Digital Audio – Sample Rate and Bit Depth

Digital is all the same, right?  If you are recording, mixing and editing audio in the digital domain for a CD you are going to work at 44.1kHz, 16 Bit, right?

NO.  Of course when you copy digital audio you don’t lose any quality, or what we call generation loss in analog.  But here’s the catch.  Digital is just ones and zeros.  Think of it like a record.  A record has a groove with hills and valleys that the needle rides over and the change follows the sound curve, creating sound.  Digital is the same, except it just has on and off.  So a sine curve in digital cannot be as smooth as it is in analog, unless.

Unless you chop that curve in so many little pieces that the on and offs start to create a smooth line from one to the other.  Therefore a sampling rate of 48 is smoother than 44.1, and 96 is even smoother, because it takes twice as many samples of the same period of time than at 48.  More samples = smoother curve = a better representation of natural sound.

Here’s another point.  In computer language a word, or sample, always has “junk” at the end of the chunk.  It never fills the entire available space.  So if you record at 16Bit depth you are not using all 16 Bits.  That’s why if you record at 24Bits you have more space for information, again a smoother curve.

This is why it is so critical to record at the highest rate possible, and to stay at that highest rate as long as possible.  When you sample down to CD (or even to mp3) your curve is smoother than if you started at 44.1.  Also, when you down sample from 24Bits to 16 the first thing that gets chopped off is the “junk” at the end of the file.  That could leave you is 16 full, complete Bits, which provides a cleaner sound.  The down-sample removes the excess at the end of the file, so your end product sounds more realistic.

Why do you think the “audiophile” recordings you can by have 96kHz 24Bit (or higher) plastered all over it?  I’ve read over the years that many professionals feel that a sample rate of 96kHz brings digital pretty close to the curve that humans can hear.

Another reason for higher sample rates: they can reproduce frequencies that lower samples can’t.  The human ear may not “hear” these frequencies, but the human can “feel” them.  They round out the experience that you would experience in a live performance.

All of this comes together in the recording, mixing and editing.  After the recording, working at a higher rate provides the same higher quality. Mixing is smoother, editing is finer, and any effects you add are more realistic at higher rates.

I know what you are thinking.  Why should I take more processing power and storage space when “everyones” listening to mp3s and internet streaming on YouTube?  Here’s a test.  Take a CD, and listen to it in a pair of earbuds (but through a CD Player, not an mp3 player).  Now listen to the same CD through a good pair of headphones.  Hear the difference?  Same source, different listening styles.  Same thing.  If you record a band at mp3, and I record the same group at PCM wav 96kHz/24 Bit – and I even then down-sample to the same rate you are using, mine will sound better.