Sample rate and bit depth are the two numbers every digital audio conversation eventually reaches, and both of them are routinely discussed in a way that does not match what they do. Raising your session to 96kHz does not make your mix better. It changes what is possible at certain stages, and it changes the noise floor of your converters. Those are different claims, and conflating them is how people end up doing expensive work for no audible gain.
What sample rate actually controls
Sample rate determines how many times per second the waveform is measured. It sets the highest frequency that can be represented, which is half the sample rate in theory and a good deal lower in practice. A 44.1kHz session cannot describe frequencies above roughly 20kHz, which is fine for almost everything you make.
The practical reason to go higher is oversampling. When a plugin applies a steep filter, a limiter, or any process that generates energy above the current ceiling, the result can alias into audible frequencies below the ceiling. Oversampling gives the filter room to work before the signal is sampled back down. This is not about capturing more air from the microphone. It is about giving processing headroom.
What bit depth controls
Bit depth determines how much dynamic range you have before quantisation noise becomes a problem. It is not a quality setting you choose in the moment. It is fixed by your converter, and 24-bit is standard for a reason: it puts quantisation noise so far below anything you can hear that it is irrelevant.
Going to 32-bit float buys you headroom in processing, not final quality. It means you can run gain-staging that would clip a fixed-point format without wrapping. It does not make the finished master any better, because the master gets dithered back to the delivery format on the way out.
Where a retuning plugin sits in the chain
A pitch-shifting plugin is a nonlinear, computational process. It analyses the signal and reconstructs it, which means it generates content that was not in the original. Placing it early on individual sources gives you the cleanest result, because there is less accumulated material for the algorithm to untangle. Placing it on a master bus is faster and more convenient but the algorithm has a much denser signal to work with.
If you use a plugin such as 432Hz MASTER per track, that is the higher-quality route. If you use RASTA 528Hz on the master, you will hear the result more clearly, but you are asking a single pass to do the work of many.
What changes and what does not
| Setting | What it genuinely affects | What it does not affect |
|---|---|---|
| 44.1kHz / 24-bit | Everything required for standard delivery | Nothing you are missing |
| 88.2kHz / 96kHz | Oversampling headroom for filters and limiters | Recorded air, source quality |
| 32-bit float | Processing headroom during gain staging | Final master quality |
| Dither on export | Correct bit depth for the delivery format | Anything audible in the session |
| Low-pass filtering | Removing real high-frequency content | Hiding monitoring problems |
Monitoring is the part that matters
The most common reason people cannot hear what their plugins are doing is simple. They are not hearing the plugin at all. If your monitoring is too loud relative to the level at which you are judging, or if you are monitoring through speakers far below the frequency range where the action is, you are making decisions on incomplete information.
After a retune, listen to the same passage twice: once with the plugin bypassed and once with it engaged, at a level where you are not leaning forward. If the change is inaudible at a comfortable level, you are not in a position to judge it properly.
The bottom line
Sample rate and bit depth are solved problems in most workflows. The differences that matter are the ones you can actually hear, and those come from where plugins sit in the chain and how carefully you listen rather than from the numbers on your session settings.
Frequently asked questions
Does a higher sample rate improve audio quality?
Not on its own. It gives filters and limiters more headroom before the signal is sampled back down, which reduces aliasing during processing. The source recording is not improved, and a mix at 44.1kHz is not inferior simply because of the rate.
Is 32-bit float worth using?
For processing headroom, yes. It prevents clipping and allows gain staging that would wrap a fixed-point format. It does not improve the finished master, which is dithered to the delivery bit depth on export.
Should a pitch-shifting plugin go on every track or the master?
Per track gives the cleanest result because the algorithm has less dense material to reconstruct. On the master it is quicker and easier to hear, but one pass is doing the work of many. Use a plugin such as 432Hz MASTER either way, but know which trade-off you are making.
Do I need dither when exporting?
Yes, if you are reducing bit depth. Dither converts the truncation error into a low-level noise floor shaped across the frequency spectrum, which is far less audible than the distortion that hard truncation produces.