DJI Mic 2 32-bit float - Overhyped?

In-depth look at the 32-bit float recording capability of the new DJI Mic 2 system, including overview of the technology and its benefits, the workflow of using it and a side-by-side comparison of standard 16-bit audio to 32-bit audio.

Keywords:
DJI Mic / Mic 2 wireless microphone system, bluetooth, 2.4 GHz, transmitter, receiver, range, vlogging, content creation, YouTube, instagram, 16-bit, 24-bit, 32-bit float, noise cancelling, DJI Osmo Pocket 3

Chapters:

00:00 - 02:05 Introduction
02:05 - 08:45 What is 32-bit Float?
08:45 - 13:00 The 32-bit Workflow
13:00 - 17:10 Audio Comparison Tests
17:10 - 18:50 Conclusion / Wrap-Up

Product Links (I may earn a commission):

Other Recording Equipment

[AlexTheTennisPro]

finally some info, not just sells info from sponsored videos

[wilri]

Audio level on cameras often have an adjustment for compressing the sound that is too loud so it doesn't clip. But then the dynamic range is limited. So if you want a high dynamic range, then the 32 bit float is the way to go. But as you say, for most situations, the compressing of loud sounds is perfectly acceptable. Perhaps it's a good comparison to D Log video which requires extra work in post and for most people is not needed. I'm glad this camera has both for those times when it matters, even if that's 1% of the time.

[erbartlett]

Great comparision, super helpful

I dont think your audio in your "loud" test clipped enough to test the 32 bit float feature. Unless you fully peaked and distorted your 16 bit audio... you probably weren't loud enough.

[OvertonWindex]

Pretty great when youre talking AND shooting guns.

[ThomsenTower]

I see not so much the 32bit part, but the independent recording directly on the mic 1) as a backup should your mic lose connection with the camera, 2) as an option to use the mic all by itself to record yourself in different settings. Am I mistaken?

[ChadWork]

Did you have noise canceling turned on? The audio had annoying artifacts that sort of garbled the high end.

[MrJohnmanjiro]

Good job. I'm interested in this camera, and I've watched a bunch of reviews. Most of them parrot the 32-bit float capability, but this is the first time that I've seen anyone explain it in a logical way. The mic 2 looks good from a convenience point of view, but 32-bit float is basically marketing hype.

[denniswong]

Thanks for the informative video.

I'm not a sound expert. I'm listening to the sound comparison with a set of earphones. I couldn't hear any clipping on the 16bit audio. At least I couldn't hear any electronic high pitch sound as a result of the clipping.

[Thatgirlgigixo]

So 32 bit only works if you manually match up the audio.. and on top of that not forget to press record also on your mic..? And if I understand correctly this 32 bit doesnt work if you record on your phone and use a simple editing app like inshot or capcut right? So its better to turn it off then? Or does it not even matter?

[devilik]

is the soud from DJI so bad as you be talking to some metal can or did you set something wrong... or did you changed the mic you use during the video?

[MasonAyresCreative]

Dude I literally just made a video about this, and this pops up on my feed the next minute. I've capped out the mics many times while screaming, it seems like the hardware has a limitation that bottlenecks the 32 bit float anyways.

[EdwardKilner]

Not a bad video, at all.

You forgot the sign bit, to indicate positive and negative values. The exponent is called the mantissa. I have never seen how many bits are used by DJI for it. Nor if the other manufacturers use the same number of bits. The remainder of the 32 bits are a binary fraction, to be used with the mantissa and sign bit to record the value of a single digital sample of the audio input.

You correctly pointed out that 32 bit float usage involves extra work. And, it might not be worth it! Or, it might be, to some of us, sometimes, as a safety net for those unexpected wide dynamic range events.

When you did the 32 vs 16 bit tests, something was wrong with the audio I listened to. Made the comparison useless. Audio was noticeably inferior to the preceding video.

Finally, I find the transition you favour and use to be annoying. A simpler one would be fine.

When you did the