How do you like to track: bit depth, clock, etc.? Sweetspot?

In the past few years I'm usually tracking at 24/96.  I do hear a difference vs. 24/44.1 (or 48) when I've got more tracks and more plugins (and maybe it's aliasing or whatever- don't really care - I like it at 96).  When I'm just tracking drums for a client I default to 24/44.1 and for most that's what they seem to want (sometimes 24/48) and if they want 24/96 then it's been a slight upcharge (because it does take 2x the disk space, more upload time, etc. - although to be honest that was more significant 10 years ago).  I thought about changing that, but almost everyone has seemed fine with 24/44.1 so haven't changed it yet.

Like I said, I use 24/96 these days... but my default used to be 24/44.1 (and yes, CDs were still a thing) and when I get projects from people requesting drum tracks, a lot of them still are using 24/44.1 and some send me 16/44.1 WAVs (I still send them 24/44.1) - I think this is the fault of Logic, or maybe even Garageband, which as I understand is not intuitive about setting your sample rates? At least that's what I was told.  And I think some of it is projects that people start (at home) on consumer interfaces and they maybe don't know what they are doing, or they started it 10 years ago, and yes, it defaulted to 24/44.1.  It doesn't really matter to me - if that's what my client asks for (or if they don't ask) then that's what they get.  But here's the thing, I've done enough projects in both 44.1 and 48 to know that there is no perceivable difference (and it's been argued to death on the purple site and more).  I can definitely hear the difference at 96 (when it's more tracks - tbh I can't usually hear it on a single track comparison).   

This is probably for a separate thread (if there is not one already) but when I deliver masters to a client (meaning I did the final product and not just drum tracks), I deliver the hi-res Wav (usually @96), CD quality (16/44.1) WAV, 320KBps MP3 (some streaming sites and internet radio use this), and if requested, and any other formats they might need for whatever the final product is.

24/88.2k. I just don't hit the converters hard enough to peak. I suppose it's around 6db down or so.

Mix down to 24/88.2k as my HD master file.

No. Good plugins are optimized for all sample rates. They either oversample or have other distortion prevention mechanisms like advanced, windowed average attack smoothing filters in dynamics, decramped eq filters, unique rms smoothing, smoothed zero crossings in rectifiers, or can only apply distortion to the low frequencies to not alias, etc.

take Oxford eq and Oxford dynamics. They are old plugins but well designed. The smoothing low pass to the dynamics that smooths the rate of attack release switching changed based on sample rate to not alias heavily even if the lookahead time that smooths the attack stays the same. So it can compress higher frequencies the higher the sample rate it without merely ramping them down beforehand. Renaissance eq and compressor are similar but I don’t know if they change based on sample rate.

Also modern processors like Weiss and TDR will over sample the audio path entirely different than the side chain to ensure low distortion and lower cpu usage.

Even old distortion plugs like Vintage Warmer and Decapitator have ways to minimize high frequency distortion at single sample rates.

And your converter chips are all running at a high sample rate in the mhz now and are linear to their dithered noise floor, clock, and analog electronics theoretically. They just filter it to the desired sample rate. The digital part is solved but cheap ones still screw up or do dumbass stuff like dsd or do not use dither, which breaks the sampling theorem. A lot of the depth increase into gear without hysteresis is just ignorant users not dithering, dulling everything, making the sample values slightly wrong. Dc filters in clean transformer less gear will make the peaks slightly higher and more forward from phase shift just running through it when level matched. Depth and width increases break down when you consider component mismatches in l/r and unwanted hysteresis murking up the signal even with jensens.

I asked you this on another thread, but if you ran higher sample rates would that mean that you could do less oversampling on plugins? For example... would I still want to run TDR on Insane mode if I was working with 96khz files? I think you're going to say "yes" but I'm curious if it would at least reduce the need for that type of thing even if it doesn't completely eliminate it.

To me the sonic benefit is lost if my DAW gets bogged down on a mix and I'm already getting bogged down at the very end in some cases. This is due to plugins where I have jacked up the oversampling such as Soothe 2, TDR, and LiquidSonics.