The auto-correlation function in freq2accor is slightly different from what I have normally seen. Normally I have seen this done with zero padding the time domain sample and then taking the DFT, squaring the magnitude and then the IDFT to get the auto correlation. But in the Odas code you take the DFT, square the magnitude, zero pad the result (in the freq domain) and then perform the IDFT. I am guessing that this may be mathematically identical and I was going to test this. Is it equivalent? it appears to work well so I am guessing it is !
The auto-correlation function in freq2accor is slightly different from what I have normally seen. Normally I have seen this done with zero padding the time domain sample and then taking the DFT, squaring the magnitude and then the IDFT to get the auto correlation. But in the Odas code you take the DFT, square the magnitude, zero pad the result (in the freq domain) and then perform the IDFT. I am guessing that this may be mathematically identical and I was going to test this. Is it equivalent? it appears to work well so I am guessing it is !