Hello welcome back to another episode of
Are You Listening? My name is Jonathan Wyner and today I'm going to talk to you about limiters. So the limiter is one of the most essential and often used pieces of kit--processors that you'll find in the mastering engineers toolkit Back in the day the limiter was created to prevent signals from going into distortion in the context of radio broadcasts, typically that was an analog compressor, with a very very high ratio. With a high ratio it takes so much energy to get through the limiter or the compressor to get just a little bit more gain coming out,
that it is effectively a ceiling. But in analog it's not really entirely a brick wall in the way that it is in
the world of plug-ins and digital. So when you're using a limiter as a plug-in, I think it's better to think about the limiter as a safety an absolute way of keeping the maximum
level at a point where it will never exceed zero or whatever you set the
threshold in the limiter to and it will keep you away from getting into distortion. We're not usually thinking in that sense about the sound of the limiter but we're thinking about the limiters ability to do something to the
sound to get us to where we want to go When I'm mastering a record whether it's
a single or a collection of songs, One of the things I'm thinking about is what
does the resulting level need to be? We think a lot about level and mastering
you've probably all heard about the loudness wars, maybe you've heard about
the idea of different levels for a peak normalized format like a CD or WAV files
that get played directly from a disc as opposed to loudness normalized material. In any case, the limiter is the thing that allows us to specify the target level. In order to be clever or smart or hopefully adequate about how I set up a limiter I'll put my file, whatever it is, into my session, before I do anything
else I'll start to play with the limiter and get the signal close to my target
level so if my target level for instance is something like minus 14 dBFS
loudness normalized, I will set my limiter. Take a look at what the
limiters behavior is going to be to get me to that target level and then observe
what's going on if I've got too much limiting taking place then I know that I
probably have to do something else to get the track to conform to that level. And we'll talk a little bit later about how you know when there's too much limiting going on. The thing that you want to make sure of,
is that the limiter is guiding you towards a solution This is not to say that the limiter is
early in the signal chain. The limiter is always the last thing in the signal
chain so while it's the first thing that I deploy it's the thing that's at the
very end that gives me that safety and any of the signal processing I'm doing
beforehand happens upstream. There are two modes for
using a limiter they both result in the same effect but they force you to work a
little bit differently so one is an instance where you have a fixed
threshold and you increase the gain going into the limiter and what you'll
notice when you do that is that the peak level at the very top of the scale never
increases above where the threshold is set but that the level of the signal, the
average level if you will is moving up or down depending on what you're doing
with the gain going into the limiter. The second mode is where you adjust the
threshold and the limiter tool gives you some makeup gain. That is to say if you drop the threshold
down by 5 DB you will hear an increase in level of 5 DB coming
out of the output. You'll still have that absolute ceiling if you will in the limiter itself but it's a different way of interacting with the limiter. In either case and they're both totally valid ways of working but in either case one of the most important things to do if you want to hear whether the limiter
is giving you too much gain reduction or too much of an effect is to make sure
that when you listen with the limiter on and with the limiter off you listen in a
way that's level matched. In Ozone that facility is fairly easily affected by
clicking on the ear tool that's on the right hand side underneath the meters
and that will level match the playback so that you can bypass the limiter and
put it back in and really hear a true a/b difference One of the core
practices of mastering and really I would say mixing as well is something
that we call level matched A/B comparing. What this refers to just like
it sounds is to make sure that when you're listening for differences in your
signal processing you eliminate the variable that is playback level and this
is especially relevant when it comes to limiting if you've increased the gain by
5 DB as a result of limiting and then bypass and the level drops 5 DB and
increases 5 DB each time you negotiate navigate turn the the bypass on and off
you'll have a hard time really hearing what the difference is between the
unlimited or the original signal and the limited or the processed signal. There are lots of different ways of doing it you can do it in your DAW by setting up a
flat version of a file and a limited version on two tracks adjacent to each
other and then you can go back and forth between them or if your limiter, and many
limiters have this function built into them, but if your limiter like Ozone has
something like the level matched A/B comparing tool and you'll see there's
little ear just underneath and to the right of the meters you can turn that on
and then anytime you bypass the limiter it will level match the processed and
unprocessed version for you and you can decide whether you've got too much of an
imprint from the limiter or at least just hear what it is changing. For the purpose of understanding how to
effectively use a limiter, it's important that you understand the difference between peak and average level. It's also important in this modern music distribution paradigm, that you understand the difference between the momentary or short-term LUFS measurements, and something that we call integrated LUFS. I'm not going to go
