Among other things, the Loudness Curve module’s settings can be customised to adjust the Duration (length of time displayed on the horizontal axis) and to show a target Reference Integrated loudness level.
Lower a mix’s integrated loudness with Squasher.
The ‘LUFS’ measurements that are used by streaming services to normalise loudness on playback, and which are now catered for by so many meters and limiters, aren’t just weighted (with an EQ curve that accounts for the sensitivity of human hearing to different frequencies) but they’re also ‘gated’: any portion of a signal where the momentary loudness falls below a defined level (‑10LU below the integrated loudness) won’t count towards the final integrated loudness.
When undertaking DIY mastering duties, the presence of this ‘floating gate’ raises two significant questions. First, for any given track, are there significant portions of time where the momentary loudness falls below the gate’s threshold? Second, if the answer is yes, might there an opportunity to raise some of those quieter sections just above the level of the gate, both to add energy or clarity, or to lower the integrated loudness — for example, to reduce the amount a track might be turned down — without negatively affecting the louder sections?
When I recently (SOS July 2024) reviewed iZotope’s RX 11, I was particularly impressed by its Loudness Optimize feature, which was conceived to address precisely those questions. As a Cubase user, that got me wondering... could I achieve something similar using only the tools that come bundled with Cubase?
Analyse This
For Cubase Pro users (v11 and above), the Loudness Curve module of the SuperVision analyser plug‑in can help with the first question. Inserted at a suitable position in your project, this provides a real‑time visual display of your audio’s loudness. The Loudness Curve module settings (first screenshot) allow you to customise the display. You can, therefore, set the time resolution of the display to suit your needs (I used a 45‑second window here) and choose which loudness parameters are shown (in this example, just the blue momentary and green integrated loudness curves are displayed).
SuperVision’s Loudness Curve module can provide very useful visual feedback on the loudness and dynamics of your audio.
As shown in the left‑hand panel of the second set of screens, these curves update in real time as the track is played. After an initial playback, the integrated curve will stabilise and give a reasonable sense of the overall loudness. You can then dip back into the Settings panel and set the Reference Integrated value to 10LU below the observed integrated loudness value. In this example, the integrated loudness is about ‑18LUFS, so I set the Reference Integrated value (it appears as a horizontal pale blue line) to ‑28LUFS. This isn’t what this parameter is really intended for (I’ll come back to that later) but it does provide you with a simple visual guide to spot sections of the track where the dynamic range may be great enough for the momentary loudness to fall more than 10LU below the integrated loudness curve.
Incidentally, for any Artist and Elements users, who won’t have SuperVision, APU Software’s Loudness Meter offers a free‑to‑download alternative to SuperVision. It provides some pretty sophisticated features that include a real‑time momentary loudness curve similar to that found in SuperVision’s Loudness Curve module.
The Squasher plug‑in, available in Cubase Pro, Artist and Elements, can apply upward compression.
Going Up
This visual feedback is a useful check to know if the gate might influence how streaming loudness normalisation is applied to a track, but it’s still your decision to do something about it or not. If you do, the focus needs to be on raising the level of the quieter sections of the track while leaving the louder parts unaffected, and there are a number of approaches that might be considered. One, for example, would be parallel compression. But the Squasher plug‑in, available in Cubase Pro, Artist and Elements, can apply upward compression, and this is super‑easy to experiment with.
The final screenshot shows some Squasher settings that might serve as a good starting point. For simplicity, the plug‑in has been set to operate in single‑band mode. Within that band, the Down knob has been set to 0% and the downward compression Threshold (the small downward arrowhead) has been set to 0dB. These settings should ensure that no conventional (downward) compression is applied, which would otherwise affect the louder sections; these should pass through the plug‑in with their dynamics intact.
Squasher is capable of upward compression, which allows you to manipulate the loudness dynamics of your track in some interesting ways.
In contrast, the Up knob, which controls the ratio of any upwards compression, has been set to 100%. With this set, you can then gradually raise the Up Threshold value (the small upward facing arrowhead) until you see the upward compression kick in, as indicated in real time by the pale bar at the base of the Threshold graphic.
