Say we have number of samples to be 128 and a signal sin(2*pi*1000*n*Ts)+sin(2*pi*2000*n*Ts) sampled at 8kHz.
Now say we have same signal but only N = 64 samples. But in order to create 128 samples, we pad the signal with 64 zeros.
Each of these scenarios the frequency resolution is 8000/128 = 62.5Hz, but the spectrum of the first one is much better, as in, you will have two clear peaks. Why is that?
Thanks in advance