curve of BER in ldpc codes in matlab

hello every body.
I wrote a plan for ldpc code an get these result but it is different the result that recieved from papaers.can any boby tell me why the result is different?


I wrote these code for n=1210 and frame_num=2000, code rate=0.5;
I used these code for channel
[m,n]=size(H);
sigma=sqrt(10^(-SNRindB1/10));
msg = randint(1,n-m);
c=rem(msg*G,2);
modObj = modem.pskmod('M',2,'InputType','Bit');
modulatedsig = modulate(modObj, c);
receivedsig = awgn(modulatedsig, SNRindB1(I), 0); ...
% Compute log-likelihood ratios (AWGN channel)
demodObj = modem.pskdemod(modObj,'DecisionType','LLR', ...
'NoiseVariance',sigma);
LLR = demodulate(demodObj, receivedsig);

and I expect nearly these result(fast QC codes figure) :at low value SNR ,bit error rate value be about 10^(-1)

can u write BER performance of non binary LDPC min sum decoding???pls help me....i tried a lot..

hi. I just worked on binary decoding

it differs from binary....

Don't get what you mean. We normally compare BER at 10^-5. If you have a simulation results similar to the paper one at 10^-3 to 10^-5. Then, likely you're doing something right. Bear in mind, BER of coded system is often worse than uncoded system at low SNR.

Hello every body,
any one can help me how to write a code can plot y axis in log scale and x axis in normal scale.

thanks,

- - - Updated - - -

What i am trying to say is i've got results displayed for BER using matlab simulink but i don't know how to plot curve of BER.

i would be so appreciated if someone could help out,

thanks

Use Microsoft Excel.