@11025Hz from the microphone input
there is a dc - offset on the sampled values. The dc offset is up to 5% and is
dependent on the amplification factor of the programmable gain amplifier.
Autocalibration does *not* help, there is still the same DC offset after
calibration. How can we get rid of the DC offset?
The inputs to the AD1845 are internally biased around the reference voltage.
Microphone signals are generally AC coupled into L_MIC and R_MIC to remove any
DC bias level. I understand that in some speech recognition systems, it is
necessary to DC couple inputs. You can DC couple inputs but you need to ensure
that the input signal is referenced to Vref . If you couple the signal into the
AD1845 directly and do not perform the level shifting function, you will see a
DC offset in the output.