Introduction
This tutorial is all about
- Advanced dynamic manipulation
- Transient life cycle
- Tone control
- Maximizing Loudness / dynamics
- Thoughtful mixing decisions / a necessary correction
- Abusing aliasing for maximum transparency
**This mix contain 3rd party plugins, no limiter.
Sometime, you got a off vocal / instrument track and for making covers, remix etc but it sound like unfinish mix because the file that you recieved from producer / composer are unmastered.
You may see this title for example
- off vocal マスタリング無 / マスタリング無 (no mastering) or (unmastered) which mean, this file does not recieved a final polishing by engineer.
- off vocal マスタリング有 (mastered) which mean, this file are polished by mastering engineer or the producer himself.
Some producer give you both type files which is generous, or you only get a mastering file or no mastering file.
You may curious why some producer offers a unmastered file as a option and this option doesn't make much sense but both files have it's own advantages.
For a mastered file is self-explanatory, complete and ready to be use. But for a unmastered file they offers a materials for anyone who want to mix himself or learn to mix but need a proper materials. Unmastered file offers more room to play around and enhance by the mixer.
Example (similar loudness)
Before mix
After mix
- Song name Palette
Originally produce by Yuyoyuppe
- Remix & Arrangement by yosimokazu
Understanding about genre and why it is important.
This is a rock song and very busy too. Which mean compatitive loudness are mandatory and because of this, we have to maximize the loudness, also need to control the tone and dynamics of the song at the same time. Unfortunately loudness and dynamics doesn't like to go along. So i have to make decisions and i choose to prioritize the maximum loudness.
Some people might disagree that making music loud is unnecessary. Although this topic is insanely controversial, but at the end of the day, we do what's best for the song and it seem like loud song sell the best.
Setting your project sample rate
Usually it's not really important but In this tutorial i will use96kHz or more (if possible) when i want to push a song to it's loudness limit this setup is important.
Tonal control (super important)
This is the first step, although it seem like a cakewalk because i only use 3 band EQ but it actually the most difficult one, since it require a thoughtful consideration. When using a broad range EQ for tonal correction it will affect an entire song. It rely on experience and developing taste. If you don't like rock song or not familiar with this genre, you probably stuck at this first step.
I using this EQ since it offer a very gentle EQ curve, which is well suit for a broad range tonal correction.
I decided to remove midrange around 1.5kHz at -2dB to soften the guitar sound and open more space for vocal tracks. I put a little bump around 4kHz at +1dB and 350Hz at 0.7dB to help add more weight to the track. Adding little bit of saturation inside this EQ plugin and it does a good job. You can use any EQ that offers a broad range of EQ.
Advanced dynamic manipulation
First part of dynamic control is to add compression, it's depend on the song, the way you choose compressor plugins may varies but in this song i need a highly programmable compressor that can fully control the behavior of the unit. You can use any plugins that suit your best however.
The first compressor on the list is programmed for RMS control
This setup offers a very clean and very smooth compression , although it may introduce transient due to RMS behavior.
The highlight of this plugin is the Unique "S-Curve" and on top of that you can control the amount of program dependency of the attack and release which make it so transparent even an extreme amount of compression.
The way i use this plugin is to setup all of necessary parameters and then only use mix knob to blend wet/dry signal to taste and 72% of wet signal is seem on point.
You can use any compressor that you are know it well or well engineered for such task, in this case i recommended Klanghlem DC8C 3 it's very cheap and sound amazing, another alternative like TDR Molot or free Molot compressor is also a really good choice for this type of work.
Second compressor is for a very fast attack/release but with RMS Behavior
This setup offers fast attack while retain RMS like behavior, problem with fast attack compressor is that it choke sound and we are applying fast attack compressor to entire song which mean this is like ducking entire song down during DJ talking. Drum parts are prone to this kind of problem. This type of song when guitar are the king so drums part are the queen of the song, you need to treat them with consideration
In this set up i choose to use Fastest attack and release time as this plugin allowed me to use. Then i set this plugin to behaved like RMS compressor by turning Peak Crest to the right (almost fully) with rather hard knee and very gentle compression ratio of 1.3:1
This will shape the peak to make it less peaky and in turn make percussion part sound more beefy but this also create another problem and this set up will make percussion lost it's aggressiveness like clicky sound and impact of the kick and snare.
This kind of problem is call "Transient life cycle" and there is no matter how many compressor plugin you use how many setup you have, it will repeat this "transient shaping" cycle endlessly. If you use fast acting compressor it will introduce flat sound with benefit of reduce the peak, if you use slow acting compressor it will introduce Transient so on and so on.
