The -6dBFS Rule

Performance & Techniques Music Production
1

The -6dBFS Rule

Original blog post here → http://innerportalstudio.com/the-6db-rule/

Been seeing a lot of new producers asking why they should keep their DAW master meters around -6dBFS (or -3dBFS, -4dBFS, etc). While a lot of mastering engineers are the ones asking for mixdowns this way, there seems to be some confusion that this is something that is needed for mastering. In truth, mastering engineers could care less what the headroom of your tracks is, as long as there is some.

The real reason you want to try and keep some headroom in your mix downs is to make sure that you’re not inadvertently clipping. A lot of synths and effects use random modulations, even some dynamics processors are designed to mimic analog processors and thus might have slight variations each time you process audio through them.

By aiming to keep roughly 6dB of headroom on the master channel as you do your mix, you’re just ensuring that some of this randomness doesn’t clip. Just because your mix peaks at exactly -1dBFS one time when you play the song, doesn’t mean it will peak to that same value each and every time. Leaving some headroom just eliminates having to worry about this, it’s a safety net, nothing more.

In a perfect world if you’re 100% sure your mix is not clipping, you can render it as close to 0dBFS as you want. But that doesn’t mean that you’re going to gain anything by doing so, your mix won’t sound better. So there’s really no need to push things that hard and risk clipping your file permanently.

As you can see, I tend to tell people to aim for roughly -6dBFS, but that’s just a personal preference. Some people recommend -3dBFS as well, either will work fine, just be sure you’re also working at 24bit or higher.

So what if your mix is too hot, and you don’t know the best way to fix it? In general I tell people it’s best to get in the habit of leaving the master fader at 0 and just lowering all the tracks by the same amount until you have the headroom you want. But there’s really nothing wrong with just lowering the master fader too, if that works better because you have a lot of track volume automation for instance. Use whatever is easiest for you, the key is just to get that safety in place.

I hope that helps clear some of this up, let me know if you have any questions!

2

Good stuff as always bud!

Thx for posting.

3

No problem!

4

Interesting!

5

Good info thank you for sharing!

6

Great advice! Im new to making music any other tips/advice you could give me?

7

Try here:

http://innerportalstudio.com/guides/

http://innerportalstudio.com/blog/

8

This is generally a safe thing to say, but isn’t completely accurate. With 16 bit sound, that means you are going to have a range of about 65000 for the value of each individual sample amplitude. By cutting off the high end, you are reducing the possible range of values. Whether or not the average person will be able to hear that difference is an entirely different matter, but you do gain something by getting as close to clipping as you possibly can without actually clipping in the first rendering of your track.

If you are going to leave headroom, then you may also find it desirable for your master to be 24 bit (or 32 bit float is actually probably best), which should help to alleviate this problem (assuming your eventual target will be 16 bit).

This does not mean that I do not advocate headroom, I totally do, just that you do lose a small amount of amplitude accuracy in your recording that way.

9

If your tone generators have enough random element that they may manage to clip unexpectedly, then you should perhaps consider a compressor/limiter for those edge cases.

10

Thanks for the links! Now I have a lot of good stuff to read and help my with making music!

11

@deathy,

  1. You’re “right”, except that -6dBFS corresponds to losing 1 bit. 15 bit audio is still worlds beyond the dynamic range of tape, vinyl, or basically anything else. Plus, modern music is so compressed that you’ll never touch the noise floor to begin with.

Raising the noise floor by 1 bit is a MUCH better compromise than risking clipping.

12

Yeah, that’s why I said it’s generally a safe thing to say, but not entirely accurate… I do shoot for -6dBFS headroom myself.

13

I think sometimes we have to accept there’s going to be practical trade offs between having a good workflow, and established digital audio theory. These days I definitely have gone more towards encouraging the former, versus trying to get people to do things strictly from a theoretical standpoint. But thanks for the clarifications, I admit I steer well away from the maths side of things these days :slight_smile:

14

Basically it’s don’t mush your mix because when it comes to mastering they won’t be able to do the best on a mushy clipped mix. When you’ve finished your mix invariably you will want it to sound as loud as others when you put it on Mixcloud/Soundcloud etc, and to do it you limit. Just make sure you do a pass un-limited keeping all the dynamic range should you wish to get it professionally mastered.

15

I do feel that it’s worth pointing out - 1 bit of loss means halving your accuracy, so this isn’t exactly a small loss. Going from 16 bit to 15 bit means going from 65,534 possible values to 32,767. One bit sounds small when you think of it on terms of just a 0 or 1, but it starts to mean a lot when you string them together.

16

It’s still the quietest parts that you’re missing.

And unless you want to be able to hear things that peak on the meter at -90dBFS in full detail while nothing else is going on…it doesn’t matter.

Everything is still there, the noise floor is just 6dB louder. If you’re running your DAW in 24-bit or 32-bit float like you should be, it’s even less relevant.

17

Cheers for posting this