gswitz
He didn't talk about capturing the recordings and the fact that most of the data about fidelity depends on using every available bit. If you are only using 6 of your 24 bits b/c you have your levels set dangerously low, then you might get a recording like an analog cassette tape at 44.1. Now I've loved my analog cassette tapes but... just saying.

 
Once your quantization error is sufficiently below the other noise in your environment and you have plenty of headroom, you don't gain anything by adding more bits. And 24 bits already leaves lots of room for error here, so it is likely that the other noise (both acoustic and electronic) is a much bigger problem. The rule for 24 bit converters is "leave plenty of headroom and then worry about acoustic and analog circuit noise, not the converters".
 

I know you don't fix this by doubling the sample rate, btw. :-)

Actually you can trade bit depth for sample rate by using noise shaping. 

This is probably a good place to ask this question. I did my first-ever sampling last weekend, and ended up using Audition to do the task. I recorded them 44.1KHz mono (drums), but Audition defaulted to a 32-bit float on saves so I just used that. Is that 32-bit buying anything at all, or just wasting space on me?

bitflipper
Claude Shannon was one of the greatest minds of the 20th century and deserves credit for not only providing the proof that elevated Nyquist's concept to the status of "theorem" but also for proposing the idea of using binary numbers as the basis for electronic computing.

 
But Shannon gets an epic Wikipedia entry, so it all balances out.  
 
(Apparently, he created a motorized pogo stick and a flame throwing trumpet.)

"This is probably a good place to ask this question. I did my first-ever sampling last weekend, and ended up using Audition to do the task. I recorded them 44.1KHz mono (drums), but Audition defaulted to a 32-bit float on saves so I just used that. Is that 32-bit buying anything at all, or just wasting space on me?"
 
 
For recording yes its wasting space. Keep in mind that your converters are incapable of recording anything above 24 bits. So the file that was created is 24 bits plus a lot of padding. This adds nothing useful to the recording at all.  
 
Now for processing it a very different story. However, I am of the opinion that the file on disk doesn't need to be greater than 24 bits even after processing.   I am sure I am alone in this view.  

Thanks John. I assumed as much but wanted to ask.

I'm now happy that my stupid Tascam interface refuses anything beyond 44.1. All is good. 
But that said I always thought my DAT recorder sounded better at 48. 

Oh Lord.
Another vain attempt to throw mathematical logic at an argument where opinion is the controlling factor. I am waiting for the golden ears to chime in with how much crisper/sharper/fuller/more realistic sampling at 6x Nyquist frequency makes their music.
3...
2...
1...

Cactus Music
I'm now happy that my stupid Tascam interface refuses anything beyond 44.1. All is good. 
But that said I always thought my DAT recorder sounded better at 48. 

So did mine....

ampfixer
I guess the Cliff notes for that article would read something like "44.1 or 48k is just fine for most things".

Ah! That's what I was looking for.

Sigh... wotta buncha hooey. Facetious scientist indeed.
 
Surprise! ADCs actually sample at frequencies in the MHz even if they only output PCM streams at 44.1/48kHz. And DACs oversample 44.1kHz audio streams (in MHz) too! That's not ultrasonic, it's radio frequency (and relates to why the use of a CD player is prohibited at times on airliners). Not to worry though, decimation and lowpass (and often highpass) filtering fortunately keeps the out-of-band nasties from getting through (at least, its supposed to if things are designed and working properly).
 
Another surprise: the cheapo onboard 'high definition audio" codec chip inside the typical PC/Mac can handle 192kHz digital audio (such as one might find as a primary audio stream on a Bluray disc) just fine (by design).
 
Here ya go Noel, he's written all about bit-depth too:
 
http://www.sonicscoop.com/2013/08/29/why-almost-everything-you-thought-you-knew-about-bit-depth-is-probably-wrong/