Version française


? | Home page | Tutorial | Mixing in practice | Mastering

Homepage
Home page


MASTERING

Reminder: we work on this song I wrote: Life


(mp3 file)

Quick definition

We should say "pre-mastering" instead of "Mastering", but everyone says "Mastering", so I'll do the same.

Theoretically, the mastering is the process that results in a "Master", which is the final version of a recording before it's transfered to a physical medium (vinyl, cassette, CD, etc.), while controlling that all established norms are respected for the type of medium. Therefore, it's the last creative process as well as the last purely technical process of a recording.

But most of the time, when someone speaks about mastering, it's merely the last creative step of a musical work, the one that comes right after mixing and before broadcasting it to the public, whether that public be yourself and your family, or millions of people!

On this page

Before anything else... | What do I need? | Multiband compressor | Limiter | Conclusion

Top of page

Before anything else...

Before you start mastering anything, make sure you follow these rules:

- The mix must be completed. Don't start the mastering process unless you are happy with your mixing and there is nothing more to do in that area. Because mixing is the art of polishing each instrument individually to get a consistent whole, while mastering is the art of polishing a whole tune as a single entity. You don't deal with various parts anymore, you deal with the whole mix altogether. Which implies it must be finished.

- In order to be objective once the mix is over, I advise you to wait one day or two before starting the mastering. As the hearing memory is rather short, you will rediscover your work with a fresh ear if you wait that long.

- During the mastering process, it's better to work with studio monitors rather than with a headset, even more so than for the mix. But as always, it's better to work with a very good pair of headphones than with mediocre monitors. If you work with headphones, be aware that your perception of stereo sounds, as well as the balance of lows and highs are different with headphones. But if you know your headphones well, if you use reference songs that you know as well, you will be able to fix any possible mistakes by yourself.

- In professional studios, the mastering engineer is not the same as the mixing engineer. That gives him a fresh and neutral look over the song. Mastering monitors are also particularly accurate and placed in a well-studied room, sound-wise.
In a home studio, you will most probably mix and master yourself, with the same monitors and/or headphones (and not necessarily the best ones), in an untreated room of your appartment or house, with daily life sounds around you (cars in the street, noisy neighbors, family members in the next room...). And you're going to have to deal with it!

Top of page

What do I need?

Not much, actually, but here are the essentials, sufficient for most home-studio owners:
- A multiband compressor
- A limiter

As an option, you may also use the following gear/plugins:
- An equalizer. If your mix is good, you probably don't need to balance frequencies when mastering, but a few (very few!) corrections might be necessary here and there.
- A reference song. To help you reach your goal, keep a song you like within reach (use a song in the same style as the song you're working on, don't refer to Mozart if you're mastering a Death Metal song...). Your aim will be to obtain a sound similar to that song.

Other tools may help you give your song the atmosphere you're looking for:
- Stereo enhancement plugins
- Spectrum analyser plugins that will analyse your song and compare it to your reference song
- Plugins that will let you know about the dynamics of the song, so you don't get a sound that is too compressed and crushed. Like TT Dynamic Range Meter available on this page.

Personally, I only use a multiband compressor and a limiter. The first one helps me to highlight some frequency ranges to shape and enhance the sound. The difference with and without a multiband compressor is staggering.
The limiter will boost the perceived sound volume and make the song more energetic. But don't overdo it or you might lose some dynamics and make your song too aggressive and tiring to listen to.

Top of page

MULTIBAND COMPRESSOR

To be clear, on the Master bus, the bus towards which all other main buses converge (guitars, bass, vocals, synths, drums), I put a multiband compressor plugin on top.

Mise à part cette caractéristique, cela reste un compresseur, avec les réglages suivants :

A multiband compressor will apply different compression settings on frequency ranges defined by the user. For example, you may want to compress a lot the low frequencies, leave the mediums untouched, and slightly compresse the highs.

- Threshold: defines the level from which the compressor starts reacting.

- Ratio: defines the compression level you want to apply, from 1 (no compression) to infinite (in this case, the compressor becomes a limiter).

- Attack: defines the speed at which the compression reacts once the threshold is reached.

- Release: defines the speed at which the compressor stops reacting once the volume level goes back below the threshold.

- Knee: defines whether the compression starts reacting precisely when the threshold is reached, or starts reacting gradually when the volume level comes close to the threshold.

