every mic is analogue, unless it is a laser microphone that is sampling the diffraction of the beam or something. For example the Neumann Solution D microphones are analogue microphones, but marketed as digital. There are a few companies that offer AES 42 digital mics - the AES specification that has been embraced as the physical transport of digital signaling for microphones - such as Sennheiser and Schoeps. The mics that are marketed as digital have a minimal matching preamp or no preamp that exists between the capsule and the ADC and output an AES42 signal to a recorder - but the sound capture is all done in the analog domain with a mechanical membrane like it has been done since the dawn of microphones.
now the Question of 32 bit audio... the purpose of 32 bit audio wasn't really to improve sound quality per se... but to make it sound better in wider ranges of situations. Back in the day when we had 16bit CDs, it was sort of acknowledged that 16 bit was more than enough for the average listener to hear "perfect" sound reproduction. As technology progressed, some emerging formats tried to increase the bit depth in order to improve 16 bit "redbook" digital audio. So for example, 20 bit HDCD allowed for greater than 16 bit reproduction, where the average person probably wouldn't really notice the difference, but for audiophiles who had good rooms and great equipment, and had a personality disorder to spend a lot of time A/B comparing their music over and over again, it was possible to detect a difference. So for the average population, our ears' resolving power probably lies somewhere in the 14 to 18 bit range. When recording studios moved away from tape and into digital recording formats, we didn't want a format that was equal to CD, wanted something with greater dynamic range so that raw sounds could be recorded, but then manipulated in post (much like recording in 12-15 stop log so that we can deliver a 7 stop consumer render)... so 24 bit was selected as the bit depth that would give more than enough audio quality. The other problem is that electronics are not completely quiet, so even though a 24 bit file is recorded, maybe 18 to 20 bits of that is actually useful information. Here is one area where AES42 digital mics help, by putting the ADC converter right at the capsule, it gives the mic the best possible outcome in signal to noise ratio. It's certainly not 24 bits of resolution, but better than cabling to a preamp and adding noise from gain and subsequent interconnects and line stages. So when we record 32 bit audio, we aren't departing into some new realm where we've invented a better way for audio engineers to hear, what it does allow us to do, though, is to unpack a wider dynamic range back into a 24bit workflow. In the most simplistic terms, it allows the gain pot to be removed from the equation, so that instead of matching the gain to an appropriate level, the sound is simply recorded at a very low level and then when working in a 24 bit workflow, the signal can be digitally gained back up without dangerously pushing against the noise floor. (technically not true, but close enough for illustration). Now theoretically, if a signal was recorded in plain old 16 bit audio, more or less the resolution of the ear, but at perfect levels, for all intents and purposes it is just as good as 32 bit audio. Maybe not for the "air" around cymbal brushes or delicate nuance of the wood mallet hit on a bell, etc... but for good old human dialogue, which is a relatively low-fi sound source of limited dynamic range. In other words in-camera audio can be just fine, even for critical professional work, but just don't mess up the gain levels (and use a professional camera with good audio circuits - talking about ENG style broadcast cameras, not DSLRs or prosumer cams). When you hire a sound recordist on set, he's recording 24 bit audio. So he can intentionally record at 10 to 20 dB lower than he knows the material will be final mixed at and still not have an issue with. Now 10-20dB is a pretty huge range. I can have someone whispering, peaking at -40dBfs and then have that person scream at the top of their lungs and unless a drill sergeant or opera singer, probably not going to clip. More percussive things like gunfire can certainly clip, but quite frankly modestly clipped gunfire sounds just fine and there are post tools that can technically declip the audio. Now you may get the feeling I'm about to poo poo 32 bit audio and your mostly correct. What 32 bit audio does allow you tot do is have no idea what you are doing and not care enough to even check your modulation levels, as you can just hit record and as long as you are on board with the idea of the point of 32 bit audio and didn't crank the gain up, you'll be fine.
Now as far as this NUX company, I don't know them at all, know nothing about their product. It could be great stuff. My intuition is that it is crap and just happened to build their gear on a 32 bit platform. I could be wrong, and its not appropriate for me to make guesses about how good it is or isn't. My intuition is based on conceit and arrogance having the luxury of working with proven gear from Lectrosonics and Zaxcom on an average day. We are talking about a lavaliere mic transmitter / receiver, which are quite frankly crappy mics when compared to big expensive condensers (shotguns and studio mics). Rode does make a couple of good mics, NTG4+ is probably one of them.... haven't used one, but NTG3 was great other than durability. Zoom F6 is actually a great recorder too, they have a history of making hissy prosumer stuff, like H4/H6, but F6 is solid. As great as your stuff is, there's probably room for improvement, like maybe upgrading to a better mic or making sure you have good shockmounts, stands, holders, and extra cable for interviews to get the mic just right, wind protection and secret sauce accessories for handling finicky situations or wardrobe. These are the things that I'd be looking at first over picking gear which most of us have never heard of and selecting based exclusively on the 32 bit concept somehow presuming it sounds best because of it.