Professional Analog Mastering.
Professional Analog Mastering.
226K+ subscribers
I see this advice come back every once in a while - in short, the idea is if you put a high pass filter on the sub frequencies you improve headroom.
If an equalizer worked perfectly, that would be the case.
But the area around the cutoff will be amplified due to the phase rotation needed to create the filter. This boost either negates any benefit of the filter, or it actually makes the situation worse by amplifying more than what was attenuated.
Additionally, I’ve seen it argued that this filter decreases measured loudness, which causes Spotify or streaming services to read the track’s LUFS as lower, in turn causing a track to sound louder when normalized.
None of that is true. The algorithm that streaming services use high pass all signal before measurement to avoid this type of manipulation.
So in short, introducing this filter either does nothing, or it does the opposite of what you’re trying to achieve.
I get particularly annoyed by this one since I included it in a video a long time ago as something to maybe do very subtly, and then I saw it turned into some of the worst mixing advice on Youtube.
In short, the idea is you put Soothe 2 on the instrumental, side chain the vocal, then attenuate the beat or instrumental whenever the vocal is present so that the vocal sits on top.
If it’s super subtle with gradual curves, that’s fine. A bit lazy, but fine.
If someone is advising you use it on every song, and do it aggressively, that’s rough.
Soothe 2 is just attenuating what’s present in the signal. It cannot differentiate between what’s musical content and what isn’t.
So if you use it on a complex signal, and you use narrow filters, you’re just turning down the highest amplitude parts of that signal, likely the overtones which are define the timbre of any recording.
It’s soothing the signal, by literally turning everything that’s musical down. If you want to do this, a free alternative is lowering the instrumental’s level with the channel fader.
This way I’m not just turning down musical aspects, and leaving unmusical aspects alone, since they’re lower in amplitude and don’t trigger the attenuation. A channel fader attenuated everything uniformly.
If you use soothe 2 with sharp filters, you’re adding audible artifacts, messing up the transient response by affecting the ADSR, and introducing unmusical distortion - all the while turning down parts that don’t need to be turned down.
Every video about creating a wide mix or master has a section about expanding high frequencies.
There’s nothing wrong with it, but I always see the same Izotope Stereo plugin, with settings that are way too high, especially in the highest range.
What’s worse is when the delay function is engaged, which uses phase interference between the left and right to further widen the signal.
Correlation in the highs isn’t seen as important - especially when compared to correlation in the lows, but there is an incredibly important reason not to greatly expand highs.
In short, when the wave file is converted into a lossy formant, like an AAC or MP3, the highs are affected the most.
This is where the algorithm deletes information to save on space, since it’s assumed it won’t be too noticeable.
However, if the left and right channels have differing info when this deletion takes place, that’s when we get those noticeable artifacts.
That phasing sound is almost entirely created by info being deleted on the right but not the left, or vice versa.
So, if you expand the stereo width of the highs, which is the same as increasing differing info in the left and right channels, then you’re ensuring your master will sound phasing in the highs once the track is converted for streaming.
If you want wider highs, try a subtle high frequency shelf on the side image. You’ll get a wider sound without creating the conditions for lossy artifacts.
There are entire video series about why equalizers, compressors, and so on, are scams. It makes for an interesting thumbnail, but these videos are more or less gross simplifications.
The curves created from analog emulated EQs can be recreated with fully parametric EQs - that is if you took the time to figure out the exact curve, and then replicate it with the EQ.
That doesn’t make them a scam, it just means you could spend a lot of time trying to create every nuance with a plugin designed to be super flexible. Or, you could use the analog emulation to expedite that process.
A video I saw about compressors being a scam is even worse - the argument was that compressors don’t do anything unique, they attenuate the signal given specific parameters and that’s it.
In short, that’s a complete lie.
Compressors have completely different behaviors that alter the ADSR uniquely, distort the signal uniquely, even equalize on occasion.
Some compressors have dual release stages. Some have attack times so quick they can literally change the frequency by wave shaping.
Some chop the signal up into on/off cycles like a transistor and turn off part of the signal, which is the case with a Pulse Wave Modulator or PWM compressor.
Now, I’ll admit I don’t know everything about compressors - I’m not smart enough to know every distinct nuance of every make and model. However, I’m smart enough to know when I don’t know something, or am incapable of fully knowing it - which is unfortunately, something that’s often lacking on YouTube.
So if you see one of these XYZ is a scam videos, watch it for entertainment, but please take it with a big grain of salt.
It’s true that lookahead will cause a compressor or limiter to distort the signal less, but this doesn’t mean it’ll sound better.
From a technical perspective, yes, using lookahead will reduce wave shaping which reduces distortion, but that’s not the full picture.
When a compressor wave shapes a peak, we can emphasize it.
Distortion will amplify the signal with higher-order harmonics, while the lower frequency peak is attenuated.
Sometimes the attenuation is greater than the distortion, sometimes the distortion is greater than the attenuation, but the point is, there’s a complex relationship at play.
When lookahead is used, the full peak is attenuated, without distortion or with very little distortion.
So you achieve clean compression without distortion, but that doesn’t mean it’s perceived as cleaner, or more detailed. Peaks are attenuated more accurately, but the result is a muted, tonal sound.Low Pass Filters and Aliasing
Aliasing occurs when harmonics form above the highest supported frequency, causing them to be reflected back down the frequency spectrum in equal and opposite measure.
The idea is, if you use a LP before a saturator, or after it, the harmonics are attenuated.
However, the harmonics occur entirely within the saturation plugin so to speak.
While being processed, any aliasing that would occur occurs. If say 15kHz is present, and a harmonic is generated at 30kHz, then the reflection or aliasing will occur before the signal is routed to the next processor.
This means a LP after the fact won’t change the aliasing that’s already occurred.
Now, if you used a LP before the saturator and attenuated a good amount of the highs, this would likely reduce aliasing by removing frequencies that could cause aliasing when saturated, but that’s pretty counter productive unless you don’t want the highs in the signal at all.
So in short, stick to a plugin’s internal oversampling function if you’re concerned about aliasing.