- Gain: the more the sound is compressed, the lower the volume gets. To compensate for the volume loss due to compression, use the gain.

Multidynamics (the multiband compressor plugin that I use) has two gain knobs; lo-gain and hi-gain. Lo-gain is the gain that is applied to the sound signal when it is below the threshold, whereas hi-gain is the gain applied to the signal when it is above the threshold. The ratio determines the transition from one to the other.

To use Multidynamics like a compressor, lo-gain must be higher than hi-gain, which will reduce the signal gain when it gets over the threshold. If you do the opposite (hi-gain higher than lo-gain), the compressor becomes an expander, that is to say the opposite of a compressor. For now, we are looking forward to using the plugin as a compressor and not as an expander.

- Output: the overall output volume.
The Wave Arts multiband compressor Multidynamics can use from 3 to 6 bands. Here, I defined 4 frequency ranges (bands):
- From 0 to 50 Hz, for infra-basses
- From 50 to 175 Hz, for lows
- From 175 Hz to 2 KHz, for mediums
- From 2 to 20 KHz, for highs.

Why these specific frequencies? Because after various trials, I found this was the best setting for this song. Nothing magical about it. Trials, errors, more trials...

Settings used for the first frequency band (from 0 to 50 Hz):
- Lo-gain: 0 dB
- Threshold: -24 dB
- Hi-gain: -5 dB
- Ratio: 2
- Attack: 35 ms
- Release: 40 ms

Settings used for the second frequency band (from 50 to 175 Hz):
- Lo-gain: 6 dB
- Threshold: -21 dB
- Hi-gain: -8 dB
- Ratio: 2
- Attack: 16 ms
- Release: 25 ms

Settings used for the third frequency band (from 175 Hz to 2 KHz):
- Lo-gain: 3 dB
- Threshold: -21 dB
- Hi-gain: -6 dB
- Ratio: 2
- Attack: 5 ms
- Release: 10 ms

Settings used for the fourth frequency band (from 2 to 20 KHz):
- Lo-gain: 6 dB
- Threshold: -18 dB
- Hi-gain: -8 dB
- Ratio: 2
- Attack: 5 ms
- Release: 10 ms

Common settings for all bands:
- Knee: Medium
- Crossover slope: 30 dB/octave (behavior of transitions between frequency ranges)
- Lookahead: 5 ms (used to analyse sound peaks that can occur during play)
- Mode: Vintage

Here is the Master bus sound without any plugin:


(mp3 file)

Then the sound obtained with the multiband compressor Multidynamics:


(mp3 file)

Obviously, the multiband compressor allows for a much better sound clarity and all instruments find their place. Everything seems to breathe better, the sound is less confined, more open. Just better!

Wave Arts Multidynamics
Wave Arts Multidynamics is a commercial plugin. You may replace it with a free alternative plugin like 5orcery from Platinum Ears Plugins, a free 5 band compressor.

Download 5orcery
(5orcery.zip, 2.09 MB)

The zip file contains the PDF user manual, as well as a preset file named "5orcery_master.fxp". This is the preset I use for the song Life, and corresponds to the screenshot below.


With a free plugin...

I obtained a very similar result using the free multiband compressor 5orcery from Platinum Ears Plugins. Simpler than multidynamics, it offers all the settings you need. You get 5 frequency bands, you can manually choose the frequency that separates two bands, and each band has threshold, fain, attack and release knobs. The peak detection is achieved by selecting an RMS level for each band (RMS is an average level, similar to what the human ear can perceive), or the peaks (Pk). You can also activate or deactivate each band separately if you don't want to use all of them. You can adjust the Input signal, and you will find a Saturation knob, which does not add saturation, but is made to attenuate unexpected peaks. Finally, you have an overall Output level knob.

Using the settings shown on the screeshot above, I obtain a result pretty close to what I get with Multidynamics. It's a bit less accurate, a bit less defined, but it's definitely close.