Third compressor is purposely programmed to protect attack of the low frequency (kick drum)
Although this compressor is barely do anything but their only job is to protect kick drum getting round off by next compressor that i will use next, this to ensure that kick drum will survive and having it's own space to live for the rest of the song. This compressor is programmed for low frequency ehnancer (mono) with full RMS behavior and control how compressor react to any given frequency band by setting the frequency dependent ratio(FDR)
This setting mean compressor will react to noting but frequency around 50Hz which is where kick drum fundamental frequency live for the rest of the song.
Using YIN mode to increases the amount of low frequency "even" order harmonics that produce during compression. A subtle kicks enhancement.
The last compressor is for quick auto RMS
I use this compressor to emulate hardware version unit of a 1960s PYE 845753/03 Compressor/Limiter That i've been using it once back in the day.
This compressor mimic the Limiter mode of this Unit, it may introduce a little bit of saturation, using this as leveler or final RMS control. Feel free to copy this setting.
Final polishing with EQ (very important)
This part is probably one of the main reason why most people hate mixing music, it require experience, constant training, good playback system and good pair of ears. I must tell people who read this guide that it's practically impossible to teach and translate to people. Every person are not the same... A bright light may seem darker for some people or too bright for them. Experience will help you make a better decision during mixing. Skill and technique can be teach and translate, no problem. Meanwhile experience and taste can not be teach.
I EQ for what i hear. Please, alway keep that in mind.
All process i've been put in the signal chain will always alter the signal, sometime you got a good result, lucky for you.
Unfortunately, i have to do a final polishing with EQ. This is what i do with it.
Low shelf at 60Hz, -0.8dB to reduce the bloom of the bass by tiny amount, but the main purpose is to make bass guitar more pronounce
Bell curve at 110Hz, +0.6dB to enhance bass guitar
Bell curve at 280Hz, +0.3dB with wide Q to add more attack to the snare
Steep bell cut at 1.15kHz, -5dB(Dynamic EQ mode) to reduce the resonance of the guitar sound, it can easily trigger tinnitus.
Bell curve at 2.2kHz, +0.4dB to add more bright to the guitar sound. This is a rock song after all.
Bell cut at 10kHz, -0.6dB with wide Q to soften the harshness and all splashing cymbals sound.
- Maximizing Loudness / dynamics
This purpose of this part is aim for Maximum loudness while maintain transparency. For this purpose i decided to use a cliper and it can be any cliper, so i will use a free one.
You may confuse that using clipper for transparent mix? is that even possible? If you read or watch most of the tutorial online they will tell you the same thing that using limiter for transparency mix, espcially a final mix and use clipper for dirty sound. I'm not saying that they all wrong but clipper are much more precise compared to limiter.
In the analog world, clipper is living inside all maximizer, AD/DA converter, it's the part of the unit.
If you look at the picture, i don't clip a huge amount. Mostly kick, snare and all other transient and i do not use oversampling because oversampling always apply filter on top and it and these filter will affect how sharp that clipper can cut the sound, although oversampling may reduce aliasing artifact. This is when project sample rate come in to play, most of the aliasing artifact will stay at high frequency and because we already set a project sampling rate at 96kHz or more which mean we have a massive headroom for aliasing before it start to affect inside our hearing range 20Hz-20kHz. This is how i abused aliasing for maximum transparency.
To this point you need to control the clipper precisely in order to Maximize loudness and transparency.
Clipper are known for making snappy sound you get more click / attack out of the kick and snare using clipper and because of that it feel more transparent.
The reason why i refuse to use oversampling.
This topic is heavily technical and controversial, it's bascially an entire ant colony.. not a rabbit holes
Most people only know that aliasing is bad and that's it. And.... they are correct! But how much of these artifact is considered "bad".
Generally oversampling offers very little benefit such as prevent aliasing and that's pretty much it.
Meanwhile it have a lot of negative downside such as
- Reducing headroom due to phase ripple that cause significantly higher peaks on material.
- It cause high frequency roll off because of brickwall like fillter that oversampling apply.
- It can cause phase changes at high frequency, if your cymbal sound so brittle, this might be the problem.
I'm not try to convince people or create a weird cult like anti-oversampling society. I'm just saying that there is no such thing like super duper state-of-the-art oversampling technique that will solve all aliasing issues without any downsides. In reality all solutions are imperfect, they all based on compromises.
I still use oversampling, sometime on purpose, sometime plugin force me to use it because i can't turn it off... i have no hate toward oversampling (why should i?) but with my honestly, most of the time oversampling is unnecessary.
I know in the Recent trends most people focus on oversampling way too much (some people go as far as call it "mandatory"). My recommendation is to use oversampling if it make an audible difference and oversampling as few times as possible (x2 oversampling would be ideal.)
Thanks for reading.
