? | Home page | Tutorial | Mixing in theory

MIXING IN THEORY

During the mixing step, you will balance instruments, place them in the sound space, set their volumes, modify their frequencies, fix or hide the flaws. The purpose is to define each instrument's location, make them separated from one another and create a song that is pleasant to listen to.

Dynamics
Equalization
Panning
Compression
Reverb
Delay
Depth
Conclusion

DYNAMICS

Dynamics is the interval between the lowest and the highest volumes in a sound. OK, fine, but we are doing Rock music, who cares about low volume? We want to kick ass! Drums must beat, guitars must yell, the bass must make the walls shake and the vocals must be loud! Pump up the volume! Wooo... calm down!

If everything is loud, then nothing is

Yes, that's logical. If everything is loud all the time, nothing stands out. But it's the contrasts that bring both subtleties and raging moments to light. How do you expect to surprise your listeners with a sound hurricane if you've already given them thunder and lightning without a break? You have to spare some quiet moments before the storm. Let's compare this with drawing. A black dot on a white background will stand out more easily than the very same black dot on a dark-gray background.

Obvious? Yes, but many musicians just care for that "powerful sound" to such an extent that it eventually becomes tiring to listen to.

You can increase the contrasts through the very writing of the song (guitar/vocal bars, followed by a passage with 3 guitars, 1 distorted bass, big drums...), or through interpretation (the musicians play louder, faster...), or through the use of effects (saturation, compression...).

Contrasts don't have to be huge, it all depends on what kind of emotion you wish to convey, on the music genre, on the topic of the song, on what you are looking for...

Nowadays, too many songs just look (and sound) like this:
Over compression

Do your ears a favor: don't do this

EQ

EQ stands for equalization. It's a tool that allows you to set the level of the frequencies that make the sound. Basically, it's like a volume knob, but instead of changing the volume of a whole song, you lower or raise only the frequencies of your choice. Just like when you use a stereo amplifier for instance: you can find "bass", "medium" and "treble" knobs. They are a simple EQ that help you to tune the sound.

Well, you are going to apply this same principle to each of the instruments of your song.

The purpose of equalization is simple: changing the sound, so that each instrument primarily occupies the frequency range that defines it the best, without encroaching on the other instruments.

In other words, lower the level of the frequencies that are not necessary to an instrument without distorting it, and see to it that the various instruments main frequencies do not overlap. Each instrument will thus be alone in its own sound space without encroaching on the others, ensuring all instruments are clearly distinct from the rest.

The hard part is to do this without distorting the sound of the instruments, and also to see to it that instruments which have common frequencies by nature be well-marked.

You will deduce that the more instruments there are, the more complex mixing becomes.

PANNING

Music is usually mixed in stereo. Not always, but we have two ears, we use headsets or speakers that usually come in pairs, so it only makes sense to use left and right channels when mixing.

That's what panning is: acting on the width of a sound, from left to right.

On top of the EQing described above, we can also part instruments from one another by placing them in different spaces. If two guitars play simultaneously in your song, you may want to place one on the left and the other one on the right. This will help to distinguish them and it will also balance the sound.

COMPRESSION

Compression affects the dynamic range of a sound signal.

It reduces the amplitude modulations of the sound, that is to say the gap between the lowest and the highest levels.

I will use the above compressor as an example. All compressors have more or less the same basic functions, so it will be easy to reproduce the manipulations on alternative models.

Let's see how this works:

- Threshold: this defines the minimum level from which the compressor will start affecting the sound. 0 is the highest sound level.
- Ratio: this is the level of compression you are going to apply. A 1:1 ratio has no effect, whereas an infinite ratio will limit the sound to the threshold level. In this case, the compressor acts like a limiter and nothing will go higher than the threshold. The higher the ratio, the more compressed the sound will be, and the volume level will be reduced in proportion, until we use the output button.
- Knee: this knob sets how much the compression will affect the sound right before and after the threshold. The softer it is, the softer the transition will be. A hard knee will result in a more abrupt transition. It is rather difficult to hear the effect of this function.
- Attack: this knob indicates the time it takes for the compression to start acting once the threshold is reached. Set it short (0 ms) to compress the sound as soon as the threshold is reached, and set it longer to avoid compressing the attack of an instrument. For example, with a solo guitar, you may want to preserve the string attack and thus set the attack knob on 10 or 20 ms. It all depends on what you wish to do.
- Release: this knob indicates the time it takes for the compression to stop acting once the volume level goes back lower than the threshold. A long release allows for a less brutal change of gain.
- Output: it can also be called "Gain". This knob sets the output volume level after compression of the sound. It compensates for the loss of volume due to the compression.

So basically, compression takes all the sound energy there is between the threshold and the highest level of the original signal and crushes it in a smaller gap. Then you raise the whole signal with the output button, which eventually renders a denser, richer sound compared to the uncompressed signal. This can be subtle or obvious, depending on your settings and the original sound.