At this stage, Squasher’s main Input and Output controls have been left at zero (no overall gain adjustment applied; I’ll come back to the Output control in a moment). In addition, the Drive control (in the Parameters panel) has been left at zero so no saturation is added, while Gate is set at its lowest available value (‑60dB), effectively bypassing Squasher’s internal gate. In this case, the Attack and Release are both on fast settings but, as with conventional downward compression, you can experiment with adjusting these based on the tempo of the track, letting your ears judge what produces the most transparent results.
Spot The Difference
From this starting point, you can adjust the two main Mix controls, the Up Ratio and the Up Threshold values to vary just how much (or how little) upward compression is applied. Again, your ears are the key here, both in terms of just how much louder the quieter sections of your track become but also in terms of how transparently any loudness change can be achieved.
However, SuperVision can also be a guide: place a second instance of the plug‑in immediately after Squasher and, if you use the same display settings as in the pre‑Squasher instance, you should easily be able to see the difference in the loudness curves pre and post upward compression.
The right‑hand side of the first screen shows an example using exactly this configuration. While the momentary loudness curves are the same in the louder parts of the audio, Squasher’s upward compression has bought up the loudness during the quieter sections. While I’ve perhaps overcooked the processing here for illustrative purposes (and could lower some combination of the Mix, Up Threshold and Up Ratio controls for a more subtle effect), Squasher’s potential to manipulate the dynamics of your audio, increasing the loudness of quieter sections while leaving louder sections untouched, is easy to see.
Target Practice?
I mentioned above that I’d return to the Loudness Module’s Reference Integrated setting and Squasher’s Output control. Almost all streaming platforms have now published the integrated loudness reference levels used by their normalisation algorithms. They’re not all the same, but they are similar and, as an example, Spotify uses ‑14LUFS‑I as its default — on playback, songs will be turned either up (as much as peak levels allow; there’s no limiting applied in the default mode), or down to ensure they’re played back at ‑14LUFS‑I.
This is an automatic process, so it’s not critical that the integrated loudness of your mastered track is ‑14LUFS — the whole point is that loudness normalisation plays it back at that level for you. But perhaps you’re working for a client who has specified delivery of files at a certain LUFS‑I. Whatever the reason, if you feel the urge to nudge your track towards a target level, you can.
First, use the Reference Integrated setting for its intended purpose, and set it to the required target value (‑14LUFS for Spotify, for example). Then you can simply use the visual reference this provides and adjust Squasher’s Output knob up or down until your track’s integrated loudness curve sits on or around the Reference Integrated target line. For the example used here, it would require a setting of about +4dB. Providing this doesn’t also raise any peaks close to clipping, on playback your final mastered track might give Spotify’s loudness normalisation algorithm very little to do.
And Finally...
There are a few caveats to keep in mind. First, it’s not all about the numbers: in raising the loudness of quieter sections of your audio, you’ll be changing the track’s overall dynamics. It’s always worth making loudness‑matched comparisons, so that you can judge whether the processed version sounds better or worse than the original!
...in raising the loudness of quieter sections of your audio, you’ll be changing the track’s overall dynamics. It’s always worth making loudness‑matched comparisons, so that you can judge whether the processed version sounds better or worse than the original!
Second, you’ll probably also be raising the level of low‑level noise in the source material. This may not be an issue if the original recordings were of decent quality, but do make the effort to listen out for unwanted side‑effects. If noise becomes an issue, you might be able to counter that by de‑noising your sound sources, whether with a spectral noise removal tool or strategic use of noise gates or expanders.
Finally, not all mixes and mastered tracks can benefit from this sort of loudness dynamics adjustment. If your initial SuperVision visuals don’t show the momentary loudness dropping 10LU below the integrated loudness curve for significant portions of time, further loudness manipulation might well be unnecessary or unfruitful. If it’s not broke, don’t fix it!