantizer. Then the Vref is what , 1 too?

Here is the program:

function [snrdB, ptotdB, psigdB, pnoisedB] = calcSNR(vout,f,fB,w,N,Vref)

% SNR calculation in the time domain (P. Malcovati, S. Brigati)

% vout: Sigma-Delta bit-stream taken at the modulator output

% f: Normalized signal frequency (fs -> 1)

% fB: Base-band frequency bins

% w: windowing vector

% N: samples number

% Vref: feedback reference voltage ———————— If the bitsteams are not ±1, for example, they are 2bits Quantizer. Then the Vref is what , 1 too?

%

% snrdB: SNR in dB

% ptotdB: Bit-stream power spectral density (vector)

% psigdB: Extracted signal power spectral density (vector)

% pnoisedB: Noise power spectral density (vector)

%

fB=ceil(fB);

signal=(N/sum(w))*sinusx(vout(1:N).*w,f,N); % Extracts sinusoidal signal ——————why we should extracts sinusoidal signal? is it in common use ?

noise=vout(1:N)-signal; % Extracts noise components

stot=((abs(fft((vout(1:N).*w)'))).^2); % Bit-stream PSD

ssignal=(abs(fft((signal(1:N).*w)'))).^2; % Signal PSD

snoise=(abs(fft((noise(1:N).*w)'))).^2; % Noise PSD

pwsignal=sum(ssignal(1:fB)); % Signal power

pwnoise=sum(snoise(1:fB)); % Noise power

snr=pwsignal/pwnoise;

snrdB=dbp(snr);

norm=sum(stot)/Vref^2; % PSD normalization

if nargout > 1

ptot=stot/norm;

ptotdB=dbp(ptot);

end

if nargout > 2

psig=ssignal/norm;

psigdB=dbp(psig);

end

if nargout > 3

pnoise=snoise/norm;

pnoisedB=dbp(pnoise);

end

Last edited: