How to A/B Your Mix Against a Reference Without Lying to Yourself
The short answer
A reference comparison is only useful if it is instant and level matched. Dragging a reference file onto a track and unmuting it fails both tests, because the reference is mastered and louder, and the two second gap while you fumble for the mute is long enough for your ear memory to evaporate. Chris Shaw solves this with MCompare on the master bus, fed by a reference track inside the session and by a system audio aux routed back into Pro Tools over Dante, with source switching mapped to single keystrokes.
Why most reference comparisons are worthless
You know the routine. Client sends a reference. You drop the file on an audio track at the bottom of the session, solo it, listen, unsolo, listen to your mix, nod thoughtfully, and change nothing.
Two things went wrong before you even started.
The reference is louder. It is mastered. It is 6 to 10dB hotter than your mix, and louder always sounds better. Always. This is not a matter of discipline or golden ears, it is how hearing works. Unless you match levels, you are not comparing mixes, you are comparing volumes, and the reference wins every time.
The gap is too long. Auditory memory for timbre is brutally short. A couple of seconds of scrolling and clicking is enough to lose the detail you were trying to compare. By the time the other version starts, you are comparing your mix to a vague feeling about a mix.
Everything below exists to kill those two problems.
The setup
1. Give the reference a permanent home.
In Chris's mix template there is a dedicated track called reference mix. It is not wherever the file happened to land on import. It is always the same track, in the same place, routed the same way.
When a reference arrives he imports it, then uses a Soundflow script to move the audio from the imported track's playlist onto the reference mix track's playlist and delete the source track. One command. The reason it is scripted is that doing it manually means creating a new playlist, dragging audio between tracks, and cleaning up, which is exactly the sort of two minute job you will avoid doing and then not reference at all.
The same script handles the print track after a bounce, so the mix ends up where it belongs too.
2. Put MCompare on the master bus.
MCompare from MeldaProduction is free, with a paid upgrade at around thirty dollars. What makes it the right tool is that it can take sources from inside the session, so the reference track feeds an instance of MCompare sitting on the master bus, and level matched switching between your mix and the reference becomes one click.
Not solo. Not mute. One click, no gap, matched level.
3. Route your computer's audio back in.
This is the part most people have never considered and immediately want.
Chris feeds system audio, streaming, iTunes, whatever is playing, out over Dante and back into Pro Tools on an aux track called iTunes. That aux becomes another MCompare source.
Which means, mid mix, without leaving the session, he can flip between his current mix, the previous version, the client's reference, and any commercially released record in existence, all level matched and all on the same monitoring path.
If you do not have Dante, a virtual audio device such as Loopback, BlackHole or your interface's own loopback channel will do the same job.
4. Make switching cost nothing.
Chris maps MCompare source switching to MIDI triggers fired from his computer keyboard through Soundflow, with a notification in the corner of the screen telling him what he is listening to. The plugin window does not even need to be open.
That last detail matters more than it sounds. A comparison you can make with one finger while your eyes stay on the meters is a comparison you will actually make. Anything requiring a mouse and a window is a comparison you will skip at hour six, which is precisely when you need it most.
How to use references properly once the setup exists
The setup is half the job. Here is the other half.
Pick references before you start, not when you are lost. Choosing a reference at the point of panic means choosing something that flatters your problem rather than exposing it.
Use two or three, not one. One reference and you will slowly rebuild that record. Three tells you where the genre lives rather than where one mastering engineer went.
Compare specific things, not vibes. "Does mine sound as good" is not a question you can act on. "Is my snare further back than theirs," "is my low end wider," "is my vocal further forward in the chorus" all are.
Compare in short bursts. Two or three seconds of the same section, switching repeatedly. Long listens let your ears normalise and then everything sounds fine, which is your brain being polite rather than useful.
Remember your mix is not mastered. The reference has been through a mastering engineer. Yours has not. Expect it to be smaller, softer and less finished, and compare balance and tone rather than loudness and polish.
Then stop. At some point the reference stops being information and starts being someone else's record standing in the way of yours. Chris's whole template is built around getting to real decisions faster, not around auditing yourself into paralysis.
The point
Nobody is impressed by your mix at 3am. They are impressed by it in a car, next to a song they already love.
Setting up level matched, one keystroke reference switching takes about twenty minutes once, and then improves every mix you make for the rest of your career. That is roughly the best return on twenty minutes available in this job.
FAQ
How many reference tracks should I use when mixing? Two or three from the same genre. One reference tends to get copied rather than learned from.
Should reference tracks be level matched? Yes, always. Mastered references are significantly louder than an unmastered mix, and louder is perceived as better regardless of quality.
What is the best plugin for A/B mix comparison? MCompare from MeldaProduction is the most widely used, and the free version covers most needs. Metric AB and ADPTR Metric are common alternatives.
Where should the reference comparison plugin sit? At the very end of the master bus chain, after everything else, so you are comparing your finished signal path against the reference.
Can I reference against Spotify or Apple Music? Yes, by routing system audio back into your DAW using Dante, a virtual audio device or your interface's loopback function, then feeding it into your comparison plugin as a source.
Keep going
- Watch the full series: Chris Shaw: The Mix Template on Puremix
- Related: What Actually Goes in a Pro Tools Mix Template
- Related: How to Prep a Pro Tools Session in 5 Minutes
- Get the full library: Puremix Pro
