When trying to achieve advanced compression during the mastering stage, try using upward and downward compression, splitting your stereo signal into separate left and right channels, or automating various functions of your compression. Emulating analog equipment is also helpful, and can be done with separate compression and distortion plugins.
One of my favorite ways to control dynamics is from both the peaks down and the noise floor up. In other words, first I’ll use low-level compression to increase the level of the quieter aspects of the signal.
This lessens the dynamic range, but without affecting the peaks of the signal. Then I’ll insert a compressor, with which I’ll attenuate the signal downward or from the peaks down.
With these 2 processors, I go back and forth until I find just the right balance between the 2. I find that more upward compression keeps me from using more aggressive downward compression.
As a result, I preserve a lot of my transients while still achieving an impressive, full, and dynamically controlled sound.
More times than not, when we process a master, we affect the left and right channel collectively, as a stereo signal. However, if we duplicate our master, pan it left and right respectively - we can now place our processing on each channel and master the left and right separately.
When it comes to compression, this is going to give us a lot more control - granted it makes the most sense to use identical settings, but slightly different settings can give the master a very unique sound.
Even if we do process the left and right with identical compression settings, it’s less of a burden for each processor, resulting in a slightly cleaner sound.
This same concept can be applied to other forms of processing so try split-channel mastering with other effects. Then if you want to collectively process the master, insert the effect on the stereo output.
Compression naturally occurs in analog equipment when the input and output aren’t a 1:1 ratio - so when the components in the hardware can’t replicate the level of the input, resulting in lower output than the input.
Good analog emulation plugins take this into account and let you compress the signal by driving the input. The signal will naturally compress in a gradual, soft-knee fashion.
If the analog emulation plugin you’re using isn’t doing this but distorts instead, you can simulate it by still increasing the input and reducing the output.
This is going to sound a little different from plugin to a plugin so try multiple to find one that sounds best for the signal you’re working on.
Usually, when we set our release time or other compressor settings, these parameters remain the same throughout the entirety of the master; however, automation opens up a lot of possibilities.
One great way to use automation is to alter the release time of your compressor to better match different sections of the song.
The first step is to take 60000 and divide it by your BPM to get a quarter note in milliseconds. The reason this works is that there are 60000ms in 1 minute.
Let’s use this quarter note, and then automate the release time to a 16th note. To find this we’ll divide our Quarter note by 4.
Under the automation tab we’ll find the compressor, the release time, and then automate that time wherever we want the compressor to have a different timbre.
The easiest way to create saturation is to use a saturation plugin - but you can create your own by combining soft-knee compression and distortion.
First, let’s compress the signal using soft-knee compression. What’s great is we can pick just about any compressor we want - be it classic, FET, Opto, or whatever sounds good.
Then we can add our distortion. If you’re curious which distortion to pick, run a sine wave through it and measure the harmonics.
I typically like to see a second and third-order harmonic - this will add some fullness with the 2nd order and a little body with the 3rd order.
Another benefit is that with typical saturation it would be difficult to balance the amount of compression with the amount of distortion, but again since our processors are separate, we can pick the balance.
