Professional Analog Mastering.
Professional Analog Mastering.
226K+ subscribers
Compression is really where you achieve a vocal’s distinctive timbre, maybe even more so than saturation, and here’s why.
We’re all familiar with attack and release - how long before compression begins and how long compression lasts once the threshold is crossed.
But there’s a lot more happening than what this simplified definition would have you believe.
It’s easier to show with some graphics - say the horizontal white line is the threshold, and the vertical white line is our attack time - let’s say 20ms.
If this vocal is above the threshold, it’s not as if our compressor simply waits 20ms, and then causes instant maximum attenuation to occur.
Instead, imagine that once the signal crosses the threshold, there’s a timer in place set to 20ms, and by 20ms the compressor reaches the maximum amount that it’ll attenuate.
So the attack time is not telling the compressor to wait - as soon as the vocal crosses the threshold it’s being attenuated, but only after 20ms will the maximum amount of attenuation occur.
Something very similar occurs for the release - let’s say it’s 250ms. Again, it’s not as if compression occurs for 250ms but stops when the signal falls below the threshold.
Falling below the threshold initiates the release, and it takes 250ms before the signal goes back to it’s uncompressed state. This means that even after falling below the threshold, the signal is still being compressed.
This is even more interesting when we consider how different compressors have different behaviors.
In our 20ms attack example, maybe the max amount of attenuation is always 20ms, but how it gets to that max attenuation can vary.
If the attenuation over the 20ms was linear, and say 10dB is the max amount of attentuation, then at 10ms, the compressor will have attenuated the signal by 5dB.
But, if the attenuation over time is not linear, which is, I believe, always the case, then at 10ms, we can have 3dB of attenuation, or maybe 9dB of attenuation. The same goes for the release - the 250ms release might be linear, but it definitely doesn’t have to be and will vary from compressor type to compressor type.
Just to demonstrate this, here’s plugin doctor showing the effect—it’s feeding a sine wave into the compressor, then amplifying it before returning it to its original amplitude. Notice how varying the attack time makes it easier to observe the gradual amplitude adjustment we discussed.
Also, notice how even after the signal falls below the threshold, attenuation occurs—which can also be affected by varying the release time. Lastly, notice how the attack and release change behavior when we alter the compressor emulation.
And this is really just the beginning of how a compressor alters the timbre - we have harmonic and intermodulation distortion, the general reshaping of the vocal’s ADSR, the reduction of masking to quieter details via make up gain, and a lot more.
Watch the video to learn more >
The current trend for vocal production is to place them right in the front of the mix while maintaining detail and nuance.
Compression with these 5 elements will achieve this:
Soft-knee4:1 RatioA Quick Attack with 1-4ms of Lookahead5-8dB of AttenuationMakeup Gain
I can guarantee you’ll achieve an upfront sound if you utilize these settings while compressing vocals.
First, the soft knee combined with a 4:1 ratio causes more compression to lower amplitude signals but keeps the compression sounding natural and relatively transparent.
Lookahead with a quick attack causes the attenuation to occur right at the beginning of the vocal, by letting the compressor measure the signal before it begins to process.
5-8dB of attenuation is all you need to control even aggressive peaks, but when combined with the previously mentioned settings, it still avoids a noticeably compressed sound.
Lastly, make-up gain, ideally automatic makeup gain, brings all of the quieter details forward. Since we smoothly compressed, and the compressed signal includes even the initial transients of the vocal, post-compression amplification will bring a great deal of quieter details to the forefront.
Let’s listen and notice how the vocal becomes detailed and complex and sits right in front of the mix.
Watch the video to learn more >
This is a common question, but it’s a false dichotomy - in other words, we don’t need to pick one or the other when both are an option.
Use EQ before compression to attenuate any problematic frequencies or anything you don’t want amplified.
This may be plosives, aggressive sibilance, nasally tones, unpleasant resonances, unneeded low mids, and so on - basically, use this EQ to create a balanced and safe sound.
Then compress and use settings from the last chapter to amplify what’s still there - which should now be everything about the vocal you like and want to keep.
With these elements amplified by the compressor, use a second EQ to amplify everything you want more of. You could boost some of 2-5kHz to increase clarity and intelligibility. You could amplify some air frequencies with EQ or an exciter.
And if you place this EQ later on in the chain, maybe after reverb, delay, and whatnot, you can even amplify aspects of the side image, like the warm reflections around 300Hz.
So, don’t limit yourself to the so-called “right” option when you have the freedom to explore different possibilities.
Let’s listen with the compressor settings from before, but now with EQ before and after compression.
Watch the video to learn more >
Let’s say you’ve compressed the vocals, but even then, they still sound weak, far back, or kind of lifeless in the context of the mix.
First, off, don’t compress more - that’s a mistake I used to make. More compression, even serial compression, will likely change the vocal’s timbre in unpleasant ways.
Instead of controlling the dynamics from the top down, let’s use waveshaping to control them from the bottom up.
Waveshaping is commonly used for synths, but if used conservatively, you can think of it as a complementary tool for compressing vocals or other instruments.
Melda Audio makes a great free waveshaper - with it, I’ll amplify quieter details, similar to how I did earlier with the compressor’s makeup gain, but this time, I don’t need to attenuate peaks.
If I only want the quietest parts to be amplified, I can double-click to create another point, drag it to the desired ampitude range, and then drag it up to amplify that range. Meanwhile, I’ll ensure that I have a linear or 1:1 input to output relationship for everything above the affected amplitude range.
This way details are brought up, while the peaks stay the same level, causing upfront vocals without messing up its timbre.
Let’s listen to 2 waveshapers - this Melda audio one and the popular Oxford Inflator. Melda Audio’s will give you more flexibility, but the Inflator has a great sound nonetheless.