Why Record Beyond 44.1kHz 16-bit?
Audio interfaces now offer recording rates of up to 192kHz – but why record beyond 44.1kHz?
The quality of recording equipment now available to the everyday home user is, frankly, staggering. The days of noisy, hiss-filled recordings to cassette are long gone. Nowadays, the typical home set-up comprises a computer and software capable of recording track counts that can run to over 100, with interfaces, such as those from Focusrite or Roland, that are typically capable of recording at rates of up to 96kHz or even 192kHz at 24-bit word lengths.
However, when the most popular music playback formats only run at ‘CD quality’, it raises the question, why record beyond 44.1kHz 16-bit?
There are some very good reasons…
More bits = more dynamic range
Switching from recording at 16-bit word lengths to 24-bit is one of the most instantly beneficial changes when upping your file-format specs. Effectively, moving to 24-bit increases the available dynamic range from approximately 93db to around 115db.
What does this mean? Any piece of recording equipment has certain amount of background noise (hiss) that is unavoidable. This is usually called the noise floor, and is the lower limit of what can be recorded- anything quieter than this will be masked by the hiss.
At the upper end, there is a peak limit whereby if the level exceeds it, the signal will clip in a very unpleasant manner. The difference between these two (effectively the useable range of the equipment) is the dynamic range.
The difference of 22db between 16-bit and 24-bit may not seem much, but as 10db is equivalent to a perceived doubling in volume, you can see that it is far more significant than it at first seems.
There is an argument that, it most home studios, when using microphones, it doesn’t make a lot of difference- your room is likely to be noisier than the noise-floor. True, but as there is very little CPU overhead for 24-bit files, or difference in file size, so you may as well take advantage of the additional headroom. Plus, if you record directly, with line level sources such as keyboards, or use DI boxes, the ambient room noise is removed.
Many DAW software packages offer 32-bit or 64-bit sound engines. These basically ‘pad out’ the dynamic range, so that when mixing, you effectively have more headroom. So, if you run into the red, it shouldn’t clip.
However, when mixing down to playable file (CD quality, for example), you’ll still need to mix with the limits of that file format. If all your meters are in the red, your resulting file will still clip, so pull those faders down…
Sounds that only bats can hear?
The range of human hearing is 20Hz-20kHz. In digital recording, the Nyquist theory states that a recording rate should be double the highest frequency recorded. Therefore, 44.1kHz should be perfect for the human hearing range.
So, what’s the point in higher rates? First of all, there’s a bit of debate regarding whether 20kHz really is the uppermost limit of human hearing. If you’ve ever used an EQ with a range that runs to 21kHz or above, you’ll know that on some sources, boosting at 21kHz can have some profound effects on some, bright sound sources. Are you hearing frequencies at 21kHz? It’s difficult to say.
Recording at 44.1kHz ensures that you won’t hear anything above 20kHz, however, by employing a low-pass filter to remove any frequencies above. This is to prevent aliasing issues.
The real benefits of higher recording rates relate to the sheer detail and precision they are capable of. Digital audio works by taking ‘samples’ of the audio many thousands of times per second, then playing back these chunks of audio in sequence. These occur so quickly, that it sounds like a continuous sound (not lots of little pieces).
Consider a waveform like the one below.
When sampled, it might look like this.
As you can see, a certain amount of precision is already lost. Dropping the sample rate, it might look like this.
The accuracy decreases again.
You can see that the basic ‘shape’ of the wave is still apparent, however. (When the sample rate exceeds half a wave cycle, though, then the accuracy starts to really drift – the aliasing problems outlined by Nyquist theory).
This inaccuracy means that higher frequencies will never have as much detail at 44.1kHz as they would at higher rates.
Still, when the track is ultimately to be mastered at CD quality, what difference does it make?
Well, more than you might think. One reason is that when mixing on a computer DAW, plug-ins can introduce delays, which in turn can cause phase distortion, cancelling and boosting other frequencies. At higher rates, these delays are far, far shorter, and have far less detrimental effects as a result.
One of the best explanations I’ve ever heard takes a slightly view on the extended frequency range. Rather then arguing that we can hear above 20kHz, it states that the increased sampling rate means that we hear the audio in the frequency range that we are most attuned to (1-5kHz) in far greater definition and detail.
With modern dithering algorithms, this detail captured at source can be preserved more accurately when converting to lower sample rates than it could if the recording was made initially at 44.1kHz. It’s also worth noting, that many engineers recommend using 88.2kHz rather that 96kHz, due to the simpler maths involved when converting to 44.1kHz (simply halving).
So, do I abandon 44.1kHz?
You might be thinking, ‘well, my recordings sound fine at lower rates- what’s the problem?’ As Joe Meek (I think…) once said, ‘If it sounds right, it is right’.
If 44.1kHz delivers results that you’re happy with, without your computer running out of steam, that’s great.
It is worth experimenting with higher rates, however. Subjectively, they have a far smoother, open and more natural sound. And, as just about any engineer will tell you, you should aim to keep the quality as high as you can, for as long as you can throughout the recording process- once the quality is gone, there’s no bringing it back…
View a complete range of recording equipment over at the Dawsons website.