Some good reading in this thread. I'll give you my take for what it's worth as far as bit and sample rates are concerned, float points etc.
I'm of the belief of "if I can't hear a difference, there is no difference and there is no reason to tax my system unless something makes me go "wow". I don't believe in floating points of any kind...32 bit float, 64 bit mix engine.....when someone can prove they can really hear a difference, I might buy into it.
The way to tell for sure is to record a project the exact same way using what you think is so much better verses what is not. Then, instead of you trying to do the comparisson yourself, let someone else run the media player and play the songs. This makes it way different than cuing up the stuff yourself to where it can be mind over matter.
I've only encountered higher sampling rates making a difference 2 times in my life. Once was a project I recorded for a client in 24/88. He was a stickler for that setting, so I went with it. For some odd reason, this project had a much different sound than the one I decided to do using 24/48 as a test project for myself. Everything just sounded thicker on the 24/88 than it did on the 24/48 (especially my guitars) and I made sure I had someone else press the buttons on the media player when I wasn't looking. I picked out the 24/88 every time in under 5 seconds. I played the exact same things using the same instruments as the 24/88. This actually bothered me because I had never used 24/88 before and haven't since then. This thread has reminded me to revisit that to see if I can still tell a difference.
Just recently I did 2 test projects of a 2 minute clip. One at 24/48, the other at 24/96. I hear no difference at all. The other time I HAVE heard a difference was an orchestra I recorded live one time. I was recording them at 24/48 and they decided they wanted to try a song they were just starting to work with just to see how it would sound. The told me they'd probably not get through the whole thing so I decided to track this song at 24/96. They made it through half the song, screwed up and wanted one more shot. I switched back to 24/48 and recorded them.
When I got the material back to my studio, the 24/96 sounded better to me. The instruments were so much more lively and ambient. It is my belief that natural sounding instrumentation recorded at higher sample rates will sound better as opposed to more sonic instruments like electric distorted guitars and electric basses and anything "direct recorded". Anything electric in my opinion will not benefit from the larger sample rate. But it seems to do wonders on acoustic instruments. Try it and see for yourself. Record an acoustic guitar or something using a stereo mic rig at a lower sample rate, then pump up to 24/96 and re-record it. The capture just sounds more like the instrument is in your room and right in your face.
But I've had my best results using 24/48 for most of my stuff over here and it is what the majority of my clients have always sent to me as well when I've had to mix something or master something. I've been working with clients in the audio field for over 12 years now. The most common bit and sample rates I deal with are 24/48 and it's been that way even before I started working with clients on a business level. All the studio's I've ever dealt with recorded in 24/48 also...so I chose that for myself as well.
As for the whole floating point thing etc, I simply cannot hear it and if soundcards can only do 24 bit, wouldn't that be hype to you guys? If I can't hear something that makes an incredible difference, I could care less what the math sheets tell me about what is really going on. That's just what I believe. If I don't gain performance from stuff like this and I can't hear a blatantly obvious sound change or alteration, what's the point? I'm on a site that does these intense shoot-outs with top end mic pre's verses lower level pre's that anyone can afford. I can't tell you how many of those tests we listened to where only the n00bs were claiming they could hear a difference. The seasoned pro's said in an instant "there may be a little difference there, but not one worth $5000."
Then you get the dudes that just think their poo don't stink that just need to say they can tell a difference and are stuck in their ways. This is when you post up 3 of the same file and lie that there were 3 different pre's used...and watch them fall into a world of BS. If you guys can truly hear this stuff, God bless you. I've been told I have some of the best ears around for picking up on stuff and let me tell you, I nearly have animal hearing and can't tell on this stuff. I can walk into a room with my eyes closed and tell you whether or a not a TV is on or off even with the sound muted. I can hear the high frequency of the tubes IF that TV has tubes to the point where I can't stay in the room with it like that because it bothers me horribly.
Try some tests on this stuff and see if you really can hear the 64 bit mix engine, 32 bit float, 16/44 vs 24/48 etc. Do a few test projects playing the exact same stuff using the exact same gear and patches as perfectly as possible and then let someone else cue them up without you looking. I am willing to bet it all makes less of a difference than you believe it does...and if that is the case at the end of your tests, what's the point? The hype? The math it does? If I can't feel it and I can't hear it to where it's blatantly obvious, it simply isn't real to me. I had a respected friend tell me floating points and all that stuff reduce errors or something. What errors? The only errors are if I screw up on a take or get some clipping or an artifact that sneaks through. At the end of the day, someone that knows what they are doing will get good sound out of anything. I can drop recordings I've done on my pre-pro gear using a stock Dell and a Realtek soundcard running at 16/44 with ASIO4ALL drivers that will blow away stuff from pro studio's. I'm sure you guys that know what you are doing can do that too. If something sounds good to me, it is good. If I can't hear something, it's hype. Just my opinion though. :)