In this pro mixing and mastering tutorial for a killer sounding final master, I am going to highlight a few key concepts that will instantly create great results!
My pro mixing and mastering philosophy is that to really master the art of the mix, as an engineer you need both a simple as well as an advanced approach to it.
When it comes to simple but very powerful techniques, there are a lot of things you can do that will definitely not work against you. One such technique is to measure the perceived loudness of the MID component of the mix as well as the SIDE component of the mix. Yet it is often not measured and when it is measured it is not measured completely and accurately enough.
The perceived loudness value is achieved from the combination of RMS, integrated LUFS and average frequency and should be measured at least on the main chorus of the song. Having a loudness value on the MID component and the SIDE component of the main chorus respectively, is I would say definitely a typical sign of pro engineering.
But when you have good loudness characteristics you want that loudness to reflect minimal amount of noise, so that the listener can keep turning up the volume a lot before the mix turns harsh. Being good at that my friends, that's on a whole other level than the loudness discussion.
Noise typically comes from the room, the monitoring, the noise floor, the noise present in effects, the recorded sounds and is present maybe especially much in the attack and decay portions of the timbre.
So here is the issue. Let's say you have done some recording, mixing and mastering and you discover it does not sound so good yet, you discover part of the problem is perceived loudness, there are some translation issues too and it does not have the kind of dynamic and ambience characteristics of a top list level mix. Besides this it kind of sounds a bit noisy at loud volume. Overall you are not satisfied. This issue is common and it is the result of not having the right amount of the right frequencies in the center and on the side and typically on top of that you might have some broad stroke compression too that amplifies the issue and makes the stereo image blurry too.
So ideally, you don't really want particular loudness levels in the center on the sides, but rather you want the right balance of the frequencies at that loudness. To resolve this issue I find it is key to first of all reduce the noise A LOT. I reduce it so much that I basically mute it, any sounds that are slightly noisy - muted. When you balance a mix it becomes very distracting when you have a set of frequencies that constantly add noise into the perception. The key is to remove that distraction so that you can focus on the sweet frequencies in the mix, because you need enough perceived loudness on those frequencies so that they add enough good vibration qualities to the perception of the mix as a whole. Noisy tracks are typically room, overhead, snare, kick, fx, perc, vocals and effects like distortion... When you are left with a set of very low noise type of tracks, check if there is still room bleed you can get rid of. Secondly, check are the attack and decay portions of these sounds at a high quality level or do they bring a certain amount of bite and blur to the perception? This is what compressors are good at taming. So use the attack and decay knobs to effectively reduce the noise. The focus should be to silence the mix so that when the listener consumes these sounds, it is a comfortable and relaxing experience. A typical example of a mix that lacks those qualities is when the vocals need de-essing but does not have it. At loud volume various vocal noises on various hits can break an otherwise emotional listening experience. This is highlighting how important it is to minimize the amount of perceived noise at the amount of perceived loudness you've achieved, I would say especially on the sides. Pros are good at that.
Noise can be both fairly well isolated to certain specific narrow frequency ranges and be present across the whole frequency range. That's the whole issue, that when you have a lot of perceived noise present at the perceived loudness, it becomes difficult to like the mix. You might have great music in there, but with all of the noise it messes with the vibe and the emotions the listeners get. So perceived loudness is kind of useless when it is like that. That's the key to understand - lowering perceived noise is more important than achieving perceived loudness.
This is one reason why many engineers think it sounds good when you keep the mix a bit less loud and more dynamic. That removes some of the perceived noise when you set the playback volume fader at an average position, but although you might have a little less noise in the attack and decay, you also will have a little less vibe from the perceived modulation. This brings us to yet another key topic for a killer sounding final master: perceived modulation at perceived loudness.
So now that you are aware that at the target perceived loudness level you need an amount of perceived noise that is low enough and an amount of perceived modulation that is high enough. The perceived modulation is directly proportional to the amount of gain, with more gain simply comes more modulation. This is why gain staging is so key. Gaining a signal adds noise and modulation and loudness. Perceived modulation is also something that is raised from harmonically enriching the content.
When you have achieved low enough perceived noise and high enough perceived modulation at the desired perceived loudness, it starts to become more a matter of achieving the right type of stereo image as well as the right type of sound and balance for a good translation. That is in my view the main purpose of mastering - I view noise, modulation and loudness to be something you need to have achieved in the post-mix and when you have achieved that in the post-mix, then you are ready to input that into mastering in order morph that sound into a certain highly attractive form, from the perspective of stereo, sound and translation - all amounting to a rich music experience when using any playback device.
Therefore, yes forming the mix well in mastering is absolutely key, but the potential of that is dramatically lowered by a noisy, modulation free mix with the wrong loudness and stereo characteristics. By being aware of this, you can improve your final master by first of all improving the perceived loudness and the perceived qualities at that perceived loudness - especially the perceived noise and the perceived modulation.
By having reached a certain quality at mixing you are ready for mastering. Now, what many engineers do at this stage is to target the mix bus. That is creating noise at the crossover points. So it is very common that the mix really sounds pretty good before mastering, but the final master does not really reflect that anymore because you now have a lot of additional perceived noise in the music. So it is absolutely key that when you start mastering you need to work minimal enough on the broad stroke level relative to on the the narrow stroke level. If you for instance have 10 tracks in the post-mix, 4 in the center, 6 on the side, it makes sense to dial in a static EQ adjustment on the mix bus very lightly and then when you are done with that, let's say you boosted 87 Hz by +0.1 dB, apply that adjustment in zero-phase mode on all 10 tracks individually. That is 1 dB of gain added to the mix, now you are much closer to your target perceived loudness and the mix still sounds great. A typical example when this works great, is when you apply frequency matching. In frequency matching certain frequencies are added and certain frequencies are removed. You cannot match by 100% with a single EQ on the mix bus with a good result, because the difference between source and target is dependent on the band dynamics difference. On the other hand, you cannot match 100% by targeting the dynamics only, because the difference between source and target is also dependent on the difference between the frequencies of the individual sounds within the mix between source and target. But when you have the right perceived noise and the right perceived modulation at the target perceived loudness in the post-mix, then you are much much closer to the end result and hence can do much more mild work with EQ and compression. Therefore what I like to do is to first of all apply a dynamic multiband compressor on the two center and side tracks individually. Then add EQ frequency matching applied on all post-mix tracks. At this point the overall dynamics in each band are much closer to the target, hence I can dial in a much more gentle static multiband compression on the center and side tracks individually to nail the target integrated LUFS levels. If I would not do this, I would be forced to apply multiband compression on the mix bus directly in order to be able to dial in a gentle enough curve, but this would still damage the stereo image too much. So dynamic multiband compression is applied first of all to roughly level the bands and overall frequency response on the mid and side. Then static EQ is added to smooth out and further add to the result of that without creating much additional noise at the crossovers. Finally, as a result of the two first steps we can now integrated LUFS adjust the bands manually with much less multiband compression/expanding. The combination creates the target sound type at the target loudness. And with the target noise and modulation perception in the content as well, it sounds great! It is a highly advanced combination of qualities that are mixed, but it creates a great final result because you get the right rms and peak characteristics across the frequency range which gives the master the right type of sound, and with the minimal amount of noise and the great amount of modulation at this loudness, and having that dialed in at the right stroke/scope levels and having it all in zero-phase, this just turns out great sounding.
A takeway point here is that, if you target a certain sound type for mastering, do so already at the mixing stage. It is generally less optimal to target one sound type in mixing and a different sound type in mastering. In other words you can achieve more quality by doing both a re-mix and a re-master, than by only doing a re-master. If you want a lot of character, add that character very granuarly as narrow stroke during mixing (even better during recording), but still target that towards a particular sound type by being aware of what type of sound you are going for at mastering. This gives you a good sound because the mix has gradually, smoothly and gently shifted towards your desired sound.
