In this pro mixing tutorial I'm going to resolve some signal level confusion that in my experience many engineers are struggling with. And the reason why the struggling happens is simple - it's difficult stuff, the nature of the issue is complex, it's not something you just fix within a few mixing sessions. Why? Because it deals with the science of transient shaping through waveform combining, which is extremely low level.
In my experience there is a lot of confusion around getting the desired loudness level. What engineers mean with this is basically that when they compare their mix with let's say a hit mix in the same genre, they notice that there is some difference between them and it appears to relate to loudness. The issue is not so much loudness although to some degree it is a loudness issue too, but the real issue is really how the mix sounds and feels, simply put how it is perceived at the loudness level being compared against. And yes, 9 times out of 10 an amateur mix does not sound the way a commercial pro mix sounds at a particular level, even with the same RMS and integrated LUFS levels. I'm going to try to nail that down for you...
There are many things that cause this issue, that's also why so many struggle with this because they think if you just solve one thing it will fly. No. If you have a monitoring issue, you have say 3 sound sources way too loud or quiet in the mix, you might on that alone feed the compressors say +-15 dB wrong signal contents - and you are not aware of it at all and because of that you also have the wrong gain staging on all other sound sources. This is just a fraction of the issue, it is much more complex than this, but this at least gives you an idea of why it is not so strange that you are struggling with level.
The first thing to ensure is a flat monitoring environment. But this is kind of basic, so I won't go into that, just remember how critical it is.
The next cause of the issue has to do with - gain staging. In my experience most engineers have no clue when it comes to gain staging, not what it is, not how you do it properly and not how it impacts a mix. The reason is simple - the long term gain level and the long term peak level are treated as one, meaning that when the peak level is wrong the gain level might be adjusted and when the gain level is wrong the peak level might be adjusted. It's essentially the wrong use of compression and volume. In practice what happens is that during maximization, let's say from -16 LUFS to -8 LUFS, the engineers target the mix hard with one or more limiters that has a very fast attack time. This indirectly forces the volume faders of the tracks down. And there goes the signal, the vibe, the vitality, the separation. It's all gone.
So it's absolutely key to understand that it is the more final stages where you set the perceived loudness, but the quality of what it sounds like at the perceived loudness is set much earlier, because that is the result of combining various types of waveforms - more specifically that you have certain modulation qualities, certain transient shapes etc. that then combine to create a signal that make the more broad compression strokes great sounding. And that's what is missing, but it is missing for a good reason.
Let's now then dig into the root of the issue. There are mainly two categories of the issue: routing/separation and compression. Routing is the fact that you are working too hard with compression on too much of the overall signal. Compression is the fact that you are imposing the wrong compression curves on the content, for instance you need much more dynamic and RMS based compression. You might have a little, for instance in terms of dynamic compression you might have some side chains active and in terms of RMS compression you might have some compressors that naturally act on the content more dependent on RMS rather than Peak. What you do not have is RMS, Peak, side chain as they relate to one and two stage compression. So essentially in order to get the right attack quality you need to sacrifice the modulation quality. When you then on top of that apply the wrong routing, the mix is going to lose a whole lot of vital vibe.
To find your way out of this rabbit hole, you need to switch from a top down approach to both a bottom up approach as well as a top down approach. Early on you apply the bottom up approach, later you apply the top down approach. What this means is that you need to engage the right compression both on each sound source and combinations of sound sources first of all. Then when you have the RMS, Peak, Side Chain applied correctly on one and two stage compression using compressors with the right compression curves, and you on top of that have the right routing and signal levels, then you are feeding the broader scoped compressors the right kind of content and they can now act on the content to make it sound great at the end. The problem is that you don't have that.
I will now go a bit deeper into gain staging. There are a whole range of pro techniques that could be discussed, gain staging is a very big topic, but I will mentioned a few key ones that you are likely unaware of, but will make your mixes better sounding once you become aware of them.
The first pro technique is this one: When a sound source in the mix is too loud, that is not what is essential. The essential part is whether it sounds great or not. Once you realize that you have been making the wrong volume moves for years and you stop doing those, you are going to be surprised about the results. So you've spotted a sound source that is too loud in the mix, but when you focus specifically on the sound of that you also notice hey it sounds great. That's not when you lower the volume on that track, that's when you raise the volume of all other tracks. AHA
Another key thing about gain staging has to do with gain staging not only with volume faders, but with EQs and multiband compressors too. As you might know, applying filters on the individual track level can do amazing things, like for instance applying hi pass filtering on vocals can all of a sudden create a really sweet low end. It's not wrong to use EQ filters and multiband compressors, it's just that it's cutting out too much of the signal because of how it is being applied. First handle the dynamics, then apply the EQ filtering, with filtering I mean hi and low pass filtering in this case. Because else what happens is that you are placing the pass filter cutoff point way too high or low (depending on whether it's in the low end or high end you apply it). So this is a key really - multiband compression works great, filtering works great - when they complement each other.
So now you've learned how to reduce loud peaks, how to remove not too much of the signal when you filter and how you need to engage a whole range of gain staging and dynamics processing moves in order to yield the right waveforms to feed into the broader compression. Let's now discuss the art of separation.
