Here is a frequency response plot of a computer-generated test tone.

Original test tone

Here is the spectral plot of the same original tone.

Original test tone spectral

Here is the frequency response plot of the same test tone after feeding it into the mic input socket of the Pocket 2 handle.  It's actually pretty good.  Mostly flat, with a bass roll-off from 90Hz, but generally I'd say it exceeds expectations.

P2 line in response

But this is the spectral response of that tone fed into the P2 mic socket.  It should be a single line going from bottom left to top right.  Instead, there are multiple lines at intervals of 4kHz, and there are other lines going from top left to bottom right, which can be heard as descending tones, in contrast to the main ascending tone.  Not good - but in my experience only a problem on certain audio material.  (Compare this with the original, second plot down from the top of this post).  Instinct tells me this is happening in the AAC encoder.  All 'lossy' audio encoders mangle the audio a fair bit in data-compressing the original sound so that the file size is much reduced, but it is done in such a way as the ear is fooled into not noticing what is going on.  You can fool the ear but not the spectral analysis plot!  (Comments and corrections welcomed).

P2 line in spectral

Testing the mics themselves requires proper facilities.  All I can do is a crude comparison with another audio device.  This test is done by simply playing back the original tone on monitor speakers, while recording on the P2 mics (set to "all", "stereo") with the device held close to the speakers.  The result not only shows the P2 response, but also the response of the speakers and the acoustic character of my studio.  Here is the frequency response of the P2 mics (in that crude test).  Although it's by no means flat, it's not nasty looking, and shows that there's some activity right up to 18kHz.  It may not be a flat line, but it's a reasonably straight one.

P2 internal mic response

And here's the spectral display of the same thing.  Again, all sorts of extra stuff in there.  Hmmm.

P2 internal mics spectral

For comparison, here's the same two charts from the same test, using the mics of a Zoom H1n recorder (mp3).  I chose that because it's a popular inexpensive recorder, and I think a fair comparison.  The frequency response holds flatter up to 10kHz but then rolls off quite steeply (so it would be expected to sound duller than the Pocket 2).  

Zoom H1n mic response

The Zoom H1n spectral response of its mics is much cleaner than that of the Pocket 2.  There's still some extra stuff visible, but not nearly so much.

Zoom H1n mics spectral

Conclusion?  The external mic input has a flat frequency response and shouldn't degrade the audio quality of any sensible mic you connect to the Pocket 2.  The built-in mics have a perfectly adequate-looking frequency response with no nasty peaks (but remember my test isn't the best), but there's some pretty odd stuff happening with spurious tones generated on the external mic input and on the mics themselves.  What first stimulated my curiousity about this was when I did a key-jangle test on the Pocket 2 - you just go round the device jangling a bunch of keys - and the result on playback was all rattle and no jangle.  


But at the end of the day, what matters is what the Pocket 2's recordings of real-world environments and music and speech sound like, and on that basis, I'd say DJI have done a very good job on the audio side.

Hello there. Thank you for reaching out and for posting these information on the audio tests you have made for the DJI Pocket 2. Again thank you for your support and keep on filming.

Thanks for the info!

Great info - thanks

Very nice test of Pocket 2 audio.