There's a lot that goes into mixing, and it can be difficult to make the right mixing decisions consistently. A/B testing can help us to improve our decision-making when mixing and quickly course-correct when we may not be on the right track.
What is A/B testing, and why is it important in mixing? A/B testing or “A/Bing” is the act of comparing one thing (thing A) against a second thing (thing B) to decide which performs better objectively. As we mix, it's vital to double-check our decisions and our mix as a whole with A/B testing to ensure that what we're doing is improving the mix.
In this article, we'll discuss A/B testing in more detail, its importance in mixing, how to go about A/B testing, and 5 top A/B tests to make habitual in your mixes.
What Is A/B Testing?
As mentioned, A/B testing tests “thing A” against “thing B” to help us decide which option is better. Hence the name.
The main point is that we're testing two versions of something in the mix to make objective decisions to improve the mix.
This could be as simple as turning a track on or off. It could be toggling between two different settings on a processor or effect. It could be switching a single parameter on a single processor/effect. The options are virtually limitless.
We can also A/B test our entire mix against our reference mixes to hear the differences in the overall mix aesthetic. This will help us to make the appropriate decisions to get our mix sounding closer to the reference.
I have a video in which I discuss reference mixes in more detail. You can check it out here:
For more information on reference mixes, check out my article Mixing: What Are Reference Mixes & Why Are They Important?
The Importance Of A/B Testing During Mixing
Our ears/hearing have a natural propensity to adjust to whatever we're hearing. This means that we can swiftly lose our objectivity while mixing or, at the very least, our original point of reference (what the audio sounded like originally).
This adaptability is partially why taking frequent breaks, utilizing multiple monitor systems, and comparing our mixes to reference mixes is common practice. It's also what makes A/Bing such a powerful habit to develop when mixing.
The longer we go about adjusting the parameters of processes in our mix, the more our ears will adjust, and the further the original sound will be in our memory. Sometimes, processing will sound great in the short term but can be easily pushed too far. A/Bing such processing against the original can show us that.
So, to ensure our EQ decisions are doing more good than harm, it's important to get in the habit of A/Bing our EQ inserts between “on” and “bypass”. Turning a plugin on and off will give us immediate feedback on whether the EQ settings make the mix better or worse.
For more information on inserts, check out my article Audio: What Are Inserts? (Mixing, Recording & More).
As mentioned, A/Bing against our references is also important if we're after the specific mix aesthetic of the chosen reference.
How To A/B Test In A Mix
The best part about A/B testing is that it's relatively simple to do.
For example, A/Bing whether a specific process is improving the mix or not only requires us to turn that processor on and off. This can be done by toggling the bypass function.
When A/B testing, opt to listen to each version for a good amount of time (10 seconds or so) to re-adjust your hearing to one option before switching to the other. It can be beneficial to remove our visual bias by closing our eyes and rapidly toggling the bypass so that we're no longer certain whether the process is on or off.
Speaking of our own biases, it's also important to level match our processes for A/B testing (and for gain staging more generally). We naturally carry a loudness bias, where we tend to prefer the louder of two options, even if the louder option isn't necessarily what's best for the mix. So take the time to level match the output of the processor so that the channel's signal is at the same perceived level with or without the processor engaged.
To learn more about gain staging, check out my article Mixing: What Is Gain Staging & Why Is It Important?
If an output level control is not available, it may be necessary to insert a separate gain plugin on the track (before or after the EQ) and use it to set the levels. A/Bing will then require us to toggle two plugins on/off.
Although A/Bing can be done in solo, it's typically best to A/B test in the context of the full mix.
I have a YouTube video on the importance of mixing in the context of the mix rather than in solo. Check it out by clicking here.
Always remember that mixing is the art of combining elements together into a cohesive piece, not making each individual element sound perfect on its own.
Beyond that, A/B testing is as simple as finding a specific parameter or option to toggle between two different versions.
It could be changing a compressor's ratio between 2:1 and 10:1. It could be turning a specific EQ band on and off. It could be muting a track altogether in the mix.
A/Bing The Reference Mix
Before moving on, I want to discuss a more difficult A/B test to pull off during mixing, but a common and important one nonetheless. That is A/Bing against a reference mix.
Generally speaking, a reference mix is a commercial release in the same or similar genre as the mix you're working on. This often means it'll be more polished than our mix and also louder than our mix (both issues for mastering).
We could decide to import the reference mix into its own track within our mix session. In this case, it's vital that we send the reference to a different stereo output than our mix bus to avoid processing the reference with our mix bus processing. Different DAWs and mixers will do this differently, so I'll leave it to you to figure that out!
Furthermore, to bring back our discussion on the loudness bias, it's also important to level match the reference to our mix. We can often do that pretty easily by adjusting the stereo track fader that hosts the reference track.
Once everything is set up, we can toggle between the mix bus (A) and whatever stereo output we've routed the reference track to (B).
