The Sonnox products are top-notch, and I wanted to know exactly what the Limiter and Inflator are doing, so I checked them out with some tools.
Last Updated: 24-September-2024
In my mastering, I heavily lean on Sonnox Oxford Inflator and Sonnox Oxford Limiter as they are pivotal in my master chain. Despite my past struggles in understanding them, I am determined to master these two crucial plugins. Internet searches have not been very helpful, as I see that others have similarly grappled with these plugins.
The first step I took was to read the manuals for each of these plugins. I also read the descriptions and help sections on the Sonnox site to get an understanding directly from the plugins’ makers.
Sonnox Oxford Limiter
General Limiting Settings
Below are shown my limiter settings. I put the limiter last in my master chain with nothing after it.
Enhance and Safe Mode
I do not normally use the enhance slider or safe mode because my music is usually coming in with plenty of loudness already when it gets to the limiter.
Threshold
I strive for a true peak level less than 0, so I set my threshold at -0.10 dB, which limits the peaks so they do not exceed 0 dB. If this limiter is set to something below 0 dB, it will not allow peaks over this setting. If, however, it is set at 0 dB, it will allow peaks over. This is a subtle difference from other limiters and one you cannot overlook if you are interested in keeping your peaks below 0 dB.
Attack
I opt for a longer attack time of 0.476 ms to achieve a more musical limiting. What is meant by this is that peaks coming in are not immediately clamped off or gain-reduced, in other words, if they have rise times less than 0.476 ms. That does not mean that peaks are escaping and being allowed through. This is because the limiter has internal headroom, so peaks are preserved. Peaks are not clipped off in this limiter.
This is an important point about this limiter. Peaks are not clipped—imagine that. They are not flattened off. They are gain-reduced, which means they retain much of their original structure as peaks, but they are lowered a bit to comply with the limiter’s threshold value that you set (in my case, -0.10 dB).
In my case of allowing peaks through that are less than 0.476 ms, these fast transients are preserved and not gain-reduced. A slower attack time like this allows the fast transients to go through and not be limited, but I get these with the dithering that squares them off and sounds more like a solid-state circuit that is in some modern hardware limiters.
Slower/longer attack times allow finer detail to pass through the somewhat harsher sound that fast limiting can produce, and the longer-term events will have a more rounded peak.
Release
I choose a release of 1.06 ms to start with in my mastering. I rarely need to change this.
Short (faster) release times give a less pumping sound to my music and cause it to sound a bit louder as a result of basically “resetting” the process of gain reduction. With a longer release time, I can sometimes hear the reduction in gain as a quieting of the music going through it, followed by the subsequent increase in level as the limiter completes its release cycle.
Faster release times result in the highest perceived loudness. This is because the gain recovery occurs quickly after peak events, affecting average program levels only for short periods.
Fast release times can lead to distortion as gain recovery may begin to occur between the waveform peaks of lower frequencies (bass sounds) in the program. If distortion of the bass is noticeable and not pleasing, the answer would be to increase the release time.
Longer release times are more forgiving of gain changes and allow for greater overall compression, but they result in a quiLonger release times provide greater overall compression but result in a quieter output, making them more forgiving of gain changes.eter sounding output.
- Fast release =
- less compression
- punchier/livelier sound (less gain reduction)
- a more even-sounding output
- higher loudness
- distortion in bass (low frequencies)
- Slow release =
- more compression
- more gain reduction
- lower loudness
- less distortion in bass
- more pumping sound
There is a tradeoff between fast/short and long release times, but fortunately, that is exactly what the auto gain button does.
Auto Gain
I have the auto gain button engaged for all of my mastering and do not disengage it. I find that it creates a less gain-reduced (pumping or ducking) sound over the length of my tracks.
To audition the effect, I played my music through with and without the button engaged and listened. I find it is always a better listening experience with the auto gain button engaged.
The reason it results in a better sound is complicated, and this single button is really deceptive as to just how important it is to have it engaged.
The auto gain function compensates for large-level changes while still allowing fast peak modeling for shorter peak events. It does this by modifying the attack and release internally (inside the limiter) based on the incoming sound.
