Even high end gear will start to saturate at some point. There is a reason people like pushing Neve preamps, it's because they saturate in a pleasant manner. Other preamps, not so much. But it still distortion.
Having an average signal level of -18dbfs is not compromising the signal in any way. It really only gives 3 bits of space for the transients. Pushing the preamps into saturation to get an extra bit or two of information is a silly trade-off.
Sample rate does not affect quality beyond our ability to capture, reproduce and perceive the sound, since it only changes the possible frequency that can be recorded. If you are recording something that has no energy above 10khz, there will be no difference between a recorded sample rate of 44.1k or 96k...or 22.05k for that matter. Same goes for bit depth, which is one of the reasons 24 bit will be as far as converters will go.
The reason why Neve preamps are pushed hard is not due to their saturation, it's due to the lack of transient distortion compared to the distortions in cheaper pres at those signal levels, as well as taking advantage of the high signal levels themselves. The color/vitality comes from the components before it. It is simply put great amplification. Sound sources become vital at these levels when they are not distorting and when they mix with other sound sources they form stronger resonance. However, if you feed a non-harmonically distorted signal into a Neve preamp, there is no magic saturation in it that will harmonize that too. What comes in goes out, which is the job of a great amplifier - to stay clean at high amplification levels. But what a Neve preamp will do, beyond doing a really good amplification, is to tell the engineer when the microphone distorts. So tracking hot with a Neve is serving the engineer with the information about what the capture quality is and feeds the well captured and amplified signal to the next component in the signal chain. The combination of a great microphone setup, with a great Neve preamp, with a great monitoring setup, tracked hot, is a solution that is likely going to work.
When it comes to -18 dBFS it means +0 dBu on an +18 dBu converter and +6 dBu on a +24 dBu converter. The resolution of a sound source is directly proportional to the logarithmic Voltage RMS curvature of the dBu range, in other words you lose less and less sound quality as you reduce the signal level. I'm sure you know what only a +0.1 dB signal boost can do to your mix (in case you have a good audio interface, good monitors, good ears). This gives you a very good real world idea of how extremely precious the signal becomes at those signal levels and the potential production quality boost from getting clean transients at high signal levels and high visibility into that range.
When it comes to the aspects of sample rate, higher sample rate (when using high end converters) produces better results because it helps to combat some of the time dilation caused in the clocking. So even with the Nyquist–Shannon sampling theorem and the limitations of the ears, you have secondary reasons why you want as high sample rate as possible. The focus should be on getting as high level of clock accuracy as possible and then sample at as high sample rate as possible using that clocking technoloqy to get as little time dilation as possible and hence get as precise digital representation of the analog signal as possible (by also storing each sample with as high precision as possible). You do need a powerful DAW too as well as power stabilizers/conditioners, else you'll get the opposite effect. What happens is that the resolution simply increases, you'll end up with more of all the great qualities of a recording (given that all other aspects are efficient) - better low end, better mid range, better high end, better and wider stereo image, more 3D, more emotion/mix color, better sense of up/down and so on. When resolution increases it's like a flower that starts blooming, it starts blooming at the upper end of the dBu range and THAT is the way you should understand resolution, that's where the increase in resolution is the greatest. High end productions are like big beautiful flowers in bloom and that's beautiful.
In practice, these kinds of things are kind of fundamental to understand before you move to what truly takes care of that potential - the microphone setup. So the best thing to do is to just accept the potential of great amplification at high signal levels, then move on to focusing on microphone selection and microphone setups, using your high quality amplification and monitoring as guidance in this process.
There are many ways to production beauty, in rock music I surely know one: high signal level of a high quality drum kit sound in the mix without clipping. That's a very beautiful flower...