into a long piece right now about these different sorts of level but suffice it
to say that in most cases in music production what we're interested in is
this short term phenomenon. We're trying to address peak level with a limiter as opposed to addressing the average or the RMS level Sure when you put a limiter in place and you move the input to the threshold up and down, you'll see the average level moving up and down but the limiter isn't really interacting with
that part of the signal. The one thing I do want to say
about integrated LUFS is this is the measurement that all of the streaming
services use to determine whether your song mix file piece of audio needs to be
increased in level or decreased in level and so what it does is measures an
entire file and gives it an integrated LUFS level or number connected to it. When you're mixing you're thinking about the difference between one snare hit and
the next or the difference between a verse and a chorus or or what have you
it's almost impossible to be thinking about your integrated LUFS level or
where your level is going to end up at the end of the day according to a
streaming service. this by the way is maybe one of the reasons why we
differentiate between the mixing mindset or the mixing practice and the mastering
practice. There is one way in which you can anticipate the level that your file will probably come out at when you send it to a streaming service. If you look at the hottest the peak level of a track and for most styles of music and most
program that's usually about two-thirds of the way through or the beginning of
the third chorus or you know whatever it is the apex the emotional peak of a
recording. If you wanted to end up for instance at
a level of -14 integrated LUFS, It's a clever thing to do to imagine that, well the beginning of the track slower, so if I want to average at -14 probably my loudest portion
needs to be somewhat over -14. Something in the order of 3 or 4 DB
higher than what you expect your average to need to be is a good guideline. Now obviously your mileage may vary depending on how much dynamic movement
there is over the course of a mix but that's a good trick or rule of thumb or
guiding principle at least to start you out with where to set your limiter. Given that a limiter is a safety and is largely just interacting with peak level that's a very short-term phenomenon. As I said, we're not really listening for the
color of the limiter but a limiter can impart a change in the sound quality and
it's not just about dulling peaks it's not just about grabbing the very top of
the waveform and restricting it. It also has this somewhat hidden kind of
insidious effect. The result is when you use a limiter in most cases is that you lose a little bit of very low frequency information and you will gain a little bit of high frequency distortion. It's usually not pleasing distortion it's
usually correlated to the input signal but not in a way that relates musically to the signal. So if you think about it, take a look at the spectrogram of any mix, you'll notice that the highest amplitude portion of the signal is usually in the low end,
it's usually in the bass. and given that a limiter is an
amplitude based interaction if you will it makes sense that the bass is the
thing that's going to be driving the limiter. Therefore the harder you drive a
limiter the more the base is going to be restricted compared to anything else in
the program. Now I think it's important to point out, when I say the word bass,
and you talk to mastering engineers there's bass which is the bass instrument, and the bass instrument will certainly get louder
with everything else, but we also talked about bass that is the low end, the low frequency portion of a signal. So that could be the very bottom of your 808 kick drum or the thump the low frequency transient and that's the thing where you will
initially start to lose energy. So when I say you'll lose a little bit of bass,
that's what I'm talking about Coupled with that and this is something that I think a lot of people don't know to look for, is that limiters have very very fast release times. If they're slow, they'll hold on to the signal for a long time and tend to be dulling or blurry. So most of the time, people use limiters with
a very fast release. That very fast release can turn into
high frequency distortion. So when you're using a limiter when you
turn it on and start to drop the threshold take a moment and listen for
what's happening to the very bottom of my kick but also think--is my hi-hat
coming up? Is the sibilance coming up in the vocal? Are the cymbals getting a
little bit sharper? Those things invariably will happen a little bit. It's not to say it's a bad thing per se, but it's a change that you should know to look for to help you more effectively use limiters. One of the things to consider when using a limiter is to what extent we want to link the two channels. When I start talking about compression, I almost always recommend using a fully
correlated sidechain input into the compressor. That means that if there's a tom fill that goes from one side to the other that throws a lot of level into the compressor, the whole stereo image won't steer around. If the sidechain is taking a sum of both channels the compressor will move up and down together. In a limiter so long as we're not pushing the limiter too hard or too deep into it, we have some latitude to decouple the sidechain input, decouple the two channels as they're
interacting with the limiter. This can result in a
more or less wide sense of the soundstage and it's something that you
can just play around with and decide what works better for you. some limiters
will let you go for fully coupled decoupled or some amount of decoupling
