Preparing for Mastering: A Guide For The Mix Engineer
- Should I Compress My Mixes Before Mastering?
- How Loud Should My Mixes Be If I Am Sending Them Out To Be Mastered?
- How Can I Avoid Distortion In The Final Master? (also, a link to a thread in a mastering web board which discusses all these issues)
Should I Compress My Mixes Before Mastering?
As a basic guideline, individual tracks should be compressed where needed and the 2 bus or stereo bus should not be compressed, or just lightly compressed.
Of course this very general advice and is aimed at the engineer who hasn't had a lot of experience or hasn't had their mixes mastered at a professional mastering facility. Sometimes mixes compressed on the 2 bus work very well but have to be done carefully. An experienced engineer who has had their tracks mastered many times and those who have attended a mastering session (and who has learned from the mastering engineer's feedback) will likely have a good idea of what will help and what will make it more difficult.
Keep in mind that an experienced mastering engineer is the "king of stereo bus compression". In other words, he knows why he is using compression, what the results will be and how much and what type to use on various styles of music.
At Silverbirch, we have 5 class-A outboard compressors at hand, ready to insert and solve just about any problem, to use just for the great sound of a tube unit or to help make your tracks really loud, if they stylistically should be made really loud.
Limiting mixes is another common practice that can cause problems at the mastering studio. For the majority of music genres these days, we will have to limit the final master anyway so if it is done in advance to the mix, the dynamics are lost forever and the choices of how to make a track loud and which of our limiters would do the best job is partially taken out of our hands.
When I receive tracks to master that are too compressed and/or limited, I try to gently point out that whatever bus compressor plug-in my client used will not compare in sound quality to our out-board compressors and that next time if they don't use it, or don't use it so much, I will be able to make better use of my compressors and the results will sound better.
Just to reiterate, I am referring to buss compression, not individual track compression. The latter is definititely the job of the mix engineer! It's the same with limiters. If you want to use one too catch the odd peak, that's OK. But even better, don't hit the stereo buss hard enough to even need one!
This may surprise some people but mastering is just as much about limiting as compression. In many cases, the amount of compression is very little with more emphasis on the limiting.
And that brings us to the next topic which is very related.
How Loud Should My Mixes Be If I Am Sending Them Out To Be Mastered?
The answer is not very loud! If your tracks are as loud and as bright as a major label CD, you are making it much more difficult for the mastering engineer to do his best work. When referencing to major label CDs while mixing, do not try to match the levels. Use them as examples of balance, relative equalization/ individual track compression, stereo field depth and width but make sure that there is a little room to move for the mastering engineer, to help us balance the album. So, not too bright or overly bass heavy, maybe slightly "underdone" would describe it. The operative words here are "leave some headroom for us to work".
If you have used an eq plug-in to match the brightness of a mastered CD, I will have to lay off using my very expensive analog units on the top end. When this happens at the mastering session, I say to the client, wouldn't you rather have the mix subtly brightened with a Manley tube or Prism solid state equalizer instead of a plug-in? Then they get it! Same goes for buss eq.
If your mixes are consistently hitting "0" on your digital meters, it is likely too loud. I prefer to see peaks at "- 2.0 dB" or even lower, but not much lower than - 4.0 dB. RMS levels for mixes are typically -15 dB to 20 dB whereas mastered pop/rock/R&B/Hip Hop tracks are currently testing out at an average of -9 to 12 dB and even lower.
If your stereo mix looks like a 2 X 4, it is either too compressed or mixed too hot to the 2 bus. By the way, a 2 x 4 is a long piece of lumber, flat on all sides! When viewing the average mix that is properly prepared for mastering, you should see some space, some peaks and valleys which indicate that there are still some dynamics left. Digitally capture and examine a recently mastered major label pop master and you will instantly see what not to do!
As a guideline to how loud your overall volume should be, your mixes should be about 6 to 12 dB lower than a major label CD.
The second way mixes are compromised for mastering are by overloading the 2 bus. Just turn down all your tracks equally until there is some decent movement in the 2 bus meters and the peaks are rarely hitting "0", or even lower, as described above. The reason for this is that most workstations do not sound good when the bus is "crowded" at the top. Turning down the 2 bus is not at all the same as turning down the tracks before they reach the bus. Before running a mix through our analog gear, we will turn down the mix in order to avoid distortion. When we turn the mix down, it is usually less invasive that you taking the master fader and turning it down. In the digital world, the algorithm/quality of the device used to make a simple gain change can make your master sound better or worse!
Inexperienced engineers using Pro Tools sometimes fail to add a Stereo Master. The individual tracks may not be clipping, but the sum total of the output of all the tracks is causing the final stereo mix to overload and clip. But since they can't see it, they don't realize it. Then when asked to bring the levels down, they take the stereo mix and lower it a few dB. As stated above, this is not helpful as we can turn the file down with better gear before mastering.
