Posted on February 01, 2018 at 00:57
Based on experiments, Fs=5*Fspidfrx_clk/((WIDTH5+2)*64) gives more precise estimate.
(Amazingly enough, even for the values given in the example this formula gets closer to 44100Hz).
I understand that the granularity is still too big for any practical precise measurement of the sample rate, but why not make things as good as they can be?
JW