to give you a little bit of control over the sense of stereo width. People often wonder what the difference in the IRC modes is in Ozone. So I'm just gonna take a moment to talk about what the differences are and how I think about them. IRC I is something similar to an analogue limiter. It has very little of a look ahead to it so it it will react almost immediately to the input signal behaving a lot more the way an analog limiter does. so the result is that you get a lot more transient reduction, you notice sort of a warmer punchier sound coming from it. If I'm working on a rock record or a punk record or something that needs more punch coming through the middle I'll choose IRC I to give me a little bit more of that it's almost like a band-limited sound quality coming from it. IRC II is more like a look ahead digital limiter and a look ahead means
that there's a buffer memory that allows the plug-in to look ahead to look at
what's coming down the pike and do a better job at reacting literally and
cleanly to the incoming signal So if you want to use a limiter that's going to be a little bit smoother and allow more of the low frequency warmth to come through a track, then IRC II is a better choice IRC III is a crazy sophisticated limiter
that for every instance of limiting actually accesses a lookup table chooses
the limiting algorithm that will produce the lowest amount of distortion and in
every instance of limiting deploy that. The result is a cleaner sound, a lot of people describe it as being brighter. It is guaranteed to have the least
distortion associated with it. So as an absolute safety at the end of my chain, IRC III is great because it imparts, well, you could describe it as bright, it imparts the least amount of color to the signal. And then there's IRC IV. This is the same three modes IRC I, II, and III, but in every case it's got a spectral shaper, or you can think of it as a very sophisticated dynamic EQ that sits in front of the limiter and just pulls back a little bit where
you have peaks in the spectrum so that the limiter isn't having to do as much
of the work. So that's just a quick tour through the 4 different modes. Sometimes you can associate the different
limiting modes with genre. For me, my approach is more "What's the resulting sound that I want to have?" I could imagine a string quartet, well, might be a stretch, but sure, a string quartet that might benefit from IRC I and a punk rock record that just needs a touch of IRC III or vice-versa, so it's really a matter of-- "What's the thing sound like---what results in the most pleasing sound?" or "Where do I want to take it?" and then I'll choose the correct mode for that. The release time of a limiter has a lot
to do with how a limiter sounds As you lengthen the release time of a limiter, it will do two things. It will hold on to the signal longer, obviously producing a smoother sound overall. So it will tend to defeat the excitement a little bit, and give you a little bit more of a smoothness. So you need to decide whether you want to maintain the sharpness of the rhythm of a track, or whether you want something that's a little mellower. It also, as you increase the release time, will produce a less distorted
sound coming from the limiter. So that also reinforces this idea that I'm describing about a longer release means a smoother, maybe darker and maybe in some cases more
pleasing sound in the limiter. Thanks for watching 'Are You Listening?'. If you want to be notified about other
episodes when they come out, subscribe to the YouTube channel and then you use the little bell icon. Feel free to go to the
izotope.com website--there lots of educational materials there. There is the iZo sessions where you can download sound samples and experiment with them yourself and you can download a copy of Ozone. I look forward to seeing you
next time, hope you find this useful
Just tried to edit my comment and accidentally deleted it.
Took me 5 mins to write that.... shit!
Well, considering I’m so loved on this platform... I will hold off on re-writing my post.
Bottom line is that I was interested in other peoples way of Mastering their work.... I would like to share my personal tribulations, but only if there’s any interest.
Does anyone use two Limiters in their master chain?
With the first one doing the main intelligent limiting, hopefully adding the least amount of color with no pumping, having a light Limit (basically the main safety net).... and then following that with a second Limiter that brings up the overall volume, again in the hopes of not adding any color and no audio pumping?
For as long as I’ve been doing this, I’ve always strived to have a perfect Mixdown that didn’t need anything at the Master stage, just a small rolloff at 40hz and a light Limit... but my issue has always been slight audio pumping, mainly with Snare drums and sometimes Hi-hats.
Now I’m trying to use two Limiters in the fashion described at the top of this post. It is a lot of experimentation, because of all the necessary ABing on multiple systems with multiple renders, etc., but I have a feeling that I will soon be utilizing this method moving forward, since it seems to keep my dynamics intact much better compared to simply using a single Limiter and lightly squashing the Mix.
My main issue is that using a second Limiter is going to add twice the color and possible high frequency distortion to the end result... and it’s difficult for my ears to discern this, even after myriad listening sessions across a range of different setups, including cans.... especially in cans since that’s all I have access to for my current project, which is no where near ideal for this kind of work.
I have a Mix for my current E.P. WIP that’s uploaded to Soundcloud right now, it’s a House tune at the top of my feed.
It’s a good example of what I’m going through atm. At best the audio pumping is inaudible to most ears, at worst it’s audible to those with discerning ears. It is a simple ‘squash’ style Limit.
I’m currently experimenting with trying a dual Limiter style to get the loudness I seek, but at the same time leave all the dynamics intact, while adding the least amount of color possible, with zero audio pumping.
Have any of y’all mastered this skill? No pun intended... 😎