6

In the instructions provided to me by the service I'm using to have my album mastered, the engineers include this advice:

Include any desired processing, such as compression and equalization, but be sure to avoid using any peak limiting, ‘normalizing’, clipping, or any tools used for approximated mastering, such as Finalizer.

The maximum level of your non-limited track should peak around -3dBfs. (If your track is already "bricked", it limits our options.)

I do use compressors on my master output track, and I have turned on the limiter function on those compressors for most of my tracks. I'm assuming I must turn "limiter" to off. Is there any sense in me having the compressor running for all the tracks combined at all? Is this something where the mastering engineers would be better off working with a "clean slate?"

And in particular, how can I be sure my track peaks around -3dBfs and is not "bricked?"

Should I expect my tracks to sound quieter when I remix them and send the files to be mastered, and expect them to be returned sounding louder than when I sent the files?

  • What is "bricked"? – Dekkadeci Jul 21 '18 at 1:05
  • 2
    @Dekkadeci I had never heard the term before, my understanding based on research I did prior to posting this question indicates that it refers to a mix where there is so much normalization that the audio file has little in the way of dynamics and hence the audio file looks like a rectangle, or "brick." – RaceYouAnytime Jul 21 '18 at 1:08
  • @RaceYouAnytime: It certainly fits with 'looks like a brick', but the origin is from brick-wall mastering, so maybe 'looks like a brick-wall' or 'hits you like a wall of bricks'? – AkselA Jul 21 '18 at 16:10
  • @AkselA interesting. As anyone who views my profile might note, I enjoy the study of etymology, so I appreciate the clarification! :) – RaceYouAnytime Jul 21 '18 at 16:16
4

The sentence preceding the use of that term explains exactly what you should do. Give them at least 3dBfs of headroom to work with. As for how to do that, I don't know what DAW you are using but it should keep track of your peaks—or the highest dBfs level that it hit—and remember them until another peak is louder. Look for a number above or around the meter for the master bus. Play through the entire song and make sure that number never goes higher than -3 (the clipping point is 0dBfs and everything below that is negative).

In other words "bricked" means that your peaks are hitting 0dBfs or as loud as they can be without clipping. If that's the case then the mastering engineer has no headroom to work with.

Is there any sense in me having the compressor running for all the tracks combined at all?

Sure, a common reason for mix bus compression is to give the mix a bit of "glue" or more of sense that the tracks are sitting well together. Or maybe you want to compress an entire section of a song at the mix level differently than another section. A couple of tips though:

  • Make sure you aren't trying to fix problems you're hearing on individual tracks on the mix bus. Put a compressor on the problem track itself instead.
  • Go easy on the ratio and threshold. You usually don't need much compression on the mix bus.
  • Put the compressor on the mix bus very early in the process and "mix into into it". If you put it on at the end, you're changing and possibly undoing some of the mixing work you just did. Whereas if you put it on early, you'll react to how the compressor is reacting in your mix decisions.
  • Don't abuse mix bus compression solely to make your mix louder. The mastering engineer can and will do that but they'll do it much better than you. Instead try to only use mix bus compression to control dynamics or add glue.

Is this something where the mastering engineers would be better off working with a "clean slate?"

A mastering engineer can and would fix any "glue" issues if it was particularly a problem. But it's fairly common to use a little bit of mix bus compression at the mixing stage. So don't be afraid of to do it, just take it easy and/or have a good reason.

  • This is a great explanation, thank you. I should have mentioned that I'm using Logic Pro X, but I think your answer explains exactly what I was wondering about. – RaceYouAnytime Jul 21 '18 at 2:47
  • 1
    I would say crest factor is more important and you should give at least 20 dB of crest factor!! – Todd Wilcox Jul 21 '18 at 4:30
3

I think the existing answer is good and doesn’t go far enough.

First, I would suggest you put zero plugins on your master output track or stereo bus or whatever you want to call it. For two reasons:

  1. It will help you make a better mix, since you’ll have to make things work without processing on the stereo bus.
  2. It will let your mastering engineer do what they do best.

Second, when you’re looking at your stereo bus levels, one important thing to pay attention to is crest factor. Crest factor is the difference in dB between the average (usually RMS) and peak levels. A recording is “bricked” when it has less than say 6dB for a crest factor, although many recordings you hear have between 3 and 6 dB for a crest factor. The worst examples have close to 0 dB. Feature films are usually delivered with a crest factor of about 20 dB, and that’s after they’ve been mastered. Many of the best sounding music recordings will have a crest factor of about 12 dB, but remember that’s what your mastering engineer is going to deliver back to you.

I strongly suggest not spending any time reducing the crest factor on your stereo bus to less than 20 dB. That means since your peaks are supposed to be no higher than -3 dBFS, your RMS levels would ideally be no higher than -23 dBFS. Check your DAW’s manual to find out how to configure your meters to show RMS levels. Most can show RMS and peak at the same time.

Critical note about delivering to streaming services

Most of the major streaming services, including YouTube, iTunes, and Spotify have reduced output levels and are recommended crest factors of at least 14 dB when you deliver your master to them. If you or your mastering engineer reduce your crest factor below 14 dB, it will actually sound quieter when streamed, not louder. You should ask your mastering engineer about that if you want to deliver to streaming services and of course your can’t create crest factor, only reduce it, so you definitely should send a crest factor of at the very least 14 dB to be mastered.

See: http://productionadvice.co.uk/spotify-reduced-loudness/

Also, one could argue this column by mastering engineer Bob Katz was the first shot fired in ending the loudness wars, and his K-system that he proposes here is now widely used:

http://www.aes.org/technical/documentDownloads.cfm?docID=65

  • This is excellent advice, thank you. I will take your advice of using zero plugins on the master output tracks. I was not aware of the concept of crest factor before either. – RaceYouAnytime Jul 21 '18 at 4:58
0

They gave you instructions. Follow them. 'Bricked' refers to 'brick wall compression', otherwise known as 'hard limiting'. The waveform amplitude has its peaks flattened, as if they've hit a brick wall.

Yes, most people think 'mastering' involves increasing the perceived loudness. Some think that's ALL it means! So mastering services do that.

-1

When you send stuff for mastering, the only processing you should do yourself is what involves the "signature sound" of a piece. Usually that is stuff that you would only do to one track in particular, processing that sort-of is a part of a single instrument or voice. Stuff you'd usually also add when using a live PA.

Making for a compact sound and distribution in general is the job of the mastering engineers. If you have particular wishes or suggestions, write them down rather than trying to anticipate them in your tracks.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.