I am designing an adaptive noise cancellation system. I have the LMS algorithm running perfectly. I defined the SNR for the 'voice + noise' before adaptive noise cancellation. I am just wondering how to find the SNR for the output of the adaptive noise cancellation system (it is mostly greater than before noise cancellation)?
SNR is a quantitative value calculated to see how good ur system works. So, to test your system u should have an uncorrupted signal which is deliberately corrupted with a known noise. At the end of the learning, subtract the known input signal from the output, which technically should leave you with the noise that is not cancelled by the system. Frm there SNR of the output is a direct calculation.