As you grow into a more skilled pro, you notice certain things that it takes time to become aware of. One such thing is that you need to separate the center panned tracks and the side panned tracks, because if you don't you lose the stereo image. This is kind of the basic pro level mixing, like being aware of what an EQ does to a stereo image when it is applied to a stereo signal. Most are aware of that. So far so good. But in my experience way too many engineers are still working way too broadly on the content and in ways that makes no sense, like having a music group track, other group tracks and an all vocals group track and so on. This type of routing creates very complex balancing issues that span across a broad range of mix qualities. In my view the better route is to have more than one group for the center panned tracks and more than one group for the side panned tracks. That alone is already a much better approach in my view. Based on my skill level up to this point, I would say it is optimal to have 4 groups or slots available for routing towards the center and 4 groups or slots available for routing towards the side. That becomes an 8 bus parallel processing strategy. So for the center it could be bass kick snare vocals - yes that means they are not grouped together and yes that means they can be side chained. Similarly on the side you could have guitars, keys, bgv etc., you have a little grouping but not so much, you can apply side chaining. With an 8 bus parallel processing strategy you can also apply side chaining between elements in the center and elements on the side and even between the two groups (center, side). Stuff like this is common out there. Take for instance some CLA mixes, you will notice that the side has a fairly long compressor release side chained to the center, which means that as soon as the vocals become less active the brain will all of a sudden notice for instance beautiful guitar reverb or background vocals on the side because it moves forward as the compression is slowly released. It is this kind of stuff that is so underrated out there, and what's overrated is the final limiting. Yes, broad stroke compression with the right compression curves does great things especially in combination with tape saturation, pitch shifting, harmonic exciting and great gain staging/post-eq, but it is really all of the steps you take with routing, compression, volume and EQ earlier that is going to be what makes it sound great at the end.
Another pro technique, that I don't think many here are aware of, is that side chaining does not have to be in relation to some other element, it can be in relation to itself. Yes, you have never thought about that. You can for instance have a sound source re-shape itself towards a certain sound as soon as it becomes somewhat "ugly" sounding. Like for instance a snare that starts to sound boxy and sharp on some hits, or a group of tracks like for instance bass and kick start becoming too heavy on some hits, re-shape it with self side chaining. So how to do that. You feed the signal to a signal chain designed to enhance the quality you don't like, that becomes the side chained signal, it never colors the signal directly, it just increases the sensitivity such that as soon as the sound source has a tendency towards a certain sound the compressor instantly deals with that tendency. The threshold basically becomes the tendency amount knob. And this scales to any sound at any scope/level in any mix with any compressor that supports side chaining, so imagine what it can do in the hands of a skilled engineer.
In my experience there is a lot of confusion around getting the desired loudness level. What engineers mean with this is basically that when they compare their mix with let's say a hit mix in the same genre, they notice that there is some difference between them and it appears to relate to loudness. The issue is not so much loudness although to some degree it is a loudness issue too, but the real issue is really how the mix sounds and feels, simply put how it is perceived at the loudness level being compared against. And yes, 9 times out of 10 an amateur mix does not sound the way a commercial pro mix sounds at a particular level, even with the same RMS and integrated LUFS levels. I'm going to try to nail that down for you...
There are many things that cause this issue, that's also why so many struggle with this because they think if you just solve one thing it will fly. No. If you have a monitoring issue, you have say 3 sound sources way too loud or quiet in the mix, you might on that alone feed the compressors say +-15 dB wrong signal contents - and you are not aware of it at all and because of that you also have the wrong gain staging on all other sound sources. This is just a fraction of the issue, it is much more complex than this, but this at least gives you an idea of why it is not so strange that you are struggling with level.
The first thing to ensure is a flat monitoring environment. But this is kind of basic, so I won't go into that, just remember how critical it is.
The next cause of the issue has to do with - gain staging. In my experience most engineers have no clue when it comes to gain staging, not what it is, not how you do it properly and not how it impacts a mix. The reason is simple - the long term gain level and the long term peak level are treated as one, meaning that when the peak level is wrong the gain level might be adjusted and when the gain level is wrong the peak level might be adjusted. It's essentially the wrong use of compression and volume. In practice what happens is that during maximization, let's say from -16 LUFS to -8 LUFS, the engineers target the mix hard with one or more limiters that has a very fast attack time. This indirectly forces the volume faders of the tracks down. And there goes the signal, the vibe, the vitality, the separation. It's all gone.
So it's absolutely key to understand that it is the more final stages where you set the perceived loudness, but the quality of what it sounds like at the perceived loudness is set much earlier, because that is the result of combining various types of waveforms - more specifically that you have certain modulation qualities, certain transient shapes etc. that then combine to create a signal that make the more broad compression strokes great sounding. And that's what is missing, but it is missing for a good reason.