We could decide to reference our mix outside our mix session. This can be done with a DAW and music playback software or in the analog realm. Ensure that both the mix and the reference are ultimately routed to the same playback device (speakers, monitors, headphones, etc.) and that the two are level-matched.
This option could work, although it will generally take up time as we switch programs to A/B. This is not as effective as instantaneous switching between the two as our ears will experience some amount of silence between the switching.
As the third option (and my personal favourite), we can opt to utilize a reference mix plugin in our digital audio workstation that can do the work for us in terms of level matching and ensuring the reference is as clean as possible.
My preferred plugin for A/Bing my reference mixes is Reference 2 by Mastering The Mix (link to check the price at Plugin Boutique).
It's super-easy to use and has a wide variety of features to help me get my mixes close to the aesthetic of my reference(s). I'd encourage you to check it out for yourself!
I talk about using reference mixes in much more detail in this video.
The 5 Best A/B Tests To Incorporate Into Your Mixes
Now that we understand A/B testing, its benefits in the mix, and how to do it, let's go through what I believe are the top 5 A/B tests for your mixes:
- A/B test against reference mixes
- A/B test polarity shifting for proper phase relationships
- A/B test plugins on and off/bypass
- A/B test tracks on/unmuted and off/muted
- A/B test process/effect parameters on and off
I'll get to writing on each of these tests shortly, but I also want to share that I have a video where I discuss each of these A/B tests. Check it out here:
Okay, let's get to the top 5 A/B tests I recommend making habitual in your mixes.
A/B Test Against Reference Mixes
We've already discussed how to A/B test against your reference mix(es).
It's one of the best methods to improve your mixes, as it's anything but subtle. For example, we're not A/Bing a single EQ band of a single track in the mix. Rather, we're comparing our entire mix to a completely different mix that we would like to emulate in some fashion.
Doing so periodically will help to keep us on track throughout our mixing process.
For example, we can trend toward an overly bright or dark mix over time without necessarily hearing our mix clearly. Toggling to our reference after a while of mixing can re-adjust our hearing and show us objectively how our mix compares.
Make adjustments as necessary to reach your mixing goals relative to the reference mix.
A/B Test Polarity Shifting For Proper Phase Relationships
Phase relationships are critical to get right in the mix, particularly in the low-end.
I even have a video dedicated to the topic that you can check out here.
Phase issues are particularly important to thresh out in multi-miked recordings within the same acoustic space. Drum kits are a primary example, so let's consider a drum kit for explanation's sake.
Let's assume the following mics:
- Kick drum outside
- Kick drum inside
- Kick drum beater
- Snare top
- Snare bottom
- Tom 1
- Tom 2
- Overhead left
- Overhead right
Each mic will be primarily tasked with capturing the drum it's closest to. However, each mic will also pick up bleed from the other drums and cymbals.
The most common phase flip we'll have here is to flip the polarity of the bottom snare mic to align it better with the top. Since these two mics are pointed toward each other, phase issues are likely.
By that same token, we'll likely want to flip the kick beater mic so that it better aligns with the opposite-facing kick inside and outside mics.
However, all the other mics will be slightly out of phase when capturing any specific drum. While small timing shifts can be useful for aligning the phase (shifting the overheads to the snare top transient is common, for example), we may be able to get a great-sounding kit overall by hitting the polarity flip options on a specific combination of tracks.
Try this out for yourself the next time you're mixing acoustic drums. You may be surprised by the results.
For more information on phase issues, check out the following My New Microphone articles:
• Mixing: What Is Phase & Why Is It So Important To Get Right?
• Microphone Polarity & Phase: How They Affect Mic Signals
• Audio: What Are The Differences Between Polarity & Phase?
A/B Test Plugins On And Off/Bypass
This one's pretty straightforward. Turning plugins or other processes on and off will give us an idea of whether they're improving the mix or not.
If not, spend some time adjusting the parameters to best suit the mix. Sometimes the processor will work perfectly with a little tweaking.
On the other hand, we shouldn't be afraid of scrapping the process/effect altogether if we can't get it to work in the context of the mix.
A/B Test Tracks On/Unmuted And Off/Muted
This one might ruffle some feathers, but it may be the case that certain tracks are detracting from the mix and that no matter how much we process them to fit into the mix, we can't get them sounding as we want.
By A/Bing certain problematic tracks on and off, we may stumble upon massive improvements to the mix we should consider pursuing.
For example, quadrupled guitars panned both hard left and hard right might sound massive in theory, but in practice, cutting two guitars per side (so we're left with doubled guitars hard left and hard right) can clean up the overall guitar sound and actually increase the presence and aggression of the guitars in the mix.
A/B Test Process/Effect Parameters On And Off
This one is the most subtle of the bunch but worth considering in some cases.
Our effects and processes will have certain parameters for us to hone in the effect/process on the audio.
We can toggle certain parameters on and off. Other times we can toggle between two different settings for a given parameter.
At the other extreme, we can toggle between two completely different settings within the same effect/process, whether we set them ourselves or utilize presets.