And how come large format cameras seem to get more of it?
Sorry, but there's a three drink maximum in this bar, and it'll take me at least four Diet Cokes to explain all that. Unless I grossly simplify. And I mean grossly. Dare I? Oh, you know me, I'll take on most any challenge.
Let's consider two sensors, one very small sensor that can capture a maximum of 8000 electrons, and one that can capture 60,000 electrons (these are not arbitrary numbers—I'm kind of generalizing one small sensor I know about and one large one). Let's push things further and say that because the smaller sensor camera needs to be mass produced and made cheaply, we're going to throw an 8-bit Analog-to-Digital converter on it (8-bit ADC). On the bigger sensor, we have to charge an arm and a leg for the camera because of sensor costs anyway, so let's give it a highly accurate 16-bit ADC.
With me so far? Now let's record a middle gray. (Look away, here comes one of the simplifications.) The 8-bit sensor takes 4000 electrons and the camera's imaging ASIC places that at a value in the JPEG at 128 (actually, most cameras place it lower). It takes about 32 electrons to move to another value (127, 129). Now our big camera grabbed 30,000 electrons and put that same value at 32768. It takes about one electron to move to another value. (Yes, I'm going to ignore Poisson distribution of photons here; just look away when I simplify!)
So which camera is better at delivering information about low-lying detail? The big one, of course. Assuming that its ADC is accurate, it is distinguishing between very small differences in light that the smaller camera must just assign the same value. That's a tiny bit what micro contrast is all about. We've got data that we can move around and establish low-level tonal ramps from low-level edges.
Likewise, this answers the large-better-than-small question, too: we have better data. I'm sure somewhere on this site I've written the words "we don't take photographs, we collect optimal data." (Actually, Ansel Adams was all about collecting optimal data, too, but he did it in the analog world.)
Now, as I wrote many times, this is grossly simplified. I left out gamma, amongst a lot of other things. And lenses can impact micro contrast, because, after all, if the lens isn't passing the detail, the sensor sure as heck isn't recording it.