The First Tip: Read all the articles on this web site!
Here’s one of the secrets of the mixing engineers
To avoid squashing, if it doesn’t sound loud enough to your ears, turn up the monitor! If you find that you’ve been forced to apply limiting or compression just to keep the meters from overloading, then you’ve been going about this backwards. Instead, turn down your individual mix levels several dB, then get rid of any compression you were using to “protect” the 2-mix. Now your mix is at a lowered meter level, so turn up your monitor gain to arrive at the same loudness–only this time it won’t sound squashed. Leave the monitor at that position as you continue to mix (mark it so you can get back to it).
In 24-bit recording you can make a perfectly good mix that peaks between -3 and -10 dBFS with no loss of quality, in fact, with improved quality. So if the mix gets too loud by your ears, then turn down the elements that are too hot in the mix instead of turning down the monitor again, with no fear of mixing “too low”. In other words, a high monitor gain gives you less temptation to overcompress. High monitor gain does not necessarily mean high monitor output from the speakers–it means that the mix level had to be lower. For example, visit the CD Honor Roll and check out the great-sounding Lyle Lovett selection, which is close to the dynamics of a raw mix. Notice that in order to listen to it, you have to turn up your monitor gain. That’s approximately where your monitor control for a dynamic raw mix should be sitting (within 4-6 dB) before mastering. Obviously, a lot of today’s hypercompressed masters would require turning down the monitor, but we’re trying to show you how not to ruin the record in the mix stage (and hopefully not in the mastering, either!).
Know thy monitors
But even when you do, never be fooled. Take your mixes around and listen to them on several other systems that you know; then go back into the control room and if they do not translate, try to adjust your mixes in the areas where they do not translate. HOWEVER, be aware of the extremes. If it sounds reasonably good in a car, for example, don’t be tempted to turn up the highs for the car or it will screech (horribly) EVERYWHERE else. First of all, in the mastering we have much more experience in knowing how far to go and make sure that a recording is not made bass-shy just because it sounds boomy in a naturally-boomy car, for example.
Always mix to the highest possible wordlength
Even if the source tracks are 16-bit! Do not sample rate convert. When you’re ready to bounce or prepare files, please see our guidelines page for suggestions on making file names and file types.
Track important instruments in stereo
In the days of 8 track you had to be very careful about allocating tracks. But those days are gone. You have enough tracks to splurge now! So there’s no reason to conserve on tracks during the tracking stage. The stereo image and depth of your final product will be determined by your skill in mixdown at using delays, reverberation, effects, and your skills in tracking, how you tracked your instruments. Try to make a plan beforehand of how your soundstage might look, where the instruments might be placed. Realize that it probably will not hurt, and probably will help to record your important instruments in stereo.
For example, even a pair of bongos that are destined to be on the right side of the soundstage will sound better if one bongo mike is panned full right and the other somewhat right of center. This is because the ear decodes the natural space and delays picked up by those microphones, actually enhancing their definition in the mix (if the room acoustics are good).
Another example: Electric guitar. Capture the direct to one track. Capture the output of the loudspeaker with a close mike to another track. Capture the medium distant sound of the speaker bouncing from the walls of the room with another mike. Listen to the combination of these sources panned to different places, and also listen in mono to make sure you have not created phase cancellations. By using stereo miking and natural room acoustics in the tracking, and possibly artificial delays and good stereophonic reverberation in the mixing, your mix will sound richer and deeper. Not everything should be tracked in stereo, but don’t skimp on elements that will increase the depth and space of your recording. Of course you will need a foreground, middleground, and background in the mix, but it’s a lot harder to create a location and space for an instrument if you had only recorded it in mono.
In the mixing, use artificial reverberators that enhance depth and space and do not sound flat, plastic or “cheesy.” Use artificial delays to locate instruments in space, not just simple panners.
Try to not exceed -3 dBFS peak on a peak meter on the highest peak of the mix. Low levels are perfectly acceptable in a 24 bit system. Once you see that the highest peak is in the range of, say, -10 dBFS to -3 dBFS, then from that point on, if you can hear it, the low level passages are ok. Preserve dynamic range! Assume that if anyone is going to ruin the master, let it be me (the mastering engineer). If the mix sounds good, then soft passages automatically are NOT too soft. Of course, if you think a soft passage sounds too soft in the mixing, then of course try to fix that during the mixing. But these can easily be dealt with and often more efficiently in mastering, as we have the context of the album in mind.
If you have a VU meter, use it. With a sine wave, adjust it so 1 kHz, 0 VU is equal to -20 dBFS on the peak meter. Use the VU, ignore the peak, and you’ll start making better mixes.
Original sources, please
If at all possible, deliver a generation that is as close to the original as possible. If it’s on CD ROM, then cut a CDR directly from your hard disc files. Speed of cutting? Try to use Taiyo Yuden or other reputable blanks, and cut at 4X to 8X speed. These will PROBABLY produce the best results. Murphy’s Law: Allow for Murphy. Do not ASSUME that all the files will transfer successfully over here and that the CD-ROMs you have cut are perfect. Allow for the possibility that on the very last minute of the very last hour of the very last day, we may have to go to a backup CD-ROM, or you may have to cut another, because of some error or other problem in the transfer. Do not paint yourself into a corner. Make backups. Do not destroy or erase any source hard disc at the origination studio until the mastering has been completed.
