Not 0. The value in the first location is enough larger than the rest of the data, that the rest of the data appears close to the axes in comparison. Below I skip plotting the first point.
Remember that fft() and ifft() are close to being the same operation, so when you ifft() your time domain signal, that is close to being the same as doing fft() of the time domain signal. And, just like if you had done fft(), the first output bin will be the sum of the time domain signal, so if your time domain signal has a non-zero mean then the first output bin can end up being much higher in magnitude than the other outputs.
cp_len = floor(0.1 * block_size);
data_source= abs(round(randn(1,no_of_data_bits)));
stem (x*(1/Fm),data_source);
xlabel('time(Microsecond)','Fontsize',14);
ylabel('amplitude','Fontsize',14);
title('Transmitted Data','Fontsize',14);
qam_modulated_data = qammod(data_source, M);
nqdata = length(qam_modulated_data);
scatterplot(qam_modulated_data,1,0,'r*');
[udata, uidx] = unique(qam_modulated_data);
text(real(udata(k))-0.4,imag(udata(k))+0.4,num2str(data_source(uidx(k))));
qm = abs(qam_modulated_data);
title('MODULATED TRANSMITTED DATA','Fontsize',14);
y=ifft(qam_modulated_data);
stem(x(2:end)*Fm,abs(y(2:end)));
ylabel('amplitude of ifft');
title('without hermitian ifft','Fontsize',14);
Best Answer