Another area to examine is plug-in levels. Even though you may not be hitting the red in your track plug-ins, some of them do tend to sound grainy, saturated or slightly distorted before the meter shows any problems. If you are not looking for that affect, solo the tracks and listen carefully. See if you can detect any slight problems. Remember that these will add up track by track and may sum into an unpleasant effect. The same goes for any buss plug-ins you are using. Remember that if using buss plug-ins, they should be used because you like the sound and never for volume.
Here is an email from a mastering client who actually went to the trouble of remixing many of his songs using the above guidelines. It also contains some observations that may beapplicable to Cubase users.
From Mike Whitla of Rainbow Songs Inc 11/21/04
Thanks for your advice about how to mix down my songs for mastering.
You told me to it was best to have the highest peaks of each mix to be about at -4dB to -2dB. This was so that you have headroom to be able to tweak things at your end by using your compressors and eqs to boost the signal rather than the plug-ins in my software to do this.
I really heard this in action through the mastering process as you were able to bring up the things that needed to be brought out and enhanced as opposed to having to do corrective frequency cutting through eq and compression of undesirable sound. You were able to tighten up the bass and give some sparkle to the guitars that would not have been possible if the mixes were too hot.
This advice was also good for another reason you might not have been aware of. When I was mixing down from Cubase I was noticing that what I was hearing the program play (multi track) was different than what I was hearing in the exported files (stereo track). The problem was I was mixing too loud and when I was exporting Cubase was doing some compression of the files which affected my mix in all kinds of ways. The balance between instruments was off and there was something missing in the overall sound.
I went through and lowered each track but kept my master faders the same. This brought the level down so the peaks were at about -3db. And like magic the exported files are sounding exactly as they are supposed to.
Since the summing algorithms in digital workstations are in some cases still the weakest part of the software, an even better way to take care of this problem is to use an outboard analog summing device. When you make your next studio purchase, you may wish to consider one of these so check out the Dangerous Music web site for more info.
An article in the Dec./04 issue of Mix magazine also gives a great overview on the subject and lists a total of 8 companies that are now making summing units so this does seem to be an important trend in the studio world. Our mixing department bought the Dangerous 2-bus LT in December/04 and the engineers highly recommend it. By the way, if you are using a good quality analog board or summing unit, you can print your mixes back to the digital bus closer to digital zero. Once again, this is because you are circumventing the digital summing problem by summing in the analog domain.
Sometimes I think that engineers are mixing too loud because the volume is closer to a commercial CD and they feel that this will please their clients or make it easier for them to evaluate the mixes. If you are a commercial studio and are giving mixes to clients to check, you might consider using an inexpensive mastering program to lightly pump things up but do explain what you are doing and that it will sound better when it goes to the mastering studio. Of course your temp mastering job would be a separate file and you would leave the original mix untouched to send to mastering. Either that or just educate your clients by telling them that the first step of evaluating their mix is to first locate their volume control! That way, they won't be shocked when they hear the mix compared to a major label CD.
How Can I Avoid Distortion In The Final Master?
The obvious answer is make sure there is no distortion in your mix.
Of course following our earlier suggestions goes a long way to addressing overall distortion but there is a more subtle problem that can happen.
We receive mixes that have one or more momentary overloads in them and can sometimes cause us to have a great deal of difficulty in controlling them. They are caused by sharp transients in an instrument or sometime its a combination of instruments all hitting at the same time. Please understand that this doesn't have to be when a peak is hitting digital 0. It can happen on an acoustic piano recording and it can happen on a gentle Latin album. It also happen on vocal transients that are mixed way above the music.
Female vocals in "singer/songwriter" mixes seem to be the most difficult to control. These can sometime be heard as "bright bursts" of energy that are slightly louder than the rest of the relevant track. As soon as we run the track through the tube gear and raise the mix up to the average commercial CD volume level, those spots can turn in to audible distortion. We do have workarounds but they take time.
The problem for the mix engineer is detecting these subtle spots and cleaning them up before going to mastering. Why do they go by undetected? Probably the main reason is the lack of the accurate monitoring systems and taking the time to listen on several different speakers at different volume levels. Some distortion doesn't show up on really good speakers but does on cheap ones. The reverse is also true! Also, some distortions only show up when the music is played really loudly. Conversely, some show up only when the music is played very, very quietly. Investing in a very good set of headphones and listening to the material through a good headphone amp is a really good way to check for all kinds of subtle mix problems from noise to clicks, pops and distortion.
So, please take the time give your mixes a real quality control check, on different speakers, headphones and at different volume levels.
As a sidebar of interest, another reason these distortions get by the mix engineer is because they haven't been trained to hear these sort of problems. This doesn't necessarily mean the engineer is a poor mixer, it just means this particular part of the ear training process is not quite there yet!
In this case, we can only hope that the artist/producer, etc., will also give the mixes these same vigorous test. Many times people other than the engineer will hear problems and once it is pointed out, the engineer can usually hear it and fix it.
And lastly, please click here if you want to read a mastering web board thread about the problems that mastering engineers have to deal with when receiving "overdone" mixes to master. Check it out. It's very interesting, humourous and mostly informative, at least after you get past the first few posts!
....to be continued!