Are you comparing your uncompressed soundtrack to others that are compressed transparently for broadcast? That's a big difference all by itself: loud and soft sounds are close to the same volume in a good broadcast (within reason), relying on the change in tone that we've come to associate with volume, rather than actual volume, to convey that sense of volume.
It could also be that Windows still has its default audio processing on. It does now, and has for a little while, so that non-technical business and casual home users can call people without understanding acoustics at all, and have it "just work".
As part of exploring all of the settings anyway, as you should for a new rig, figuring out what all of them do, and setting them to work for you instead of the other way around, you need to find all of that audio junk that Windows does by default, and turn it off.
I have put the mic in cardioid mode to primarily focus on recording my own voice.
Good! Keep it there! Good audio starts with noise rejection, and cardioid (compared to omni or figure-8) has a nice null in back that you can put the worst offender in. Use that!
Or if it's inconvenient to point the mic directly away from something, you might use a different pickup pattern that makes it more convenient. Figure-8, for example, has its null all the way around the sides. You might put that between two speakers so that both speakers are in that null. And pad the wall behind, so that the equally-sensitive rear lobe doesn't get a reflection of your voice.
Super-cardioid, compared to just cardioid, tightens up the forward lobe so that the sensitivity falls off faster on the sides, in exchange for some rear sensitivity coming back. The null in this case is a cone at some angle behind. Super-cardioid is also called "shotgun", if you happen to see that term among techies.
---
In the vast majority of mics, there are only two native pickup patterns - omni and figure-8 - and the rest are formed by mixing different amounts of those two together. Equal amounts of each produce a cardioid pattern, etc.
The switchable ones probably do have multiple mic elements, and the switch determines how much of each goes into that mix. The non-switchable ones often have a single mic element, with a labyrinth port on the back to delay the sound that gets to the back of the diaphragm. The effect is the same - omni plus figure-8 - but it's achieved in a different way that involves a physical engineering term called "superposition", that says that you can analyze different sub-parts of the problem separately and just add their results together.
At any rate, if you block the rear port, it suddenly becomes omni, which might give you a clue as to how that superposition works. I've heard of sound guys dumping on rappers because they like to hold the mic in a way that blocks the rear port, and then the monitor speakers *that they're still pointed away from*, suddenly start feedback-screeching.