Let's now then dig into the root of the issue. There are mainly two categories of the issue: routing/separation and compression. Routing is the fact that you are working too hard with compression on too much of the overall signal. Compression is the fact that you are imposing the wrong compression curves on the content, for instance you need much more dynamic and RMS based compression. You might have a little, for instance in terms of dynamic compression you might have some side chains active and in terms of RMS compression you might have some compressors that naturally act on the content more dependent on RMS rather than Peak. What you do not have is RMS, Peak, side chain as they relate to one and two stage compression. So essentially in order to get the right attack quality you need to sacrifice the modulation quality. When you then on top of that apply the wrong routing, the mix is going to lose a whole lot of vital vibe.
To find your way out of this rabbit hole, you need to switch from a top down approach to both a bottom up approach as well as a top down approach. Early on you apply the bottom up approach, later you apply the top down approach. What this means is that you need to engage the right compression both on each sound source and combinations of sound sources first of all. Then when you have the RMS, Peak, Side Chain applied correctly on one and two stage compression using compressors with the right compression curves, and you on top of that have the right routing and signal levels, then you are feeding the broader scoped compressors the right kind of content and they can now act on the content to make it sound great at the end. The problem is that you don't have that.
I will now go a bit deeper into gain staging. There are a whole range of pro techniques that could be discussed, gain staging is a very big topic, but I will mentioned a few key ones that you are likely unaware of, but will make your mixes better sounding once you become aware of them.
The first pro technique is this one: When a sound source in the mix is too loud, that is not what is essential. The essential part is whether it sounds great or not. Once you realize that you have been making the wrong volume moves for years and you stop doing those, you are going to be surprised about the results. So you've spotted a sound source that is too loud in the mix, but when you focus specifically on the sound of that you also notice hey it sounds great. That's not when you lower the volume on that track, that's when you raise the volume of all other tracks. AHA
Another key thing about gain staging has to do with gain staging not only with volume faders, but with EQs and multiband compressors too. As you might know, applying filters on the individual track level can do amazing things, like for instance applying hi pass filtering on vocals can all of a sudden create a really sweet low end. It's not wrong to use EQ filters and multiband compressors, it's just that it's cutting out too much of the signal because of how it is being applied. First handle the dynamics, then apply the EQ filtering, with filtering I mean hi and low pass filtering in this case. Because else what happens is that you are placing the pass filter cutoff point way too high or low (depending on whether it's in the low end or high end you apply it). So this is a key really - multiband compression works great, filtering works great - when they complement each other.
So now you've learned how to reduce loud peaks, how to remove not too much of the signal when you filter and how you need to engage a whole range of gain staging and dynamics processing moves in order to yield the right waveforms to feed into the broader compression. Let's now discuss the art of separation.
As you grow into a more skilled pro, you notice certain things that it takes time to become aware of. One such thing is that you need to separate the center panned tracks and the side panned tracks, because if you don't you lose the stereo image. This is kind of the basic pro level mixing, like being aware of what an EQ does to a stereo image when it is applied to a stereo signal. Most are aware of that. So far so good. But in my experience way too many engineers are still working way too broadly on the content and in ways that makes no sense, like having a music group track, other group tracks and an all vocals group track and so on. This type of routing creates very complex balancing issues that span across a broad range of mix qualities. In my view the better route is to have more than one group for the center panned tracks and more than one group for the side panned tracks. That alone is already a much better approach in my view. Based on my skill level up to this point, I would say it is optimal to have 4 groups or slots available for routing towards the center and 4 groups or slots available for routing towards the side. That becomes an 8 bus parallel processing strategy. So for the center it could be bass kick snare vocals - yes that means they are not grouped together and yes that means they can be side chained. Similarly on the side you could have guitars, keys, bgv etc., you have a little grouping but not so much, you can apply side chaining. With an 8 bus parallel processing strategy you can also apply side chaining between elements in the center and elements on the side and even between the two groups (center, side). Stuff like this is common out there. Take for instance some CLA mixes, you will notice that the side has a fairly long compressor release side chained to the center, which means that as soon as the vocals become less active the brain will all of a sudden notice for instance beautiful guitar reverb or background vocals on the side because it moves forward as the compression is slowly released. It is this kind of stuff that is so underrated out there, and what's overrated is the final limiting. Yes, broad stroke compression with the right compression curves does great things especially in combination with tape saturation, pitch shifting, harmonic exciting and great gain staging/post-eq, but it is really all of the steps you take with routing, compression, volume and EQ earlier that is going to be what makes it sound great at the end.
Another pro technique, that I don't think many here are aware of, is that side chaining does not have to be in relation to some other element, it can be in relation to itself. Yes, you have never thought about that. You can for instance have a sound source re-shape itself towards a certain sound as soon as it becomes somewhat "ugly" sounding. Like for instance a snare that starts to sound boxy and sharp on some hits, or a group of tracks like for instance bass and kick start becoming too heavy on some hits, re-shape it with self side chaining. So how to do that. You feed the signal to a signal chain designed to enhance the quality you don't like, that becomes the side chained signal, it never colors the signal directly, it just increases the sensitivity such that as soon as the sound source has a tendency towards a certain sound the compressor instantly deals with that tendency. The threshold basically becomes the tendency amount knob. And this scales to any sound at any scope/level in any mix with any compressor that supports side chaining, so imagine what it can do in the hands of a skilled engineer.
Last edited:
