TLV320AIC3111: Unexpected spike in frequency response of codec

I'm using an analog mic across mic1lp and ground. My output is across SPL

Yes you're right about the setup we're using. However, our earlier project used an analog circuit with op amps and using the same setup, we were getting a proper cutoff at 400 and no spikes as observed now. We're not able to understand why this difference in output should occur.

Souvik, 

You're right, I would still expect to see a cutoff at 400.  But its hard to tell what's going on with some many unknowns.

for me its helpful to see the baseline,  ie what is the response of the ADC with a simple line input.  it might be helpful to know the speaker response and the test environment.

best regards,

-Steve Wilson

Okay I'll do this experiment once and see how it goes.

So I did the experiment today. The output I got is below, which looks pretty good.

Any idea about why there's such a big difference in output between the codec and the analog filter using op-amps?

Thanks for your help.

Souvik, 

This looks better, but still not great.  could it be a measurement issue?

-Steve

Maybe it's due to the out of band noise? I haven't put any filters to remove it and I think the instrument is picking it up.

Souvik, 

I would expect noise to cause more varied results,  this seems pretty consistent.  how are you measuring this?

best regards,

-Steve Wilson

We're using a microphone that measures the speaker output that the codec generates.

We've decided to attribute this response to the non linearity of the test speaker and test microphone for this output response.

When we test the codec directly as you mentioned earlier, we're getting a good response, which is audibly better.