Professional Analog Mastering.
Professional Analog Mastering.
224K+ subscribers
Recently, I saw someone arguing that parallel compression is more or less just regular compression. And in short, no, it isn’t. Or as we’ll see later, it shouldn’t be.
When parallel compression is introduced correctly, it should have an upward compression effect in which the combination of the dry and processed signals amplifies low-level details of the signal. Then, louder signals should align with unity or return to a 1:1 ratio.
If we use FabFilter’s Compressor and observe its dynamics with Plugin Doctor, we’ll notice that the representation of the dynamics in the plugin matches that of Plugin Doctor.
With the default settings and auto-make up gain turned off, we can observe a 4:1 ratio, with a somewhat soft-knee.
I’ll adjust this to limiting settings, with a hard knee, and the lowest possible threshold. Again, Plugin Doctor is showing the same result.
But watch what happens when I introduce dry signal using the Dry Signal dial.
Plugin Doctor’s graphic no longer matches FabFilter’s - that’s because the introduction of the dry signal in equal proportion to the processed or wet signal creates what we’d expect from parallel compression.
That is, upward processing or the amplification of quieter details. Let’s listen and notice how this drum track’s quieter details become easier to hear.
Watch the video to learn more >
With this relationship, we can vary the parameters to adjust the upward processing. For example, if I change the hard knee setting to a soft-knee setting, we’ll notice that the amplification of quieter details becomes gradual, and slightly lower in amplitude.
A slower attack also amplifies or increases the upward processing, as does a shorter release.
The same could be said about the ratio, as when less compression occurs, the included dry signal contributes more to the amplification of quieter details.Furthermore, the type of compression, or maybe better worded, the behavior of the compressor, alters this relationship as well.
A classic compressor setting will cause a differing relationship between the 2 signals, as will optical or clean.
To show this, let’s set the compressor back to default - again, we’ll notice that Plugin Doctor’s dynamic graphic matches FabFilter’s.
This time, I’m going to introduce parallel compression, but I’ll just use the mix slider instead since it’s easier.
However, there seems to be an issue. I’m no longer getting amplification of the quieter details. It looks like I’m only getting less compression.
Well, let’s try auto makeup gain and see if that fixes it. Kind of, but not really. Quieter details are being amplified, but not in the same way. Additionally, the louder aspects are still being compressed whereas before they weren’t.
What if I compensated for the change by manually adjusting the output? Again, I’m not achieveing the same charactristic as before.
Okay, well maybe it’s just this plugin - maybe they reserved parallel compression for the dry and wet signals, and not the mix slider.
I’ll load Izotope’s Dynamics plugin. And great, this one’s slider even says parallel; I’m sure it introduces the effect as expected.
Except it doesn’t - like Fabfilter’s mix slider, this plugin’s parallel function seems to introduce less compression. Testing gain match, and using idential threshold and ratio settings didn’t make a difference either. Either I was getting less compression, or in the case of gain match, the compressed signal was being amplified to unity after processing has occured.
The same thing happens with Artiuria’s 1176 Emulation. Same with Universal Audio’s compressors, and any other compressor with a Wet/Dry dial that I could find.
The only compressor I could find that behaved in the way I expected was the Weiss DS1-MK3. With its parallel compressor setting enabled, a limiting ratio, and a threshold of -60dB, we’ll notice very similar behavior to FabFilter’s parallel compression settings.
Here, let’s compare Weiss’s parallel compressor to Izotope’s parallel slider function so you can hear the difference.
Watch the video to learn more >
Why parallel compression, and parallel compression with a wet/dry function vary so greatly, is easier to visualize than describe.
When we combine our wet and dry signals with a wet/dry function, we’re introducing level compensation to the dry signal while the wet signal is being amplified or vice versa, depending on whether we’re going from dry to wet or wet to dry.
This compensation causes the final result or output to sound much more similar to traditional compression than parallel compression.
When we introduce parallel compression without this level compensation, the loudest aspects remain the same level or a similar level depending on your settings, while the quieter aspects of both the dry and wet signals combine to amplify the range.
It’s this level compensation that has been throwing us off, and causing us not to create parallel compression, but instead create level compensated compression.
Going back to our FabFilter plugin, I can create a similar effect. I’ll introduce parallel compression with the wet and dry dials, but this time, lower the overall output until the quieter levels return to unity.
This is essentially the effect we achieve with a mix or wet/dry slider. It’s as if we create the intended effect and then nullify it with gain compensation.
If we look in an actual session, we’ll notice that when FabFilter’s compression is introduced, the original level before parallel compression and the level after parallel compression match. So if I feed -12dB into the compressor and introduce parallel compression, the output will be -12dB.
The overall amplitude and perceived loudness will remain the same, but the perceived level of detail will increase. Meanwhile, if I create the same settings for Izotope’s dynamic plugin, and use the wet/dry slider, the amplitude of the signal decreases to -18dB, while the percieved loudness also decreases.
Now, even though we could fix this in a lot of compressors by adjusting the band’s gain after compression, it’s a step most of us wouldn’t think to introduce.
Additionally, actual parallel compression allows us to hear the before and after immediately with the same perceived loudness and at the same amplitude. With wet/dry parallel compression, we need to adjust the output gain until we achieve the same level and loudness.
The point is, the intended sound of parallel compression, which is upward compression, is lost through gain compensation whenever we use wet/dry sliders or dials to introduce the effect.
As we discussed, you can use the output dial in some compressors to cause the intended effect, even if it adds some guesswork.
Additionally, I would be wary of trying to accomplish it with a compressor that doesn’t allow you to control the threshold, the ratio, or other important effects that let you allocate the amplification to lower amplitude aspects of the signal.
For example, if I use PSP’s impressor, you’ll notice I can achieve a similar effect with the wet/dry dial and level compensation via the output dial. But since I can’t control certain parameters, it’s difficult to isolate the amplification to quieter parts of the signal.
Instead, I personally still like the old way. I’ll use a pre-fader send that’s set to unity and introduce whichever compressor I want while using an aggressive ratio and threshold. Then, I’ll blend the wet signal in until I can hear those quieter details. If I need more of the effect, I can increase the wet gain, the auxiliary fader, the plugin’s output, or a combination.
Although not identical to the method shown with FabFilter, at the very least, I know the original signal is not being attenuated to compensate for the level of the processed signal. I know that when I blend it in, I’m only hearing the addition of the parallel compression.
And, if you’re still set on using your compressor’s wet/dry or mix function, just be sure to adjust the output so that you can more accurately hear the upward processing that’s associated with parallel compression.