Keep your mixes consistent by using a reference
Like most of you, I work a “day job” to pay the bills. I run my own CD mastering business by the name of Resonance Mastering. Unlike other mastering houses, I don’t believe in “corrective mastering.” If I hear a problem with a mix, I’ll ask the client to go back and fix it! Many clients routinely hire me for mix consultation, in which I offer detailed suggestions on each track in order to fine-tune their release before mastering. I run into the same problems again and again, so I find myself offering up the same solutions, which I’ll present here over the course of a few articles.
TIP 1: Periodically A/B your mix against the same reference mix in each of your song projects
It used to be common for an album to be recorded in several studios, yielding an inconsistent set of mixes. Some might be thin or bright while others are boomy or muddy. These days most albums are mixed in a band member’s home studio. You’d think that songs mixed with one set of ears in the same room would be consistent, but that’s not always the case. The best method I’ve found to keep my mixes on target is to periodically A/B them against the same reference mix. Find a song with the tonal balance you’re looking for, insert it onto its own track in your project, set its fader to match the volume of the song (IMPORTANT!), and then mute it. Every twenty minutes or so, solo it. If it sounds bright, that means your mix is too muddy. If it sounds bassy or dull, your mix is too bright. If you use the same reference for every song, you’ll produce a consistent set of mixes. I’ve been using the same reference mix for years: “True Love Wars” as covered by Neuropa on their Beyond Here and Now album (originally recorded by Erasure). It’s not the cleanest recording in the world, but it’s balanced and the instrumentation stays basically the same throughout the entire song.
[Listen to a clip of “True Love Wars” by Neuropa]
Since the ear’s frequency response varies according to volume, it’s important to always mix at the same level. If you mix too quietly, you’ll turn up the bass and hi-hat to compensate for the lack of boom/sizzle (which is exactly what the “loudness” button on your stereo receiver does). If you mix too loudly, your ears will tire and you’ll find yourself creeping up the volume and craving more high end. The sweet spot is 85 dB SPL, which is safe for up to 8 hours per day. Radio Shack offers two inexpensive Sound Level Meters. I use the $50 digital version (catalog number 33-2055), but you can save $5 by going with the analog one. Set the dial to 80, the response to slow, and the weighting to C, then hold it in the mix position as you adjust your monitor level.
None of this negates the need to revisit your mixes with fresh ears over several sessions. Even in a treated room with accurate monitors, comparing to a reference mix at a consistent volume level, I don’t trust my ears after a half hour or so. I’m always bouncing from mixing or mastering to e-mail or web surfing to give them a break.
Brian Hazard is a recording artist with fifteen years of experience promoting his seven Color Theory albums. His Passive Promotion blog emphasizes “set it and forget it” methods of music promotion. Brian is also the head mastering engineer and owner of Resonance Mastering in Huntington Beach, California.
Reader Comments (16)
Wow - that is really good advice! Im using GarageBand and I always find myself referencing with songs in iTunes (which is annoying)- but of course it makes more sense to just import a track.
Thanks for the advice
chris
music: acrashinc.com
blog: musicneutral.com
WOW, that's AWESOME advice. I'm excited to start doing this.
THANK you.
Jason Silver
That's great advice. I'm currently mixing a bunch of new songs so I'll be using your tip!
I'm the same as you, I don't trust myself after half an hour or so. I normally mix a song in short bursts over a whole week.
nice to see a more technical post here! i didnt think that was even allowed ;)
if i can add anything to this though, i think in order to A/B while you're working you need to send all your tracks to something like a "group channel" before it goes to the master channel. so group A becomes your new master channel, and you will A/B it against a normal audio track that holds your song.
i wouldn't recommend loading up itunes alongside your program in case your OS has altered the volumes of the different programs. for example what comes out of my cubase might go direct to my soundcard, but some programs might have the OS interfere and attenuate, like in windows the speaker icon in the tray.
More technical posts are wanted and welcome (and they are popular). Post away on MTT Open to start.
Thanks for the kind words! I'll definitely be posting the other three articles in the series, which are even more technical than this one.
There's no need to route your channels to a group, at least in Cubase. In the screenshot, the top track is my reference. It's muted. All I have to do to A/B is click the solo button. Another click mutes it again.
Great advice! I have tried to do this myself but I'm always running into the same problem. If I put something on the master fader (I'm using Pro Tools), it colors the reference mix. Is grouping the solution or maybe bypassing the plugs on the master fader when A/Bing? This really becomes a problem if you are limiting a demo or something and want to compare it to a similar style track.
