I understand that headroom is needed between the mixing and mastering processes. For example, after a mix is done it may sit at -6dB so the person mastering the song has headroom to play with.

I have heard that if you send a slightly-below 0dB track to a masterer they will tell you to fix it and send it back to them with more headroom. That makes sense as they need headroom to play with.

I am trying to figure out if there's any difference between mixing something to -6dB and sending it to a masterer or mixing something to slightly-below 0dB and having the masterer pull the volume fader on the track down by -6dB to compensate.

I'd think that there'd be no difference, but I feel like I'm missing something fundamental about that. So, is there a difference and if so, what is the difference?

• "I have heard that" where? Please provide citations. Jul 7, 2020 at 12:07
• numerous videos and online tutorials. Sorry but I prefer not to detract from the main question with tons of links. Jul 7, 2020 at 12:09
• Well, imo the main question is based on a false premise - so having some background as to where the information came from might give us a hint as to how you arrived at the false premise. Jul 7, 2020 at 12:11
• I understand where you're coming from but I don't want to hunt down links. So your answer is that it is a false premise and there's no difference. Jul 7, 2020 at 12:33
• I agree with Tetsujin, it would help to see an example of what you're reading that talks about a mix "sitting at -6 dB so the mastering engineer has 'headroom'", because that could mean several different things and it's not clear what you're asking about. Normally, no mix would "sit at -6 dB". If "sit at" means peak levels, then -6 dB is unnecessarily low - might as well make then between -1 and -3 dB. If "sit at" means average levels, then -6 dB is way too high. If the crest factor is more than 12 dB, then it doesn't matter where the peaks are (within reason). Jul 8, 2020 at 4:48

There appears to be a slew of miscomprehension as to why mixing to -6dB is a good or bad idea. Let me throw in some quick persuasion as to why this is 'a false premise'.

1. If you mix to -6dB you immediately add 6dB of random noise to your noise floor. You've just thrown those extra bits away, wasted them for no reason.

Over-compressing the mix is going to annoy the mastering engineer no matter how many dBFS you leave him at the top.
You've already crushed the life out of it & he has nothing to work with. This is the same whether you hit -0.01dB or -18dB.

3. The Mastering engineer wants clean. He wants breathing room.
He wants headroom in the actual audio, not in the final level.
'Clean' is achieved by not doing 1, or 2. above.

4. If you were to save your mix at 32-bit float then minor considerations aside [rounding, floating point error I'm going to ignore just for this one purpose] then you could send your mix at +18dB & it wouldn't make the slightest difference.

5. We no longer in this digital era have a compulsory 'absolute' 0dB before the track is mastered for the consumer. We have floating point internally, we only go to fixed point at output. At that point what we do have these days, to combat the Loudness War, is we have LUFS. [I'm not going to explain LUFS, Google it :P]
LUFS, however, is held as 'the ultimate weapon' by Apple Music, Spotify, YouTube etc.
This is [as we're down to fixed decimal point sound at consumer level] bounded by our 0dBFS, but weighted to an average 'loudness', which these broadcasters will not let us exceed… if you do, they'll just turn your whole track down.
This bit is the mastering engineer's concern.

So, in conclusion…
Mix to -0.01 dBFS if you want, or leave a comfortable 0.3dB if it makes you happier. Your DAW metering will be able to spot any real overs.
Don't comp the sh*t out of it. This will irritate the mastering engineer & he will either do a bad job or send it back to be fixed.

• Thanks for the explanation you provided. I guess I should be more clear. Instead of mixing I should say leveling as I was not intending to say I'd compress anything (as I'd leave that to mastering) Jul 7, 2020 at 17:50
• What is mixing if it doesn't involve 'levelling'? You do need to comp unless you're doing classical or jazz. How you comp is all a part of the mix engineer's job. The mastering engineer is the guy with the golden ears who turns your mix into something fit to be released. He is not going to do the mix for you. That's your job. Jul 7, 2020 at 17:53
• Now you're getting the idea ;) You need to leave the mastering engineer breathing room. This not the same as wasting 6dB at the top. Jul 7, 2020 at 17:58
• We're now getting into the realms of gain-staging & going far beyond the remit of the original question. Just bear in mind that any mix operation needs a calculation. Every calculation makes the sound 'dirtier' by some infinitesimal degree. Your job is to reduce these as much as you can, whilst mainly just ignoring what goes on inside the DAW, which probably works at 64-bit floating point to reduce this calculation error so far as to make it negligible… you only need to worry at output. Jul 7, 2020 at 18:14
• I'm assuming the asker is a bit confused about these things, so I suggest making it clear what you mean by "mixing to -6 dB". I would normally guess that means mixing so that the average level of the mix is -6 dB, but that would be a very hot, compressed mix, so I wouldn't expect there to be much discussion about "mixing to -6 dB" other than "don't do that, target your average levels somewhere between -21 and -12 dB". So maybe you're talking about mixing so that the peaks are around -6 dB? Nothing wrong with that, but maintaining a crest factor is what you want in any case. Jul 8, 2020 at 4:44

Suppose that some of the tracks on an album spend most of their time at an amplitude which is 1dB below the peak amplitude, and others spend most of their time 5dB below the peak amplitude. If one were to start by mixing the first group of tracks with a peak amplitude of 0dB, there would be no way to avoid having the tracks in that group sound louder than tracks in the second group except by cutting the volume on them. If one were to instead start by mixing the first group of tracks with a main amplitude of -6dB, peaking at -5dB, then one could mix the second group with the same main amplitude, letting it peak at -1dB.

I don't think there's anything particularly special about -6dB as a value except that it's large enough to accommodate many scenarios but small enough not to cause too much needless loss of signal quality. Different genres of music benefit from different amounts of headroom, and how the dynamic range of the piece being mixed is expected to compare with that of other pieces that will be mixed.