This tutorial is all about
- Advanced dynamic manipulation
- Transient life cycle
- Tone control
- Maximizing Loudness / dynamics
- Thoughtful mixing decisions / a necessary correction
- Abusing aliasing for maximum transparency
**This mix contain 3rd party plugins, no limiter.
Sometime, you got a off vocal / instrument track and for making covers, remix etc but it sound like unfinish mix because the file that you recieved from producer / composer are unmastered.
You may see this title for example
- off vocal マスタリング無 / マスタリング無 (no mastering) or (unmastered) which mean, this file does not recieved a final polishing by engineer.
- off vocal マスタリング有 (mastered) which mean, this file are polished by mastering engineer or the producer himself.
Some producer give you both type files which is generous, or you only get a mastering file or no mastering file.
You may curious why some producer offers a unmastered file as a option and this option doesn't make much sense but both files have it's own advantages.
For a mastered file is self-explanatory, complete and ready to be use. But for a unmastered file they offers a materials for anyone who want to mix himself or learn to mix but need a proper materials. Unmastered file offers more room to play around and enhance by the mixer.
Example (similar loudness)
Before mix
After mix
- Song name Palette
Originally produce by Yuyoyuppe
- Remix & Arrangement by yosimokazu
Understanding about genre and why it is important.
This is a rock song and very busy too. Which mean compatitive loudness are mandatory and because of this, we have to maximize the loudness, also need to control the tone and dynamics of the song at the same time. Unfortunately loudness and dynamics doesn't like to go along. So i have to make decisions and i choose to prioritize the maximum loudness.
Some people might disagree that making music loud is unnecessary. Although this topic is insanely controversial, but at the end of the day, we do what's best for the song and it seem like loud song sell the best.
Setting your project sample rate
Usually it's not really important but In this tutorial i will use96kHz or more (if possible) when i want to push a song to it's loudness limit this setup is important.
Tonal control (super important)
This is the first step, although it seem like a cakewalk because i only use 3 band EQ but it actually the most difficult one, since it require a thoughtful consideration. When using a broad range EQ for tonal correction it will affect an entire song. It rely on experience and developing taste. If you don't like rock song or not familiar with this genre, you probably stuck at this first step.
I using this EQ since it offer a very gentle EQ curve, which is well suit for a broad range tonal correction.
I decided to remove midrange around 1.5kHz at -2dB to soften the guitar sound and open more space for vocal tracks. I put a little bump around 4kHz at +1dB and 350Hz at 0.7dB to help add more weight to the track. Adding little bit of saturation inside this EQ plugin and it does a good job. You can use any EQ that offers a broad range of EQ.
Advanced dynamic manipulation
First part of dynamic control is to add compression, it's depend on the song, the way you choose compressor plugins may varies but in this song i need a highly programmable compressor that can fully control the behavior of the unit. You can use any plugins that suit your best however.
The first compressor on the list is programmed for RMS control
This setup offers a very clean and very smooth compression , although it may introduce transient due to RMS behavior.
The highlight of this plugin is the Unique "S-Curve" and on top of that you can control the amount of program dependency of the attack and release which make it so transparent even an extreme amount of compression.
The way i use this plugin is to setup all of necessary parameters and then only use mix knob to blend wet/dry signal to taste and 72% of wet signal is seem on point.
You can use any compressor that you are know it well or well engineered for such task, in this case i recommended Klanghlem DC8C 3 it's very cheap and sound amazing, another alternative like TDR Molot or free Molot compressor is also a really good choice for this type of work.
Second compressor is for a very fast attack/release but with RMS Behavior
This setup offers fast attack while retain RMS like behavior, problem with fast attack compressor is that it choke sound and we are applying fast attack compressor to entire song which mean this is like ducking entire song down during DJ talking. Drum parts are prone to this kind of problem. This type of song when guitar are the king so drums part are the queen of the song, you need to treat them with consideration
In this set up i choose to use Fastest attack and release time as this plugin allowed me to use. Then i set this plugin to behaved like RMS compressor by turning Peak Crest to the right (almost fully) with rather hard knee and very gentle compression ratio of 1.3:1
This will shape the peak to make it less peaky and in turn make percussion part sound more beefy but this also create another problem and this set up will make percussion lost it's aggressiveness like clicky sound and impact of the kick and snare.
This kind of problem is call "Transient life cycle" and there is no matter how many compressor plugin you use how many setup you have, it will repeat this "transient shaping" cycle endlessly. If you use fast acting compressor it will introduce flat sound with benefit of reduce the peak, if you use slow acting compressor it will introduce Transient so on and so on.
