What Sample Rate Should You Record At?
(Why HIGHER Can Be WORSE!)
I’ve long maintained that the ever-increasing sample-rates offered by recording/playback equipment result in diminishing returns and that the limits of human hearing mean, as per Nyquist, that anything above 48kHz is simply increasing the file-size for no discernible improvement — that, in fact, whilst ‘CD quality’ audio could perhaps be somewhat improved with a higher bit-rate, fundamentally, 16bit/44.1kHz was as good as was likely ever going to be necessary.
Typically, I was shouting into the hurricane for the most part — I mean … what do you say when proffered a £1,500 pair of headphones along with a £2,000 mp3 player, with the recommendation that you listen and (in a peculiarity of the English language) see precisely where the money went?
So, rather than argue the point myself, yet again, I’ll let someone else do it … whom people might take a little more seriously (because they use pretty pictures and animations and Siri/Alexa/whatever on YouTube rather than simply say “There’s barely any appreciable advantage to more than 44.1kHz … and no advantage to anything more than 48kHz … because human beings can’t hear anything above 20kHz.” *sigh*)
Of course, as observed, studios these days anticipate 24bit/48KHz, so I’m pretty sure that (in the not too distant future), they’ll anticipate 24bit (if not even higher by then ¹ )/96KHz … because that’s what their customers will mostly deliver (if not 192KHz) *sigh*
¹ 28bit? 32bit? 34bit? 36bit? Who knows? (I mean why 24bit rather than the obvious doubling from 16bit to 32bit anyway? ² )
² There is actually a good reason but, potentially at least, a 32bit system could be advantageous to someone like me (who uses lots of insert effects rather than shared FX busses ³ ), because cascaded 24-bit DSP operations can create audible noise after a while and many audio sources summed together can mean a noise signature that covers pretty much the entire bandwidth of the material — so, 32btt/48KHz would be fine as far as I’m concerned, but that’s not the point (it’d be reassuring to know there were the extra headroom to accommodate my utterly ludicrous demands, but not necessary ⁴ ).
³ I like to shape each sound uniquely rather than swamp everything with a common set of effects and hope that it all mixes together nicely — I want to know that each sound fits into the sonic space perfectly, not just approximately, and, in the track I’m working on currently, for instance, I have grabbed a single phoneme from a vocal, doubled it up and applied an ‘arpeggiated’ gate to one of the copies, followed by no fewer than five delay plugins and some automated panning and volume changes to get the precise effect I want.
⁴ I make electronic music ‘in the box’, I don’t record philharmonic orchestras — it’s nice to know there’s the extra headroom for when I record something from outside the box but … as everything is either a balanced signal to begin with, or else gets cleaned up via a DI box before it reaches any recording device … once again, it’s entirely unnecessary.