This leads us to the final pro conclusion. The more of the final sound quality you can achieve earlier, with higher precision of fx application and gain, the less you are going to have to do at post-mixing and mastering and the better it will sound at the end. The processing you apply at post-mixing and mastering is still essential, but with a more optimal weight ratio at each stage, it just turns out better sounding. And failing with the loudness level perception is instantly going to mess it all up... For instance the average frequency of the loudness might be such that you have totally wrong height of the center and totally wrong height of the side, so now an entire dimension of the mix might not fit the song at all.
The good thing about this type of approach, is that you are rarely going to have to deal with translation issues. That is something one should appreciate the value of.
Good monitoring improves the performance of all of this, especially the more important steps like reducing noise and adding modulation.
Keep in mind that if you apply a mastering strategy where you do some gaining on the entire signal at the end, maybe into some clipping, keep in mind that although this increases the modulation, it also brings up the noise and depending on what is behind it, the noise:modulation ratio can become significantly worse. It is therefore better to get the gain from earlier stages instead where you can add the gain without getting a lot of noise added too.
I know a lot of effects, hardware and software, that add a lot of noise. They are, with a few exceptions, rarely worth having in the signal path.
Finally, what essentially becomes most important when you work to achieve a final master quality like this, is by incorporating proper A-B across all stages of the music production process. A-B is for instance what will ensure you have an appropriate amount of vocals and bass in the mix. Basic things like that is highly important when you want a certain type of quality and sound. For instance too quiet vocals will significantly reduce the modulation of the mix. Too loud vocals can on the other hand bring some serious amounts of noise into the perception of the mix. So working on getting a good noise:modulation ratio on the vocals as well as the right perceived loudness level is highly important... (proper A-B...)
There are situations when no matter how you gain stage the mix, you are always ending up with too much noise. This is usually because the total attack time of all compressors/limiters in the mix when combined, is too fast and the release too slow, but is also due to unbalanced frequency response. Be cautious about effects that do not let you control compression attack and release time, it can add up to a problem. I avoid limiters a lot, instead I use compressors with max slow attack and max fast release, then when I tune I tune them both at the same time, meaning the release gets slower and the attack gets faster. This means that the more "hammering" that is added by the faster attack, the less frequent it becomes, so it balances out some of the negative effects of a faster attack.
My pro mixing and mastering philosophy is that to really master the art of the mix, as an engineer you need both a simple as well as an advanced approach to it.
When it comes to simple but very powerful techniques, there are a lot of things you can do that will definitely not work against you. One such technique is to measure the perceived loudness of the MID component of the mix as well as the SIDE component of the mix. Yet it is often not measured and when it is measured it is not measured completely and accurately enough.
The perceived loudness value is achieved from the combination of RMS, integrated LUFS and average frequency and should be measured at least on the main chorus of the song. Having a loudness value on the MID component and the SIDE component of the main chorus respectively, is I would say definitely a typical sign of pro engineering.
But when you have good loudness characteristics you want that loudness to reflect minimal amount of noise, so that the listener can keep turning up the volume a lot before the mix turns harsh. Being good at that my friends, that's on a whole other level than the loudness discussion.
Noise typically comes from the room, the monitoring, the noise floor, the noise present in effects, the recorded sounds and is present maybe especially much in the attack and decay portions of the timbre.
So here is the issue. Let's say you have done some recording, mixing and mastering and you discover it does not sound so good yet, you discover part of the problem is perceived loudness, there are some translation issues too and it does not have the kind of dynamic and ambience characteristics of a top list level mix. Besides this it kind of sounds a bit noisy at loud volume. Overall you are not satisfied. This issue is common and it is the result of not having the right amount of the right frequencies in the center and on the side and typically on top of that you might have some broad stroke compression too that amplifies the issue and makes the stereo image blurry too.
