Professional Mastering - Analog & Digital
Professional Mastering
251K+ subscribers
We often think of an EQ as separate from a clipper. Or a compressor as a different tool than a temporal effect like reverb.
But the processors we use to mix, master, produce, or perform sound design are all intertwined in ways that are rarely, if ever, discussed.
The best place to start as we look into these relationships is how EQ is related to everything and vice versa.
Watch the video to learn more >
Using an equalizer is a deliberate change to the frequency response of a signal.
But just because it’s the only deliberate method to change the response, it doesn’t mean it’s the only method available to you.
Far from it.
For example, if I reverberate a signal, I’m adding multiple decay taps or duplications of the original signal.
Say we have a reverb with a longer decay rate in the high mids, when compared to any other frequency range. A longer reverb decay often requires more decay taps. More decay taps in the high mids means there’s more amplification in the high mids than in other ranges.
Additionally, if the reverb is consistently triggered by the signal, then the additional taps in the high mids will overlap and add to one another, further amplifying the range.
So, in this example, even if our intention was just to reverberate a vocal or instrument, we’ve inadvertently affected the frequency response by increasing the amplitude of the high mids.
We’ll look at more examples of this in a moment, but first, I want to discuss a concept and a plugin that helps push this understanding further.
Watch the video to learn more >
Measuring the ADSR of a signal, or its attack, decay, sustain, and release, is a reliable method to interpret its content.
For example, if we look at a snare drum, the initial transient or the snap of the snare is the attack and decay. It’s the rapid increase and subsequent decrease in amplitude that occurs within a short period of time.
The tonal elements of the snare, the snare’s ring, its resonance, its tone, or fade, is a combination of the sustain and release.
For a snare, the sustain and release are lower in amplitude than the transient, but occupy a longer amount of time.
But it’s not just amplitude, and it’s not just the time that differ between the transient and tonal content of a signal.
The frequencies of each vary mildly to greatly depending on the instrument.
Using Split EQ, we can isolate the transient elements of an instrument or mix from the tonal elements. Again, the attack and decay, and the sustain and release, respectively.
By soloing each, we’re able to hear differences in the frequency response.
This means that any processor that affects the ADSR also affects the frequency response.
If the transient has a different response than the tonal, and a processor changes the amplitude of either the transient or the tonal, then, it has changed the frequency response, and effectively, although indirectly, equalized the signal.
Cleverly, Eventide utilized this concept with a plugin that takes indirect equalization of the signal via alteration to the ADSR, and makes that change purposeful.
So, amplifying transients that occupy the high mids is possible. Or amplifying tonal elements in the lows is possible.
By breaking the frequency response into 2 time zones, so to speak, you can now adjust the frequency response directly by altering the ADSR.
Let’s take a listen to the plugin on a mix to better understand the relationship between the frequency response and the ADSR.
Watch the video to learn more >
Processors that inadvertently equalize don’t stop at reverb or specialty processors like SplitEQ.
Compression, saturation, clipping, limiting, upward processing, gating, and any other form of dynamics processing you can think of indirectly equalizes.
Say I’m using a compressor on an acoustic guitar.
I use a quick attack, a quick release, some lookahead, and a moderate knee shape to quickly capture and attenuate the initial snap of the guitar string.
If you’re thinking, well, that would affect the ADSR and, in turn, the frequency response, then you’re absolutely correct.
Now, imagine I attenuate the peaks and amplify all of the quieter aspects of the signal post-compression. If the sustain and release are lower in amplitude, then I’ve just increased the amplitude of these tonal elements relative to the transient elements.
Again, if the transients and tonal elements, or even just the high amplitude and low amplitude aspects of the signal, have differing frequency responses, then I’ve inadvertently equalized the signal.
And quick side note, if reverb reflections, modulation like phasers, or any other effect are lower in amplitude, which, unless you’ve gated the effect, is always the case, then this post-compression amplification is changing the reverb or modulation.
This is why I always tell our clients to lower temporal effects before sending us their tracks for mastering. Amplifying quiet details through limiting or upward processing will almost always increase the amplitude and perception of these effects.
But compressors, and other dynamics processors are incredibly multifaceted - not only do they affect dynamics, they wave shape - meaning, they alter the actual oscillation of the waveform they’re processing.
This is especially true for low frequencies, with their longer oscillations.
If we observe a sine wave’s oscillation, then introduce a compressor, we can see this effect in real-time. Notice how big of a role the attack time plays. Other variables matter, but the attack time has a big effect in particular.
As I alter the attack time, notice what occurs to the waveform. With aggressive attenuation and while processing low frequencies, it’s easy to see dips and changes in the oscillation.
Watch the video to learn more >
Well, each change to a waveform’s shape isn’t attenuation, or removal of a part of the waveform; it’s the generation of a new waveform or waveforms.
These new and added waveforms are higher in frequency than the original, since they’re generated from a change that occupies less time than the original waveform.
Remember, the number of oscillations per second is the frequency. 50 oscillations per second is a 50Hz wave, and so on and so forth.
So, if I attenuate a 50Hz waveform, and it shapes the wave and generates a 150Hz sine wave, I’ve equalized twice with this one change.
50Hz is lower in amplitude after compression, and 150Hz is higher in amplitude after compression.
As you can imagine, as we process complex waveforms, the generation of new waveforms is equally as complex, with hundreds, if not thousands, of variables playing a role, resulting in complex changes to the frequency response.
Now, you can typically avoid the majority of wave shaping with a compressor by introducing lookahead, and if available, oversampling.
This lets the oscillation pass through the compressor before processing, and then retroactively attenuates it without cutting into the waveform.
But this isn’t the case with saturators or clippers. Most saturators gradually saturate, and just about all clippers cut the peak of the waveform, in turn squaring it.
Only some limiters use enough lookahead and oversampling to ensure wave shaping doesn’t occur.
All of this means that if you compress, saturate, clip, or limit, you’re almost certainly adding harmonics or overtones that alter the frequency response.
Now think about this practically. Say I have a clipper on the mixbus. The high amplitude kick and bass are triggering a fair deal of the clipping, since they often cross the threshold first.
Harmonics, in turn, form from the low frequency range and occupy the range directly above it. This means the low mids will be amplified the most.
I could give countless examples or scenarios for this, but let’s take a listen to asymmetrical saturation. It generates a second-order harmonic, which will amplify the low mids.
I’ll compensate the gain, and notice how changes to the frequency response become easy to hear.
Watch the video to learn more >
We’ve talked about how all processing affects the frequency response and, in turn, equalizes the signal. But what’s just as important is:
Now, this last point is a difficult one to wrap your head around, or at least it is for me.
In short, when you alter a frequency range with an EQ, you also alter the perception of all of the processing that occurs within that frequency range.
Going back to our reverb example from the first section of the video, in which we have additional reverb reflections between 2-5kHz, if I amplify that range with EQ, I’ve also amplified those reflections.
And if I attenuate that range, I attenuate those reflections.
This occurs with any form of processing. If a clipper adds harmonics, those occupy frequencies.
Adjusting those frequencies with EQ post clipping will affect the amplitude of those harmonics, along with everything else that occupies that range, in turn affecting the timbre.
The main point is that if each form of processing alters the frequency response, and its changes to the frequency response are a huge part of that processor’s timbre, then adjusting the frequency response with EQ retroactively affects that processing.
Not that the EQ changes the compression ratio or anything like that - but it changes our perception of the effect by altering what it’s added to the signal.