Actually due to the way recording works, recording at 24 vs 16 (and keeping it that way throughout the whole production process) is a massive difference. After all is said and done and the music has been mixed and mastered then dithered to 16bit though, yeah, 24 and 16 is pretty hard to tell apart. :P We're talking about professional studios, not people dicking around in FL with soundfonts. The dynamic range of something as unassuming as a wind or string instrument (or even voice) can go >30dB just due to many factors, and you want to catch all of it with as low a noise floor/quantization error as possible. 16bit is 96dB from floor to clip, and while it's possible and even mandatory at times in the past to squeeze the dynamics into an adequate recording (tape had far less headroom) there's so much else that can go wrong in the studio that you just don't want to worry about it. 24bit allows you to gainstage very very comfortably where you know you aren't going to clip yet are going to keep the noise floor really low. On a semi-related note there's quite a bit of consternation at recording at 48kHz when you don't have to (I.e. your target media isn't film or DVD) since resampling from 48 to 44.1 introduces more problems than the original extra clarity is worth, but a lot of people record at 96kHz then resample to 44.1 OR 48, which f- things up a lot less.