REVERB

Reverberation brings a little air to a recording that would otherwise sound hollow or confined. It is not necessary to add much of it, or it could drown the sound and give the impression that the recording took place in a cathedral, unless this is what you are looking to obtain. A lot of reverb can be nice at some specific moments in a song, on one specific instrument, but I doubt anyone would apply this effect at a high level on a whole song.

It is recommended to put some on vocals. This will add depth. The more you add, the more the sound will seem to move away and get drowned in the mix. On the contrary, the less there is, the more vocals will seem to be close to the listener. A nice little touch of reverb makes vocals sound richer, deeper and more beautiful. Just find the right level.

DELAY

Delay is a sound effect that allows to... delay an audio signal and have it repeated. It's very much like an echo: shout and you will hear your voice first, then the echo of your voice repeated one or several times.

This effect is somewhat related to the reverb effect, with one major difference: delay repeats at regular intervals without generating a continuous surrounding sound. You can use it instead of the reverb to avoid overcrowding the mix.

As a general rule, set this effect by basing it on the song tempo. Choose whichever interval you wish to hear the echoes. Here is a chart to help you:

Click on the chart to download the corresponding Excel file ("Delay by Tempo & Instrument Frequencies.xls" - 64 KB)

DEPTH

By changing not only the volume, but also the frequencies or some effects like reverb, you can give the impression that a sound is more or less near the listener. This adds depth to a recording.

In everyday life, how do you recognize a distant sound?

Its volume is low. At least relatively. A motorbike engine roaring in the distance will be much less loud than if it were 2 metres away from you. Thus, the first thing to do is to lower the volume of the instrument you want to place in the background.

Distant sounds don't have much bass frequencies. Bass frequencies require a lot of energy to move. Unless you hear a very powerful sound (like an explosion), the bass tone won't get to you. So you may cut low frequencies with a high-pass filter or an EQ.

On the other end of the sound spectrum, high frequencies of a distant sound are also fainted because they are absorbed by obstacles as they're moving. So you also need to lower these high frequencies.

Distant sounds are indistinct, unintelligible. Human voice, for example: if you hear someone speaking in the distance, you will of course recognize that this is a human voice (yes, your brain is that smart), you will probably know whether it belongs to a man or a woman, but you will not understand a word. You are going to have to reproduce this lack of clarity to simulate distance. A little reverb, a little chorus or some phaser touch will do. Give it a try and see what suits your music best.

You can't always tell exactly where the faint sound you're hearing is coming from. So pan the background instruments in the middle position and leave the sides to foreground instruments.

When listening to a music, a listener will mainly focus on changing elements. The use of a repeating, short musical pattern, treated as described above, will be easily forgotten and place itself in the background. The purpose of a background instrument is to add some fullness to the song, without being so interesting that it draws attention. It should be kept simple to have it blend into the song without being noticed too much.
And of course, on the contrary, foreground instruments will be the ones bringing volume, lowest and highest frequencies, melodies, and elements of surprise.

Example of a voice that gradually goes away from the listener
(full screen for a better view)

CONCLUSION

When adding effects, check the result by activating and deactivating them as the song (or the instrument) plays. It will help you to see whether an effect is too present, or not enough, whether it is necessary, whether it is too agressive, etc. You may like heavy reverb, but too much of it just drowns it all and makes the song inaudible. Control is essential. Too little is useless, you might as well remove the effect completely, whereas too much will make everything weighty and will surely waste the whole piece.

An effect often turns out to be... effective if you miss it when it's gone, even though you were not certain to hear it when it was on.

Recording - Previous | Next - Mixing in practice

MESSAGES

(leave a message)

Messages page # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35

Fomega
le 21/08/2010 à 00h41

Salut MAN,

Je suis musicien (guitariste harmoniciste) amateur désirant me lancer dans le domaine de la MAO. Je suis vraiment émerveillé par ce travail que je juge titanesque sur ton site. Merci. En visitant certains sites de MAO, je n'ai jamais été autant satisfait que ce soir. J'ai vraiment trouvé mon compte. Bonne continuation MAN. Bon boulot...

* * * * * * * * * * * * * * * *

Grebz : Voilà qui me fait super plaisir ! Ravi de pouvoir aider, au moins un peu.

Glagla
le 20/08/2010 à 14h27

Un grand Merci !!
Je monte mon 1er home studio... Habitué des vrais studios et tables analogiques, je suis (j'étais ..) perdu pour choisir les softs et le hardware.
Je ne connaissais pas ces sim d'ampli gratuites, le matos dont tu parles rentre dans mon raisonnable budget et tes explications claires me confortent dans mes choix.
Merci.

* * * * * * * * * * * * * * * *

Grebz : De rien, content que mes propres galères aujourd'hui résolues puissent aider d'autres personnes !
Bonne continuation musicale !