Vocal Up/Down or (Even Better) Gang of Four
A) Make a lead vocal up (1/2 or 1 dB, you be the best judge) version. Do it NOW before you forget. It’s a lot easier to do it NOW than to discover in the mastering that you should have. Occasionally do a vocal down (1/2 or 1 dB) version if you think it may be useful; then again, it only takes 3 minutes to do a vocal alt version when you’re in the heat of mixing, but it takes forever to try to fix it in the mastering if you forgot.
B) Gang of Four – Stems or Splits
Professional mixing engineers never get caught with their pants down when they produce the Gang of Four. This is even better than Vocal Up/Vocal Down and is not any more work.
Here’s the key: You produce up to four synced stems. In a time pinch, you can produce only the first two or three.
Stem #1. TV (that’s instrumental plus chorus or background vocals).
Stem #2. Lead Vocal(s) (plus its reverb of course —basically muting everything else).
Stem #3. Full Mix (that’s what I will use unless there is a problem, and it’s also a reference to prove that #1 and #2 were made correctly).
Stem #4. Instrumental (by adding this to #1 we can reduce the chorus level. By subtracting this from #1 we can increase the chorus level. By subtracting this from #3 we can increase lead and chorus. And so on!
Synchronized stems are produced by running a separate mix pass from the same starting sample each time without changing any gains. You mute the tracks that you don’t want to hear. That way any reverbs or other processing which were applied on the full mix remain on the stems. Don’t be afraid if the vocal-only version has 1 minute of blank at the head, that’s part of the design!
In mastering, the sum of Stem #1 and Stem #2 at unity gain = Full mix. If we want lead vocal down, we just take the level of Stem #2 down a hair. And through other combinations we can control instruments or all vocals. In mastering, if a vocal is sibilant, we can apply a de-esser just to the vocal track, which is less of a compromise than de-essing the full mix. If a bass instrument needs to come up, we can equalize the instruments without making the vocal any bassier. And so on.
Many mix engineers argue (correctly) that if they are using bus compression the stems will not reflect the same sound they got on the full mix. This is true, if you are doing strong bus compression, the interaction between the peaks of the mix and the individual elements will not be the same when using stems. In that case, a legitimate gang of four cannot be produced!
The gang of four protects you and your clients in many ways. It gives you archive options and alternative options. It gives you the TV mix the client forgot to ask for but requests six months later! It allows cleaning up dirty words without dropping the music out. Develop the discipline to do the gang of four. You won’t be sorry!
When, Why, and How to Make Stems
I’ve definitely reached the conclusion that the less compromise you can make in the mastering process, the better the result. Mixing is mixing and mastering is mastering, and you should make the very best, finished mix that you can before sending it to mastering. Stems in mastering are not meant to be a substitute for a bad mix, nor a “mix fixer”, but they can be used in the mastering process to polish or aid a very good mix with less compromise than a full mix. The case of vocal up/down is a clear one where the mix is otherwise preserved, but the mastering engineer can tweak the lead vocal if the mastering processing seems to affect the vocal/instrument relationship, or the producer thinks twice about the vocal level, or some word somewhere needed to be punched but someone forgot. The mastering engineer is not a closet or frustrated mix engineer, we treat the stems as safeties or convenient means to an end, not as infinite “remix opportunitites”.
In more severe cases, you may have produced an otherwise great mix, but you had a little trouble with your monitoring and produced too little bass instrument, too much kick, and the lower midrange is a little bit muddy. This is a potentially bad (not lethal) combination for mastering and if the client has time, I recommend a remix. But in situations when time is tight, I have asked the client for stems, and the results have ALWAYS been better than if I had mastered from the combined two-track.
In mastering, the question comes whether I should first remix the stems in a remix session, then master—–or remix and master in one step. The more stems, the more it becomes a true remix, and the more inclined I would be to put on my mixing hat and mix first, then master. But with only 3 to 6 stereo stems, I find that I can get excellent results mastering with “supplementary mixing” going on. I’m mostly concentrating on mastering, using the stems to aid in the goals of mastering without compromise, such as using the vocal stem when de-essing is needed instead of de-essing and compromising the entire mix. Or equalizing the bass instrument instead of using overall bass frequency eq on a full mix, which has less compromise on the vocal, keyboards, etc. Of course, this kind of precision is usually not needed, and I master from the full mix more than 90% of the time. For example, the mastering processing is going to affect the clarity of the midrange and through “slop” will probably leak down into the bass region, hopefully for the better. But in the case of this lopsided mix I just cited, the mastering processing could easily make one range better while making the other worse.
Is this idea of mastering from stems heresy? It’s certainly a dangerous technique if placed in the wrong hands. You can end up with a less than ideal mix or less than ideal master if the mastering engineer does not think holistically. But if placed in the hands of an experienced mastering engineer, I think mastering from stems can produce a subtly better product or do it more efficiently. Producing stems the first time in the mix session reduces the number of calls for a costly and inconvenient remix! An ounce of prevention is worth a pound of cure.
Click here to see what other engineers had to say on the subject of Compression in Mixing.
Share this Article