So ideally, you don't really want particular loudness levels in the center on the sides, but rather you want the right balance of the frequencies at that loudness. To resolve this issue I find it is key to first of all reduce the noise A LOT. I reduce it so much that I basically mute it, any sounds that are slightly noisy - muted. When you balance a mix it becomes very distracting when you have a set of frequencies that constantly add noise into the perception. The key is to remove that distraction so that you can focus on the sweet frequencies in the mix, because you need enough perceived loudness on those frequencies so that they add enough good vibration qualities to the perception of the mix as a whole. Noisy tracks are typically room, overhead, snare, kick, fx, perc, vocals and effects like distortion... When you are left with a set of very low noise type of tracks, check if there is still room bleed you can get rid of. Secondly, check are the attack and decay portions of these sounds at a high quality level or do they bring a certain amount of bite and blur to the perception? This is what compressors are good at taming. So use the attack and decay knobs to effectively reduce the noise. The focus should be to silence the mix so that when the listener consumes these sounds, it is a comfortable and relaxing experience. A typical example of a mix that lacks those qualities is when the vocals need de-essing but does not have it. At loud volume various vocal noises on various hits can break an otherwise emotional listening experience. This is highlighting how important it is to minimize the amount of perceived noise at the amount of perceived loudness you've achieved, I would say especially on the sides. Pros are good at that.
Noise can be both fairly well isolated to certain specific narrow frequency ranges and be present across the whole frequency range. That's the whole issue, that when you have a lot of perceived noise present at the perceived loudness, it becomes difficult to like the mix. You might have great music in there, but with all of the noise it messes with the vibe and the emotions the listeners get. So perceived loudness is kind of useless when it is like that. That's the key to understand - lowering perceived noise is more important than achieving perceived loudness.
This is one reason why many engineers think it sounds good when you keep the mix a bit less loud and more dynamic. That removes some of the perceived noise when you set the playback volume fader at an average position, but although you might have a little less noise in the attack and decay, you also will have a little less vibe from the perceived modulation. This brings us to yet another key topic for a killer sounding final master: perceived modulation at perceived loudness.
So now that you are aware that at the target perceived loudness level you need an amount of perceived noise that is low enough and an amount of perceived modulation that is high enough. The perceived modulation is directly proportional to the amount of gain, with more gain simply comes more modulation. This is why gain staging is so key. Gaining a signal adds noise and modulation and loudness. Perceived modulation is also something that is raised from harmonically enriching the content.
When you have achieved low enough perceived noise and high enough perceived modulation at the desired perceived loudness, it starts to become more a matter of achieving the right type of stereo image as well as the right type of sound and balance for a good translation. That is in my view the main purpose of mastering - I view noise, modulation and loudness to be something you need to have achieved in the post-mix and when you have achieved that in the post-mix, then you are ready to input that into mastering in order morph that sound into a certain highly attractive form, from the perspective of stereo, sound and translation - all amounting to a rich music experience when using any playback device.
Therefore, yes forming the mix well in mastering is absolutely key, but the potential of that is dramatically lowered by a noisy, modulation free mix with the wrong loudness and stereo characteristics. By being aware of this, you can improve your final master by first of all improving the perceived loudness and the perceived qualities at that perceived loudness - especially the perceived noise and the perceived modulation.
By having reached a certain quality at mixing you are ready for mastering. Now, what many engineers do at this stage is to target the mix bus. That is creating noise at the crossover points. So it is very common that the mix really sounds pretty good before mastering, but the final master does not really reflect that anymore because you now have a lot of additional perceived noise in the music. So it is absolutely key that when you start mastering you need to work minimal enough on the broad stroke level relative to on the the narrow stroke level. If you for instance have 10 tracks in the post-mix, 4 in the center, 6 on the side, it makes sense to dial in a static EQ adjustment on the mix bus very lightly and then when you are done with that, let's say you boosted 87 Hz by +0.1 dB, apply that adjustment in zero-phase mode on all 10 tracks individually. That is 1 dB of gain added to the mix, now you are much closer to your target perceived loudness and the mix still sounds great. A typical example when this works great, is when you apply frequency matching. In frequency matching certain frequencies are added and certain frequencies are removed. You cannot match by 100% with a single EQ on the mix bus with a good result, because the difference between source and target is dependent on the band dynamics difference. On the other hand, you cannot match 100% by targeting the dynamics only, because the difference between source and target