Honeys
le 05/07/2010 à 17h40

Salut,

J''ai noté une grosse erreur au début quand tu dis que tu enregistres en 48 khz, c'est tout ce qu'il ne faut pas faire, parce que le gain de qualité est absolument inexistant mais par contre le calcul pour revenir en 44,1 est beaucoup plus complexe que de passer de 88,2 à 44,1 (division par 2), le 48 khz et le 96 khz sont utilisés en video uniquement, jamais en musique, donc si on enregistre en suréchantillonnage, on se met en 88,2 khz et c'est tout. Et encore, ce n'est pas non plus très utile parce que ça sert juste à repousser l'effet d'aliasing sur des fréquences qui ne seront de toute facon pas audibles (on enregistre avec un échantillonnage 2 fois supérieur à la fréquence de note la plus aiguë qu'on va enregistrer).
Or, au-delà de 20 khz et meme avant on n'entend plus rien, d'où la norme fixée à 44,1 KHz). Donc si on enregistre à plus de 44,1 KHz, ça veut dire qu'on s'occupe des fréquences situées après 20 khz, donc inutiles.
Après, je dis ça mais j''enregistre en 88,2 khz, mais c'est plus par principe parce que j'ai de la place que par réelle nécessité.
Mais en tout cas un truc est sûr, ne pas enregistrer en 96 KHz et encore moins en 48 khz car on a plus de perte à la conversion qu'autre chose :-)

* * * * * * * * * * * * * * * *

Grebz : Oui, c'est vrai, le standard de la musique sur CD est de 44,1 KHz. Il est donc plus logique d'enregistrer en 44,1 ou dans son multiple 88,2 KHz.

Cela dit, ceci est surtout valable si l'on enregistre ensuite sur CD... Or, le CD n'est pas systématiquement le support final. Beaucoup de gens n'écoutent plus que du MP3 ou du OGG, ou d'autres formats compressés et là, la fréquence d'enregistrement n'a plus d'importance.

Un fichier WAV en 48 KHz converti en MP3 donnera un MP3 à 48 KHz et un certain nombre de Kbits/s choisi par la personne qui fera la conversion.

Je rajouterai que je n'ai personnellement jamais entendu à l'oreille la moindre perte de qualité entre mon enregistrement de base (48 KHz en 24 bits) et le résultat une fois gravé sur CD (donc à 44,1 KHz en 16 bits).

Peut-être certains entendent-ils une différence, mais je ne connais personne capable de différencier les deux. Peut-être des ingénieurs du son à l'oreille entraînée, c'est possible...

Mais la fréquence choisie importe moins que la profondeur d'échantillonnage. 24 bits plutôt que 16 bits ! Ça c'est une certitude !

CelticBreizh
le 15/04/2010 à 12h16

Concernant la repisse, j'ai vu des vidéos de séance d'enregistrement sur youtube je crois, où la chanteuse portait un casque, sans doute semi ouvert, et dont on avait bouché les deux côtés avec du gros ruban adhésif, genre tape grise. Je pense qu'ils avaient dû ajouter sous le ruban, de la mousse. Bref, peut-être une astuce pour garder ce très bon casque Beyer, mais chacun fait comme il l'entend, l'essentiel est de partager et c'est ce que fait Grebz à travers son site. Sans flatterie aucune, qu'il en soit vivement remercié.

* * * * * * * * * * * * * * * *

Merci, c'est sympa !

Le Beyer DT990 me sert à tout : écouter de la musique, regarder des films, jouer sur mon PC.

Pour la musique, sa fiabilité me permet de mixer en confiance le soir, sans déranger ni le voisinage, ni ma famille. Je réécoute de toute façon le résultat obtenu sur mes enceintes de monitoring.

Le DT990 a tendance à faire plus ressortir les sons graves que mes enceintes. À partir du moment où je le sais, je peux en tenir compte.

CelticBreizh
le 15/04/2010 à 01h39

Bonjour,
J'ai vu que vous utilisiez un casque Beyer Dynamic DT990, un excellent casque, mais je ne comprends pas comment vous faites pour vous enregistrer (voix) devant un micro statique par exemple. Pour moi, il ne s'agit pas véritablement d'un casque fermé comme le BEYER DT770 PRO 250 OHM. Par contre pour le mixage et l'enregistrement des guitares en direct, sans placement de micros devant les amplis, c'est certainement plus confortable et moins étouffant que le DT770.

* * * * * * * * * * * * * * * *

Bonjour Éric,

Le DT990 est en effet un casque semi-ouvert et laisse passer du son que peut capter un micro statique. En fait, j'utilise généralement mon SM57 dynamique pour enregistrer les voix, ce qui résoud le problème de repisse du casque.

Le SM57 n'est sûrement pas le meilleur micro du monde pour enregistrer des voix, mais le résultat est loin d'être mauvais.

Il m'arrive parfois d'enregistrer des voix avec un micro Rode NT5, micro statique, et j'utilise dans ce cas un autre casque, fermé, un vieux casque hifi de moyenne qualité, mais très largement suffisant pour m'en servir simplement en monitoring le temps d'un enregistrement.

Je n'enregistre jamais mes guitares avec des micros, je passe toujours en direct en branchant mes guitares (électriques ou électro-acoustiques) sur l'interface audio, via un préampli. L'utilisation d'un casque est donc inutile dans ce cas.

Grebz