Monty
Anything on the master bus that affects the overall sound is, by my definition at least, mastering. As you might suspect, I advise my clients to leave that to me. I personally never put anything on the master bus other than metering or analysis tools.
In the case you're describing, where you're just spitting out a quickie demo, I'd turn the volume of the reference track down by the amount you're limiting the master bus. Unless you're making your demo hotter than the reference, the reference track won't be touched by the limiter.
Better yet, nail down the mix before you insert the limiter. Unless you're pushing it way over the top, you shouldn't have to adjust your mix to compensate.
Very strong advice. I always a/b my mixes with a reference, but never thought about importing the reference track into the project. So i can't wait for the next articles and tips. Thanks a lot !
Cheers ;o)
Great tip! Can you offer advice on what to look for in a sound level meter device? I could get cheap devices for around 25 EUR here at ebay Germany, but they are very basic and I don't know if they will be up to the job.
Cheers, Chris
Wish I could help, but my experience is sadly limited to Radio Shack. Hopefully someone else can chime in with some info for you.
Actually I found an auction with the Radio Shack device for 30 EUR. I think I will just give it a shot.
Thanks for the article Brian. Seems to me that both pieces of advice are very sound. Using a fixed monitor gain, i.e. using a standard XdB SPL as a reference point for 0 VU means that you will always mix to a consistent level and avoid the perils of the "loudness wars". And referencing well mixed material frequently is like bringing everything back into focus.
There's a lot of other good advice on a site I highly recommend: Digital Domain. The link leadsto a page of excellent articles by Bob Katz. For reasons that will become clear if you read the article "Level Practices (Part 2) (Includes the K-System)" I'd go with 83dB SPL measured using pink noise at -20dBFS as a standard reference level for monitor loudness, reducing it to 77dB SPL for typical pop productions. (This reduction leads to a slight compression of the dynamic range which suits the material)
I thoroughly recommend this article by the way, and the Pt 1 piece that goes with it.
We discussed working at a fixed level in the comments at passivepromotion.com. To make a long story short, it's not for me. Each project requires a different amount of squashage, depending on the character of the mixes and the tastes of the client. Stuff with lots of low mids or sustained bass can't handle as much limiting as bright mixes with lots of short transients. As much as I admire the reasoning behind the K-System, I can't force it on my clients. With that in mind, I manually set the monitor gain for each project I master. For mixing, I try to keep the peaks under 3 dB while maintaining unity gain on the master bus, and adjust the monitor gain for each song.
Brian.
All very sound advice indeed. However, some of the comments above are interesting: I think it depends on the purpose of the mix how you go about your A/Bs. In the ideal world, we'd mix and master straight out of the software we recorded into - it's not true that you don't lose quality in the digital domain - you do.
Any recording package (Pro Tools, Cubase, Logic) uses its own massively complex and unique algorithms to mix together everything that's happening. The best mix you'll ever hear coming out of them is when you play 'on the fly'. The rendering process somehow 'muddies' things up a little. Having to render once is bad enough, and the process of exporting an unmastered mix, treating it, then rendering again takes away even more quality, especially if you've been working at 32 or 24 bit, then you dither down to 16bit, and moreover, resample down to 44.1Khz. These processes though, are necessary in the real world, as most mix engineers don't possess (what people call) good mastering skills, although arguably they should, as anyone producing music should aim to get as near as dammit to another professionally recorded track, straight off from the mix coming out of the desk, leaving as little as possible to do in the mastering.
This is especially true when you're A/Bing against already mastered recordings. So, with reference to what's been said, yeah, A/B against other artist's stuff, your favourite tracks, whatever, but remember to bypass any mastering effects on your L/R bus when you do so. For many people, their mix will never see a professional mastering suite, as it's just not necessary for the purpose, and moreover, as I've said, with some care and attention, you can achieve very professional and impressive mixes by paying very, very careful attention to your A/Bs. I'd add finally, that I A/B with a couple of tracks, no matter what the style - usually picking from a rock track, a dancey track, and some kind of crossover track. Some mixes I find useful are Hellagood by No Doubt, Empty at the End by Electric Soft Parade and (Hey You) What's That Sound by Les Rhythmes Digitales.
Anyway, I liked the post, and the replies, and I hope I've contributed something - however personal and idiosyncratic these things are!
thanx dude