Bit depth in digital recording (using a computer or digital recorder rather than analog tape) refers to how many characters are used to represent the audio in digital language - you know, all that binary "ones-and-zeros" stuff. To get an idea what that even means, see my article: 16-Bit Audio Recording – What The Heck Does It Mean? where I tell you how "bit depth is like champagne":-).
Anyway, the Reaper Blog recently published a video showing you why you might just want to stick to 24-bit recording and rendering, rather than thinking you can benefit from using something called "32-bit floating point." You can see that video at the link below. But I wanted to say a few things about it before you check it out. The video not only explains what it promises in the title/headline, but in doing so, it shows you some other cool things as well. You'll see what a "clipped" audio recording sounds like. That's when you overloaded your interface and ended up with distorted audio. It looks squared off on the top. I always say that once you have distorted audio, you cannot "undistort" it just by turning it down. The damage has been done. You'll see that in this video. You'll also see a MIDI drum file in one of the tracks, along with the virtual instrument (Addictive Drums) that the MIDI is triggering.
So go and watch this short 6 minute video here. It's full of great information: http://reaperblog.net/2015/02/video-24-and-32-bit-recording-formats-explained/