As mentioned before, fast release times can lead to distortion in the low (bass) frequencies as gain recovery may begin to occur between the waveform peaks of these lower frequencies. The speed of release and the generation of distortion in music involves a trade-off. Adjusting the release timing allows for tuning these effects, and that’s exactly what the auto gain button does.
If the auto gain function is not selected, moderate release time settings (above around 0.2 seconds) may produce audible gain “pumping” due to longer and more noticeable recovery periods. You can get around that by increasing the input gain and using the soft knee function to compress more gradually, making the transition in and out of compression gentler and less obvious.
- Auto gain in =
- compensation for wide variations in loudness of the music
- less pumping sound
- Auto gain out =
- no compensation for wide varying loudness
- need for adjusting the input gain and soft knee to reduce pumping sound
- potential pumping sound if not adjusted elsewhere
- music with a lot of fast peaks and slow peaks (tempo changes) can have pumping
This auto gain button is effectively a magic button for me that stays engaged always as it makes my music sound better.
Soft Knee
The soft knee slider is interesting. It engages the gain reduction sooner so you get a more gradual limiting the more you slide it upward (increase). What I find is that I can control my peaks and my loudness a bit with it. For example, if I am not getting good peaks, I can reduce it to zero and then the limiter is acting as a hard knee limiter with peaks going right up to the threshold and being folded back by the limiting. This allows a bit of expansion of a sort because I can increase my PLR by having the peaks hit a higher point.
With the soft knee engaged and sliding it upward, I can limit earlier and get a softer limiting function. Peaks are not so prevalent, but if they are there like with a drum it, it shapes them so while there are fewer hitting the setting on the threshold, there are some, so this is a shaped increase in PLR. Because the peaks are reduced earlier due to the soft knee, the overall loudness decreases. So I can use this to control my short-term loudness as well as my overall loudness.
It has really become a handy tool for me when I am trying to dial in my loudness numbers and really shape my music into something better.
Harmonics
Using Plugin Doctor, I checked the harmonics imparted by Limiter:
1. With no Enhance in (slider at lowest point), a small amount of harmonics are introduced – see the image below. It is down around -120 dB, so it isn’t very important, in my opinion.
2. With Enhance in, at any level above zero (in this case, 5%) then, there are odd third and fifth harmonics around -35dB (yellow and orange lines), and even harmonics are at a lower level around -65dB (red, green, and brown lines). This kind of explains to me why people say this limiter has a good low-end — they must be using it with the enhanced feature active. I ran it higher in the enhanced level and also in safe mode, and these levels just increased.
Sonnox Oxford Inflator
Harmonics
1. With the effect out, there is not much difference from that of the Sonnox Oxford Limiter enhance function off.
2. With the effect in and at 100% and the curve at 17.0, you can see the odd harmonics in the -15dB to -20dB range (yellow and green lines), while the even harmonics have a different signature showing some peaks in the midrange and tapering upward at either end (green, orange, and brown lines).
Curve
Looking only at the extremes, it is easier to describe what the curve slider does.
- +50 – in any range from 0 to +50, the loudness increases, volume goes up, fatness increases, and the dynamics drop a little bit – although they still sound dynamic.
- 0 – at 0, loudness is increased, dynamics are not lowered, presence increases mostly in the low end, and highs and peaks are dropped a bit.
- -50 – in any range from 0 to -50, the harmonics are still there, but there is not as much increase in loudness; highs are a bit attenuated, making them softer-sounding, and loud parts will stick out more from the mix.
Combining Sonnox Oxford Inflator with Limiter — What I Do
I like to have the inflator before the master EQ and then the limiter. Sometimes I put other effects between the Inflator and the EQ (tape, sampler, etc.).
Since the inflator is giving a lot of harmonics and doing the best at sweetening the sound for me, I try to never use the enhance feature of the limiter as it only adds, what I feel, are conflicting harmonics, making my sound a bit muddier.
But if I do need to use the enhance feature, I try it in crossfade (X fade) mode first, then later in safe mode if needed and starting at 0.