On Jun 30, 10:40�pm, John <sampson...@gmail.com> wrote:
> On Jun 29, 3:31�pm, "volkanizma2003"
>
>
>
> <ozduranvolkan@n_o_s_p_a_m.gmail.com> wrote:
> > �Hi, I have some problems regarding noise normalisation. I have original
> > signal and I want to add a noise.
>
> > EbN0db= 0.00;
> > SNRv=10^(EbN0db/10);
> > global sigma;
> > sigma=1/sqrt(SNRv);
>
> > b=randn(1,1200);
> > gaus=sigma*b;
>
> > x= round(randn(1,1200)); % my original signal
> > y= gaus % my noise
> > z= x+ gauss % original signal + noise
>
> > Question 1.) Do I need to normalize the noise for example multiple by 0.1
>
> > z=x+ 0.1*gauss %will it be like this?
>
> > is this noise`s phase is equal to my original signal or less than my
> > original signal?
>
> > Thanks for your time.
>
> 1) You already scaled the noise when you multiplied it by 1/sqrt(SNR)
> 2) Adding the noise does not change the phase, only the amplitude
>
> John
Actually it does.
Reply by Greg Heath●July 2, 20102010-07-02
On Jul 1, 3:48=A0pm, "volkanizma2003"
<ozduranvolkan@n_o_s_p_a_m.gmail.com> wrote:
> >On Jun 29, 3:31=3DA0pm, "volkanizma2003"
> ><ozduranvolkan@n_o_s_p_a_m.gmail.com> wrote:
> >> =3DA0Hi, I have some problems regarding noise normalisation. I have
> origina=3D
> >l
> >> signal and I want to add a noise.
>
> >> EbN0db=3D3D 0.00;
> >> SNRv=3D3D10^(EbN0db/10);
> >> global sigma;
> >> sigma=3D3D1/sqrt(SNRv);
>
> >> b=3D3Drandn(1,1200);
> >> gaus=3D3Dsigma*b;
>
> >> x=3Dround(randn(1,1200)); % my original signal
> >> y=3Dgaus % my noise
> >> z=3Dx+ gauss % original signal + noise
>
> >> Question 1.) Do I need to normalize the noise for example multiple by
> 0.1
>
> >> z=3Dx+ 0.1*gauss %will it be like this?
>
> >> is this noise`s phase is equal to my original signal or less than my
> >> original signal?
>
> >> Thanks for your time.
>
> >If x is complex why aren't you adding complex noise?
>
> >Greg
>
> Dear Greg, thanks for your time and interest. X is not a complex x is a
> digital signal. it contains 0 and 1 bits.
Look again. You are using round(randn), not round(rand).
Also, why do you reference phase?
> And I add analog noise signal
> into my digital signal. should my noise's amplitude is equal to the my
> original signal's amplitude or less than original signals amplitude. do I
> need to adjust the noise by multiple with 0.1
>
> By the way, when I use Ofdm, I use complex signals as you know. Do I need
> to add random complex noise, or just a random noise. thanks volkan.- Hide=
quoted text -
>
> - Show quoted text -
I really don't know why you are doing this.
If you want a real-valued binary signal
why not
sig(sig>=3D0.5) =3D 1
sig(sig<0.5) =3D 0
Greg
Reply by volkanizma2003●July 1, 20102010-07-01
>On Jun 29, 3:31=A0pm, "volkanizma2003"
><ozduranvolkan@n_o_s_p_a_m.gmail.com> wrote:
>> =A0Hi, I have some problems regarding noise normalisation. I have
origina=
>l
>> signal and I want to add a noise.
>>
>> EbN0db=3D 0.00;
>> SNRv=3D10^(EbN0db/10);
>> global sigma;
>> sigma=3D1/sqrt(SNRv);
>>
>> b=3Drandn(1,1200);
>> gaus=3Dsigma*b;
>>
>> x=round(randn(1,1200)); % my original signal
>> y=gaus % my noise
>> z=x+ gauss % original signal + noise
>>
>> Question 1.) Do I need to normalize the noise for example multiple by
0.1
>>
>> z=x+ 0.1*gauss %will it be like this?
>>
>> is this noise`s phase is equal to my original signal or less than my
>> original signal?
>>
>> Thanks for your time.
>
>If x is complex why aren't you adding complex noise?
>
>Greg
>
Dear Greg, thanks for your time and interest. X is not a complex x is a
digital signal. it contains 0 and 1 bits. And I add analog noise signal
into my digital signal. should my noise's amplitude is equal to the my
original signal's amplitude or less than original signals amplitude. do I
need to adjust the noise by multiple with 0.1
By the way, when I use Ofdm, I use complex signals as you know. Do I need
to add random complex noise, or just a random noise. thanks volkan.
Reply by Greg Heath●June 30, 20102010-06-30
On Jun 30, 11:33�am, "third_person"
<third_person@n_o_s_p_a_m.ymail.com> wrote:
> Can you tell me why you are normalizing it by sqrt(SNR)?
>
> I generate AWGN as
>
> Noise = (randn(1,10)+j*randn(1,10))/sqrt(2)*10^(-SNR/20) �% quadrature
> noise
>
> Am I missing something?
>
>
>
> > Hi, I have some problems regarding noise normalisation. I have original
> >signal and I want to add a noise.
>
> >EbN0db= 0.00;
> >SNRv=10^(EbN0db/10);
> >global sigma;
> >sigma=1/sqrt(SNRv);
>
> >b=randn(1,1200);
> >gaus=sigma*b;
>
> >x= round(randn(1,1200)); % my original signal
> >y= gaus % my noise
> >z= x+ gauss % original signal + noise
>
> >Question 1.) Do I need to normalize the noise for example multiple by 0.1
>
> >z=x+ 0.1*gauss %will it be like this?
>
> >is this noise`s phase is equal to my original signal or less than my
> >original signal?
>
> >Thanks for your time.- Hide quoted text -
>
> - Show quoted text -
There is
SNR = (average signal power)/(average noise power)
and
SNRdB = 10*log10(SNR).
Hope this helps.
Greg
Reply by third_person●June 30, 20102010-06-30
Can you tell me why you are normalizing it by sqrt(SNR)?
I generate AWGN as
Noise = (randn(1,10)+j*randn(1,10))/sqrt(2)*10^(-SNR/20) % quadrature
noise
Am I missing something?
> Hi, I have some problems regarding noise normalisation. I have original
>signal and I want to add a noise.
>
>EbN0db= 0.00;
>SNRv=10^(EbN0db/10);
>global sigma;
>sigma=1/sqrt(SNRv);
>
>b=randn(1,1200);
>gaus=sigma*b;
>
>x= round(randn(1,1200)); % my original signal
>y= gaus % my noise
>z= x+ gauss % original signal + noise
>
>Question 1.) Do I need to normalize the noise for example multiple by 0.1
>
>z=x+ 0.1*gauss %will it be like this?
>
>is this noise`s phase is equal to my original signal or less than my
>original signal?
>
>Thanks for your time.
>
Reply by Greg Heath●June 30, 20102010-06-30
On Jun 29, 3:31�pm, "volkanizma2003"
<ozduranvolkan@n_o_s_p_a_m.gmail.com> wrote:
> �Hi, I have some problems regarding noise normalisation. I have original
> signal and I want to add a noise.
>
> EbN0db= 0.00;
> SNRv=10^(EbN0db/10);
> global sigma;
> sigma=1/sqrt(SNRv);
>
> b=randn(1,1200);
> gaus=sigma*b;
>
> x= round(randn(1,1200)); % my original signal
> y= gaus % my noise
> z= x+ gauss % original signal + noise
>
> Question 1.) Do I need to normalize the noise for example multiple by 0.1
>
> z=x+ 0.1*gauss %will it be like this?
>
> is this noise`s phase is equal to my original signal or less than my
> original signal?
>
> Thanks for your time.
If x is complex why aren't you adding complex noise?
Greg
Reply by John●June 30, 20102010-06-30
On Jun 29, 3:31�pm, "volkanizma2003"
<ozduranvolkan@n_o_s_p_a_m.gmail.com> wrote:
> �Hi, I have some problems regarding noise normalisation. I have original
> signal and I want to add a noise.
>
> EbN0db= 0.00;
> SNRv=10^(EbN0db/10);
> global sigma;
> sigma=1/sqrt(SNRv);
>
> b=randn(1,1200);
> gaus=sigma*b;
>
> x= round(randn(1,1200)); % my original signal
> y= gaus % my noise
> z= x+ gauss % original signal + noise
>
> Question 1.) Do I need to normalize the noise for example multiple by 0.1
>
> z=x+ 0.1*gauss %will it be like this?
>
> is this noise`s phase is equal to my original signal or less than my
> original signal?
>
> Thanks for your time.
1) You already scaled the noise when you multiplied it by 1/sqrt(SNR)
2) Adding the noise does not change the phase, only the amplitude
John
Reply by volkanizma2003●June 29, 20102010-06-29
Hi, I have some problems regarding noise normalisation. I have original
signal and I want to add a noise.
EbN0db= 0.00;
SNRv=10^(EbN0db/10);
global sigma;
sigma=1/sqrt(SNRv);
b=randn(1,1200);
gaus=sigma*b;
x= round(randn(1,1200)); % my original signal
y= gaus % my noise
z= x+ gauss % original signal + noise
Question 1.) Do I need to normalize the noise for example multiple by 0.1
z=x+ 0.1*gauss %will it be like this?
is this noise`s phase is equal to my original signal or less than my
original signal?
Thanks for your time.