Third compressor is purposely programmed to protect attack of the low frequency (kick drum)
Although this compressor is barely do anything but their only job is to protect kick drum getting round off by next compressor that i will use next, this to ensure that kick drum will survive and having it's own space to live for the rest of the song. This compressor is programmed for low frequency ehnancer (mono) with full RMS behavior and control how compressor react to any given frequency band by setting the frequency dependent ratio(FDR)
This setting mean compressor will react to noting but frequency around 50Hz which is where kick drum fundamental frequency live for the rest of the song.
Using YIN mode to increases the amount of low frequency "even" order harmonics that produce during compression. A subtle kicks enhancement.
The last compressor is for quick auto RMS
I use this compressor to emulate hardware version unit of a 1960s PYE 845753/03 Compressor/Limiter That i've been using it once back in the day.
This compressor mimic the Limiter mode of this Unit, it may introduce a little bit of saturation, using this as leveler or final RMS control. Feel free to copy this setting.
Final polishing with EQ (very important)
This part is probably one of the main reason why most people hate mixing music, it require experience, constant training, good playback system and good pair of ears. I must tell people who read this guide that it's practically impossible to teach and translate to people. Every person are not the same... A bright light may seem darker for some people or too bright for them. Experience will help you make a better decision during mixing. Skill and technique can be teach and translate, no problem. Meanwhile experience and taste can not be teach.
I EQ for what i hear. Please, alway keep that in mind.
All process i've been put in the signal chain will always alter the signal, sometime you got a good result, lucky for you.
Unfortunately, i have to do a final polishing with EQ. This is what i do with it.
Low shelf at 60Hz, -0.8dB to reduce the bloom of the bass by tiny amount, but the main purpose is to make bass guitar more pronounce
Bell curve at 110Hz, +0.6dB to enhance bass guitar
Bell curve at 280Hz, +0.3dB with wide Q to add more attack to the snare
Steep bell cut at 1.15kHz, -5dB(Dynamic EQ mode) to reduce the resonance of the guitar sound, it can easily trigger tinnitus.
Bell curve at 2.2kHz, +0.4dB to add more bright to the guitar sound. This is a rock song after all.
Bell cut at 10kHz, -0.6dB with wide Q to soften the harshness and all splashing cymbals sound.
- Maximizing Loudness / dynamics
This purpose of this part is aim for Maximum loudness while maintain transparency. For this purpose i decided to use a cliper and it can be any cliper, so i will use a free one.
You may confuse that using clipper for transparent mix? is that even possible? If you read or watch most of the tutorial online they will tell you the same thing that using limiter for transparency mix, espcially a final mix and use clipper for dirty sound. I'm not saying that they all wrong but clipper are much more precise compared to limiter.
In the analog world, clipper is living inside all maximizer, AD/DA converter, it's the part of the unit.
If you look at the picture, i don't clip a huge amount. Mostly kick, snare and all other transient and i do not use oversampling because oversampling always apply filter on top and it and these filter will affect how sharp that clipper can cut the sound, although oversampling may reduce aliasing artifact. This is when project sample rate come in to play, most of the aliasing artifact will stay at high frequency and because we already set a project sampling rate at 96kHz or more which mean we have a massive headroom for aliasing before it start to affect inside our hearing range 20Hz-20kHz. This is how i abused aliasing for maximum transparency.
To this point you need to control the clipper precisely in order to Maximize loudness and transparency.
Clipper are known for making snappy sound you get more click / attack out of the kick and snare using clipper and because of that it feel more transparent.
The reason why i refuse to use oversampling.
This topic is heavily technical and controversial, it's bascially an entire ant colony.. not a rabbit holes
Most people only know that aliasing is bad and that's it. And.... they are correct! But how much of these artifact is considered "bad".
Generally oversampling offers very little benefit such as prevent aliasing and that's pretty much it.
Meanwhile it have a lot of negative downside such as
- Reducing headroom due to phase ripple that cause significantly higher peaks on material.
- It cause high frequency roll off because of brickwall like fillter that oversampling apply.
- It can cause phase changes at high frequency, if your cymbal sound so brittle, this might be the problem.
I'm not try to convince people or create a weird cult like anti-oversampling society. I'm just saying that there is no such thing like super duper state-of-the-art oversampling technique that will solve all aliasing issues without any downsides. In reality all solutions are imperfect, they all based on compromises.
I still use oversampling, sometime on purpose, sometime plugin force me to use it because i can't turn it off... i have no hate toward oversampling (why should i?) but with my honestly, most of the time oversampling is unnecessary.
I know in the Recent trends most people focus on oversampling way too much (some people go as far as call it "mandatory"). My recommendation is to use oversampling if it make an audible difference and oversampling as few times as possible (x2 oversampling would be ideal.)
Thanks for reading.