Here is the sound I get with the multiband compressor 5orcery:


(mp3 file)
Platinum Ears Plugins - 5orcery
Platinum Ears Plugins - 5orcery
Top of page

LIMITER

We are one step away from the final result, and we are going to use a limiter. This is a tool that will boost the volume of a song, by raising up the volume from a defined threshold, up to a ceiling level you define manually. This is a great tool, but it can be damaging to the song. If you set a threshold too low, the sound is altered, it can become very aggressive, because the sound dynamics is crushed and everything will sound loud all the time. So, be gentle with it. With modern music (such as Dance or many Electronic musics), you can compress the signal much more than you would with musics that should have more constrast, where volume variations are desired and benefit the song, making it lively.

For the song Life, with the multiband compressor activated, the average RMS level is rather low, around -15 dB, with a dynamic range between -12 and -14 dB for the most part. This means we have quite a lot of margin, the dynamics is rather large, and we will be able to apply much compression without damaging the song, in order to raise the average level while maintaining good dynamics.

This is the dynamics we have before applying the limiter, measured with the plugin TT Dynamic Range Meter:
TT Dynamic Range Meter
In the top center, you can see the dynamic range at the moment the screenshot was made while playing the song (14.3 dB). At the bottom left and bottom right, you can see the average RMS level of the left and right channels (-16.3 dB and -15.6 dB).

Once the limiter is activated, we still have a dynamic margin between 9 and 11 dB. For your information, the album Nevermind by Nirvana, released in 1991, had a dynamic range of about 12 dB, as well as the album Wish You Were Here by Pink Floyd, released in 1975. The Fame Monster from Lady Gaga, released in 2009 has a dynamic range of 5 dB. Way more compressed!

The trend nowadays is to compress a lot, reducing the dynamic range as much as possible, which is bad for your ears and bad for musical finesse, but high compression levels allow for a high sound volume. It has been proved that a song is more attractive when played loudly, so record companies try to get the highest possible volume by compressing more and more. People are getting used to such high compression levels, they have become a standard and if you don't follow that trend, you may very well be looked down upon. Within a playlist, if the volume of a song is much lower than that of other songs, chances are the listener will just skip it, especially if they don't know it well and like it a lot. Yet, it would be so easy to simply raise the volume..

Here, the Wave Arts limiter FinalPlug is set with a threshold of -6 dB and a ceiling of -0.1 dB. This means that when the sound level reaches -6 dB, the signal will be boosted up to -0.1 dB. Release is set to Auto.
Here's the sound obtained with the limiter FinalPlug:


(mp3 file)
Wave Arts FinalPlug
Here is the dynamics measured after applying the limiter, at the same position in the song as the previous screenshot:
TT Dynamic Range Meter
As you can see, the RMS level goes from about -15 dB to about -9 dB, which corresponds to the 6 dB added with the limiter.
As for the dynamics range, it is reduced by about 3 dB, but we maintain around 10 dB.

Wave Arts FinalPlug is a commercial plugin. As an alternative, you may want to use the Classic Master Limiter from Kjaerhus.

Download Classic Master Limiter
(classic_master-limiter_v106.zip, 356 KB)


And with a free plugin...

Classic Master Limiter from Kjaerhus is a good limiter, very simple to use as there is only one knob. The only setting is the threshold. There is no ceiling setting, but it is in fact predefined to -0.2 dB. The sound obtained is slightly different from Wave Arts limiter's, but the difference is not that huge.
Here is the sound obtained with the Kjaerhus Classic Master Limiter:


(mp3 file)


Kjaerhus Classic Master Limiter
Top of page

CONCLUSION

Voilà... this tutorial is over. Mastering was the last stage.

From the begining, we went from...

a sound recorded in the box, with no effects:


(mp3 file)


We got to this version after mixing:


(mp3 file)


And obtained this final result after mastering:


(mp3 file)


Remember that this tutorial is not a magic formula that should be strictly followed all the time, with all songs of all genres. Mainly because it is not perfect, but also because there are many possible ways to mix a song to get various results. Finally, because each music, each song must be treated differently. Some plugins are useful for some songs and not for others, settings will need to change, etc.

I hope that with this tutorial, you understand better - if it wasn't already the case - how to mix a project, how to organize it, what is the purpose of the various plugins we used, how to set them, why you should use an EQ, a reverb, a compressor...

There are tons of other things to find out and learn, by yourself, with other people, through the Internet, this is a vast subject. Everyone has their habits, their hints and tips, but there's always more than one solution to a problem. Don't forget that the best way to improve your skills is to practice, again and again. So, experiment and have fun!!!

