I have tried to estimate the SNR of the sigma delta conversion after the sinc^3 filter and got lots of noise at the low frequency end. It looks like the opposite of noise shaping, see attached. When fft of the bitstream is plot the SNR is almost twice as high for the same frequency. Am I using the sampling rate a lot higher than the signal I am looking at?
Thanks!
Never mind, I have found the error in my code - needed to account for the transition delay of the decimation filter. After I discarded those initial bits from the bitstram (the higher the decimation rate - the more bits need to be discarded) the SNR is correct.