Professional Analog Mastering.
Professional Analog Mastering.
226K+ subscribers
You’ve seen an air filter before, I’m sure - made popular by that Maag EQ4 with the filter that extends to 40kHz.
When you’re not looking behind the scenes, this seems to make sense - you turn it on, boost the filter, and suddenly it sounds clearer.
However, what we’re hearing is a shelf filter that extends from the center frequency of the air filter down into the audible range.
For example, if I enable a 40 kHz shelf, then what sounds good or what I’m hearing is a mild shelf to roughly 12 kHz and above, depending on how greatly I amplified the shelf at 40kHz.
The same thing could be accomplished by creating a filter that boosts 12 kHz and above by the same amount.
If this were the only issue, I wouldn’t care - but boosting 40kHz will likely increase aliasing, noise, and introduce other unwanted artifacts like intermodulation distortion.
Say I place this before a saturator and boost frequencies near or above the Nyquist cutoff - I’ll absolutely increase aliasing distortion.
Or, say my recording environment includes a subtle high pitch whine from an AC unit, or a cheap dimmer switch, that’s above the cutoff and is reflected down the spectrum. It’s inaudible, but boosting it increases it enough until it's audible.
The point being, I can create the same beneficial effect of an air filter by using a lower frequency shelf filter, while avoiding any of the drawbacks that come from using the air filter.
Let’s compare a high shelf at 12kHz and above, amplified to the same extent as an Air EQ’s filter. Notice that they sound the same, but one won’t exacerbate aliasing.
Watch the video to learn more >
I used to work at a really well-equipped studio back in 2019 - they insisted on recording at 192 kHz, 32-bit floating point.
Sounded fine; however, once the session got above 40 or so tracks, the latency would interrupt any overdubs we were doing, since we had to increase the buffer size to keep it running.
Technically, recording at 192kHz should reduce latency, but once the computer is at capacity, so to speak, there’s nothing you can do to give either yourself or the performer an immediate signal to perform to.
When this limitation is coupled with the fact that there are no studies that support that a higher sampling rate has any perceivable effect, then it makes you ask, ‘what’s the point?’.
If you plan on saturating high frequencies in the box, then I can understand a higher sampling rate - but in the rare instances in which this is needed, oversampling may be a better option than running the entire session at a higher rate.
If that higher rate will impede your ability to produce, then oversampling the one or two tracks that need it is a better solution.
Furthermore, microphones, hardware units, mixing and recording consoles, and interfaces are all band-limited.
Meaning, they reproduce the signal up to 20 kHz before there’s a steep drop-off in amplitude.
For example, say I have this Apollo Twin. The conversion goes up to 192kHz, meaning it can reproduce frequencies up to 96kHz.
But, I’m using a Neumann U87; its response is 20Hz to 20kHz.
That mic is plugged into the Apollo’s unison mic preamp, with a frequency response of 20Hz to 20kHz; you see where I’m going with this.
No frequencies will ever go high enough to necessitate this high of a sampling rate - the one exception being if I record a vinyl record, which can go higher if the turntable used supports that.
So, despite the trend of using higher and higher sampling rates for the sake of increased fidelity, it doesn’t make much sense to record higher than 48kHz except in special cases.
Watch the video to learn more >
Now, this is, of course, subjective, but there are some technical reasons that analog sounds better.
It isn’t about the harmonics - we can replicate these with plugins. The same could be said for variable slew rates or small nuances.
The real reason it sounds better is a concept I covered in the last chapter - band limiting.
Imagine I’m recording an electric guitar and I use 2 different setups.
In the first, I record the guitar directly into the interface. To get a distorted sound or a sound that’s more indicative of a guitar amp, I use saturation plugins, or pedal emulations, and maybe an impulse response for cab modeling and reverb.
In the second, I used a real guitar amp and pedals - a mic is placed near the speaker, then run into the interface.
Ignoring the additional variables that come with recording the amp, like the air temperature, the relative level of the amp and mic, the room’s influence on the recording, and so on, why would one sound better than the other?
In the digital setup, the distortion I use will likely send harmonics above the max supported frequency, creating aliasing.
The convolution plugin I used for the cab modeling used FIR filters, meaning it creates pre-ringing distortion.
Any electro-magnetic noise from the guitar will be recorded, since the direct input supports frequencies up to 20 kHz.
These aren’t huge issues, but they add up.
Now let’s compare this to the guitar cab.
The distortion and guitar’s noise are capped at the guitar speaker’s frequency limitation, which is usually 6 to 7 kHz. This means less perceivable noise and no aliasing distortion.
The reverb and EQ introduce no pre-ringing distortion, meaning the transient response of the guitar is more indicative of the instrument - it’s not altered by digital filters and their limitations.
Now, apply this concept to any instrument you record. What can be considered a limitation of analog equipment helps minimize unwanted artifacts.
If my outboard gear can saturate, reverberate, equalize, or perform any other task without the addition of aliasing and/or pre-ringing, then the overall project will benefit greatly.
To show this, let’s listen to 2 guitar recordings. One performed with an amp and outboard gear, the other DI.
Notice that I can attempt to match the DI recording as closely as possible, but even in this isolated recording, there’s a noticeable difference.