Mixing in practice : Preparing the session | Bass | Drums | Rhythm guitars | Solo guitars | Keyboards | Vocals | Mastering
Top of page


MESSAGES

(laisser un message)

Page de messages n° 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



lgwn
le 08/05/2012 à 18h42

Merci pour tout, super boulot.



Malandain
le 24/04/2012 à 14h37

Bonjour,

J'ai un dr4d akai qui enregistre en 16 bits (44/48khz).
Je relie numériquement mon dr4d à ma carte son (spdif) qui est reliée a mon ordi en usb. Ma D.A.W. est Mixbus qui enregistre en 32/24/16 bits.
Pour un bon enregistrement, faut-il mieux mettre tout en 16 bits ? (dr4d + carte son + D.A.W.)
Peut-on enregistrer le 16 bits du dr4d en 24 bits carte son + D.A.W. ?
Ou 16 bits du dr4d vers 16 bits carte son et 24 bits Mixbus ?
J'ai l'impression que tout en 16 bits, c'est mieux.

Merci d'éclairer ma lanterne.

A. Malandain

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

<em>Bonjour,

Bienvenue sur mon site !

Puisque le dr4d enregistre en 16 bits, il semble logique de penser que tu ne gagneras a priori pas en qualité en réglant carte son et séquenceur sur 24 bits.

C'est vrai si tu te contentes de "transférer" les sons enregistrés avec le dr4d vers ton séquenceur sans les traiter ensuite.

En revanche, si tu fais du mix depuis Mixbus, en utilisant des effets, là il vaut mieux être en 24 bits, car les calculs effectués par le logiciel seront plus précis.

Il est très probable que la différence sera imperceptible, et travailler en 16 bits de bout en bout ne va pas dégrader le son obtenu.

Je ne pense pas que beaucoup de gens soient capables d'entendre la différence entre un traitement effectué en 16 bits et le même traitement effectué en 24 bits.

Néanmoins, la différence existe au niveau de la précision des calculs, et l'accumulation des effets sur différentes piste peut entraîner en 16 bits un risque (très faible) d'avoir quelques petites approximations qui vont se traduire par une légère dégradation du son. Encore une fois, c'est théorique, ça peut arriver, mais ça ne sera pas forcément quelque chose qui va s'entendre.

Donc si tu es pointilleux, je te conseille de régler ta carte son et ta D.A.W. sur 24 bits.

Grebz</em>



Franck
le 02/04/2012 à 03h34

Bonjour Grebz,

J'ai vu que tu utilisais Sonar X1 Producer.
J'aimerais bien si tu veux que tu me donnes quelques truc pour ce logiciel car sur Internet, il n'y a que des vidéos anglo et je ne les comprend pas toutes.

Merci d'avance.

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

<em>Bonjour,

Pour ceux que ça intéresse, j'ai fait ce petit guide rapide pour savoir comment intégrer et utiliser un VSTi (synthé virtuel) dans Sonar X1. Ce ne sont que les bases, mais ça pourra servir à ceux qui sont perdus.

Télécharger "<a href="/Downloads/tutorial_homestudio/Synthés virtuels dans Sonar X1.pdf">Synthés virtuels dans Sonar X1</a>" (Fichier PDF de 1,20 Mo)

Grebz</em>



Molten
le 13/03/2012 à 01h27

Merci mon ami : Excellent !!!



Alain6465
le 08/03/2012 à 00h54

Merci pour ton site qui est très utile pour un novice comme moi dans la MAO, même si je fais de la musique depuis longtemps.

J'ai l'intention de revendre mon matériel hardware et d'acheter un séquenceur en soft. Je penche pour Sonar car j'aime bien l'interface graphique de studio instrument, mais je ne sais quelle version choisir, Sonar Studio ou Producer ou X1 etc.

De plus, je n'ai pas vu la VS-100 de Cakewalk dans tes choix de carte son, qu'en penses-tu ? Serait-ce un bon choix, car elle a été conçue pour être utilisée avec Sonar je crois.

Merci pour ta réponse et pour tes conseils très pédagogiques !

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

<em>Bonjour,

Merci de ta visite sur mon site.