is also dependent on the difference between the frequencies of the individual sounds within the mix between source and target. But when you have the right perceived noise and the right perceived modulation at the target perceived loudness in the post-mix, then you are much much closer to the end result and hence can do much more mild work with EQ and compression. Therefore what I like to do is to first of all apply a dynamic multiband compressor on the two center and side tracks individually. Then add EQ frequency matching applied on all post-mix tracks. At this point the overall dynamics in each band are much closer to the target, hence I can dial in a much more gentle static multiband compression on the center and side tracks individually to nail the target integrated LUFS levels. If I would not do this, I would be forced to apply multiband compression on the mix bus directly in order to be able to dial in a gentle enough curve, but this would still damage the stereo image too much. So dynamic multiband compression is applied first of all to roughly level the bands and overall frequency response on the mid and side. Then static EQ is added to smooth out and further add to the result of that without creating much additional noise at the crossovers. Finally, as a result of the two first steps we can now integrated LUFS adjust the bands manually with much less multiband compression/expanding. The combination creates the target sound type at the target loudness. And with the target noise and modulation perception in the content as well, it sounds great! It is a highly advanced combination of qualities that are mixed, but it creates a great final result because you get the right rms and peak characteristics across the frequency range which gives the master the right type of sound, and with the minimal amount of noise and the great amount of modulation at this loudness, and having that dialed in at the right stroke/scope levels and having it all in zero-phase, this just turns out great sounding.
A takeway point here is that, if you target a certain sound type for mastering, do so already at the mixing stage. It is generally less optimal to target one sound type in mixing and a different sound type in mastering. In other words you can achieve more quality by doing both a re-mix and a re-master, than by only doing a re-master. If you want a lot of character, add that character very granuarly as narrow stroke during mixing (even better during recording), but still target that towards a particular sound type by being aware of what type of sound you are going for at mastering. This gives you a good sound because the mix has gradually, smoothly and gently shifted towards your desired sound.
This leads us to the final pro conclusion. The more of the final sound quality you can achieve earlier, with higher precision of fx application and gain, the less you are going to have to do at post-mixing and mastering and the better it will sound at the end. The processing you apply at post-mixing and mastering is still essential, but with a more optimal weight ratio at each stage, it just turns out better sounding. And failing with the loudness level perception is instantly going to mess it all up... For instance the average frequency of the loudness might be such that you have totally wrong height of the center and totally wrong height of the side, so now an entire dimension of the mix might not fit the song at all.
The good thing about this type of approach, is that you are rarely going to have to deal with translation issues. That is something one should appreciate the value of.
Good monitoring improves the performance of all of this, especially the more important steps like reducing noise and adding modulation.
Keep in mind that if you apply a mastering strategy where you do some gaining on the entire signal at the end, maybe into some clipping, keep in mind that although this increases the modulation, it also brings up the noise and depending on what is behind it, the noise:modulation ratio can become significantly worse. It is therefore better to get the gain from earlier stages instead where you can add the gain without getting a lot of noise added too.
I know a lot of effects, hardware and software, that add a lot of noise. They are, with a few exceptions, rarely worth having in the signal path.
Finally, what essentially becomes most important when you work to achieve a final master quality like this, is by incorporating proper A-B across all stages of the music production process. A-B is for instance what will ensure you have an appropriate amount of vocals and bass in the mix. Basic things like that is highly important when you want a certain type of quality and sound. For instance too quiet vocals will significantly reduce the modulation of the mix. Too loud vocals can on the other hand bring some serious amounts of noise into the perception of the mix. So working on getting a good noise:modulation ratio on the vocals as well as the right perceived loudness level is highly important... (proper A-B...)
There are situations when no matter how you gain stage the mix, you are always ending up with too much noise. This is usually because the total attack time of all compressors/limiters in the mix when combined, is too fast and the release too slow, but is also due to unbalanced frequency response. Be cautious about effects that do not let you control compression attack and release time, it can add up to a problem. I avoid limiters a lot, instead I use compressors with max slow attack and max fast release, then when I tune I tune them both at the same time, meaning the release gets slower and the attack gets faster. This means that the more "hammering" that is added by the faster attack, the less frequent it becomes, so it balances out some of the negative effects of a faster attack.
Last edited:
