Factor used to scale Noise in 4QAM simulation ???

Hi,
i am simulating the 4QAM in Matlab. there i need to scale the noise to get the SNR vs BER graph. Can any one of you plz tell me the factor by whcih we can scale the noise...and also how u get this factor?

P.S. my noise if of type ... noise = randn(1,N) + i * randn(1,N);

Waiting.....

hi
refer to my posts in this link : "Channel Capacity in AWGN"
h**p://

regards

i have checked the link but am not still cleared about it. I have got the theoratical SNR vs BER graph by eq 0.5*erfc (sqrt(10^(snr/10))) but the graph of pratical SNR vs BER is not following the theoratical. The only thing which comes to my mind is that may be i am not scaling the noise correctly. I would be grateful if you help me out and tell me the scaling factor for 4QAM.

Waiting....

bee said:

Hi,
i am simulating the 4QAM in Matlab. there i need to scale the noise to get the SNR vs BER graph. Can any one of you plz tell me the factor by whcih we can scale the noise...and also how u get this factor?

P.S. my noise if of type ... noise = randn(1,N) + i * randn(1,N);

Waiting.....

Click to expand...

You can refer to Proakis's book. Basically, you just need to calculate the average energy of your constellation (for you 4QAM, the energy is same for each point), denoted Es, then you simply use your Es/No value to calculate the noise power.

For a numerical example, suppose you want to simulate the BER at SNR=10db. You generate the 4QAM points (I am not sure if Matlab normalizes the energy or not, probably does. Please check it out. We assume it does so that Es=1). Convert SNR from dB to linear, which gives you 10. Then, the noist power is given by
SNR=Es/No --> No=Es/SNR=1/10=0.1. Then you split the noise power equally to its real/imag parts.
noise=sqrt(0.1/2) * (randn() + j*randn())
Done!

bee said:

i have checked the link but am not still cleared about it. I have got the theoratical SNR vs BER graph by eq 0.5*erfc (sqrt(10^(snr/10))) but the graph of pratical SNR vs BER is not following the theoratical. The only thing which comes to my mind is that may be i am not scaling the noise correctly. I would be grateful if you help me out and tell me the scaling factor for 4QAM.

Waiting....

Click to expand...

hi

practical BER vsSNR always is worse that theoretical curve , because in practice many unwanted patameters affect the system which are not taken into account in theory . In derivation of the theoretical curve , we always assume an AWGN channel and do not consider other effects.


regards