When making a loud master be sure not to push not into a limiter aggressively - instead, boost a little of 3kHz and dip 250Hz to emphasize important frequencies. Also, when masking a loud master that’s not squashed, pick saturation that amplifies high frequencies, but leaves lows unchanged.
The primary mistake engineers make when trying to make a loud master is relying too heavily on the limiting stage - it goes without saying and you probably know this if you’re watching this video, but the more you push a signal into a limiter, the quieter it’s actually going to sound in the long run.
This is especially true if the master gets normalized. So let’s take a listen to a mix being ruined this way, and then we’ll move on to some better methods for achieving loudness.
If we look at a graph of where our ears are most sensitive, we’ll notice that 3-5kHz sounds the loudest compared to other frequencies - we can subtly boost this area on our master to make it sound louder. This next graph shows how much 250Hz. masks other frequencies.
We’ll notice it has the greatest effect on this same 3-5kHz. range - this means we can subtly attenuate 250Hz. to reduce its masking effect, in turn creating a louder master.
Let's add both of these subtle filters, and notice how our master becomes a little louder.
Saturation is going to cause harmonics, which will emphasize certain frequency ranges - the trick when trying to make a loud master is finding a saturator that emphasizes the mids and highs, but has a subtle effect on the lows. It’s easier when you can emphasize saturation on higher frequency ranges, like with this PSP saturator.
Just be sure to use oversampling and use its smoothing function to avoid aliasing.
Let’s listen to how saturating the frequencies we’re most sensitive to causes the signal to sound louder.
Maximization, upward compression, and inflating all work by capturing quieter details of the signal and amplifying them - this is what we’re actually trying to do with aggressive limiting, but with these types of processors, we don’t need to sacrifice our peaks. Granted if used too aggressively, we’ll still lose dynamics.
Regardless, this is a great alternative to downward compression when mastering, especially when making a louder master. Let’s take a listen.
Like with limiting, there are a lot of mistakes we can make - for a loud master, don’t use lookahead and avoid a long release time. Additionally, and it goes without saying, don’t heavily compress the signal - instead, consider using subtle mid-side compression to cause dynamic stereo expansion.
Aside from that singular use, I find compression doesn’t help too much when mastering unless it’s corrective like with de-essing.
Let’s listen to how even subtle compression can negatively impact loudness if we use an extended-release and lookahead.
Both linear phase processing and oversampling will affect the loudness of your master in the long run - unless used for parallel processing, linear phase should typically be avoided due to how it’ll affect transients. Oversampling on the other hand will reduce the effect of aliasing distortion.
Aliasing distortion causes phase cancelation to our high frequencies, which subtly reduces loudness by attenuating important high-frequency ranges, and by lowering the overall level through destructive interference.
Let’s take a listen to high latency linear phase processing since its effect is more pronounced, and notice how higher latency settings negatively impact the dynamics and high-frequency transients.
Since YouTube, Spotify, and other services normalize audio to roughly -14 LUFS, many engineers make the mistake of mastering to -14 LUFS - something I’ve been guilty of myself. In truth, if you master to a little above this loudness, your track will still sound louder than -14 LUFS tracks, even after normalization.
Granted, if you destroy your dynamics as we covered in chapter 1, it’s not going to sound louder, but if you use upward processing and thoughtful limiting to get to -10 or -9, the quieter parts you push to the forefront will make a difference.
To show this, I’m going to play a track mastered to -14 LUFS, and then one mastered to -10 or -9 LUFS, but normalized to -14 LUFS, and we should be able to hear a difference.
If I’m trying to make a loud master, I often separate my signal into left and right channels on which I can insert limiters with identical settings. This way, each limiter is only causing attenuation when the individual channel is crossing the threshold - this is similar to delinking channels.
With this setup, I find I’m able to push the signal to a much higher level and still avoid aggressive attenuation. This is especially true with this Oxford limiter in particular, where I can use the enhance function to amplify quieter details in addition to peak-down limiting.
Let’s take a listen.
Aside from using limiters too aggressively when trying to make a loud master, another big mistake is using lookahead - this will cause the limiter to capture peaks early, in turn attenuating transients and reducing important dynamic high frequencies. Just as importantly, avoid true-peak limiting.
This also reduces the impact of transients, causing the master to sound quieter. Instead, monitor true peaks and use oversampling, as well as a slightly lowered output to avoid them.
Let’s listen to a limiter being used with true peak limiting enabled, and notice how it negatively impacts our transients.
Last up, let’s look at how we can use clipping to increase a master’s loudness - in short, clipping causes a form of high-frequency noise to be added whenever the threshold is crossed. This acts a little like an exciter in that it’s adding high-frequency content to transients.
This makes the master sound brighter overall, and also, somewhat counter-intuitively helps us retain transients since the added noise makes them easier to identify.
Let’s listen to how clipping creates a louder master, and pay attention to how the transients still sound bright and present when they cross the clipper’s threshold.