Pour l'interface V-Studio 100, je ne l'ai pas listée car je ne l'ai jamais utilisée et j'ai simplement référencé quelques marques connues, mais ce n'est pas du tout un choix exhaustif. Je suis un simple passionné de musique, pas un professionnel du test, et je ne me permets de parler en détail que du matériel et des logiciels que je connais !

J'ai été voir divers avis d'utilisateurs de la VS-100 et ils sont globalement très positifs, tant en ce qui concerne la qualité des pré-amplis que de l'aspect pratique des différents boutons présents sur l'interface, qui donnent un accès direct à de nombreuses fonctions du séquenceur utilisé. Effectivement, la carte a été conçue par Roland, qui est aussi l'éditeur de Sonar, et donc Sonar reconnaît immédiatement la VS-100 et attribue directement les bonnes fonctions aux bonnes touches de l'interface. Mais la VS-100 fonctionne aussi très bien avec les autres séquenceurs, avec lesquels il est tout à fait possible de configurer les boutons pour leur attribuer les fonctions voulues. Conclusion : même si je ne l'ai jamais utilisée moi-même, cette interface me semble être un bon choix.

Pour le choix d'un séquenceur maintenant... En tant qu'utilisateur de Sonar depuis plusieurs années, je vais évidemment te dire que c'est un très bon choix. L'interface me convient, je le trouve assez intuitif à utiliser et largement assez riche et puissant pour moi. Point important : Sonar est livré avec beaucoup de plugins (VST, VSTi) et outils qui permettent de travailler sans avoir impérativement besoin d'acheter des outils complémentaires. Mais restons objectif, les Cubase et autres Reaper ou Presonus Studio One sont aussi d'excellents logiciels. Je crois que c'est surtout une question de feeling, de coup de cœur... et de prix ! Tous les séquenceurs sérieux permettent de parvenir plus ou moins aux mêmes résultats, chacun avec sa méthode et son organisation.

Dans la gamme Sonar, le dernier en date est Sonar X1, qui existe en 3 versions : Essential, Studio et Producer.
La différence se trouve dans le nombre d'outils livrés avec le logiciel.

<strong>Essential</strong> est la version la moins complète et la moins chère. Elle est disponible uniquement en version 32 bits et est limitée à 64 pistes audio (ce qui suffit à la plupart des gens, moi y compris). Elle ne propose que les fonctions de base (pas de synthés, batterie logicielle, outils de compression évolués, etc.). Environ 90 euros.

<strong>Studio</strong> est disponible en versions 32 ou 64 bits, n'a pas de limitation du nombre de pistes, et possède des outils complémentaires (divers VST, VSTi, V-Vocal pour corriger les problèmes de justesse des voix ou instruments, Audio Snap pour caler des enregistrements sur le tempo). Environ 180 euros.

Enfin, <strong>Producer</strong> est la version la plus complète et la plus chère bien sûr. Elle contient en plus différents outils d'excellente qualité comme le ProChannel (compresseur, EQ, saturation), Session Drummer 3, un logiciel de batterie plutôt bien foutu, les VSTi Dimension Pro (très bon synthé, très complet) et TruePianos (version allégée d'un piano virtuel de très bonne qualité à mon goût). Environ 390 euros.

À sa sortie fin 2010, X1 n'était pas super stable et plantait un peu trop souvent à mon goût. Du coup, en attendant les correctifs, je continuais à utiliser la version précédente, mais c'est très frustrant et décevant de payer pour un logiciel qui ne marche pas aussi bien que ce que vous attendez. Mais heureusement, les correctifs ont rapidement amélioré la situation et la version de X1 actuellement en vente est nickel.

J'ai encore quelques plantages de temps à autres, mais j'ai pu constater que c'était toujours à cause de certains VST que j'ai ajoutés. La faute donc aux programmeurs de ces VST qui perturbent Sonar dans certaines conditions (pas toujours très claires à déterminer, d'ailleurs). Ça n'arrive d'ailleurs qu'avec des VST gratuits. Je n'ai jamais eu de plantages à cause de VST du commerce. Normal, puisque les conditions de tests ne sont pas les mêmes... Entre un bidouilleur passionné qui développe son VST dans son coin et une société qui peut tester sur tous les séquenceurs du marché, il y a une différence. N'empêche que beaucoup de VST gratuits sont non seulement utiles, mais aussi performants que certains logiciels du commerce.

Grebz</em>

Top of page