If you have ever used a medium- to large-sized audio mixer, or an audio mixer in a DAW, you know the thing:
There’s a button, usually with the symbol like Ø. The manual usually tells you (in more or less detail) that it does something like “phase reversal”. You typically use it in situations where you have two mixer channels coming from the same audio source, but from different directions, e.g. miking a snare from top and bottom. Or to create your own M/S mix.
But what, exactly, does this button do? And why is it called “phase reversal”?
Samples from a few products
Let’s see what a few manufacturers have to say on the topic in their device manuals, in no specific order.
Starting with Soundcraft’s GB8 large format analogue live consoles from their current roundup, we get:
“Reverses the phase of the selected input” is not a very detailed description. A big gun with an almost identical description is SSL’s 9000J series:
For another example, here’s what Midas have to say in the manual for their M32 (digital) console.
Now this is strange: it starts with a rather wishy-washy explanation of the concept of phase of a periodic function, then it says to “reverse by 180°”. What is that supposed to mean?
From Steinberg’s Cubase version 9 DAW we get an even more confusing combo:
The statement in the reference guide sounds inconspicuous enough:
However, looking at the mentioned buttons both in the channel and channel equalizer view (on left and right, respectively), we get:
So while the button on the left looks like any other phase button on any console, the one in the EQ view says “Phase 180°”, just like the Midas manual did.
From Studer’s Vista X flagship line of digital consoles, we find:
This is interesting. Now we get again new terminology: “in phase” and “out of phase”, and, what’s more, actual equations: X = -A, which would mean that phase matrix in the diagram would simply multiply by -1, or invert polarity?
Mackie have always been good in writing manuals. In the documentation for their Onyx 80 series of analogue consoles, they have:
This is new. While the symbol stays the same, the word “phase” is completely removed from any text. Instead, we talk about polarity, which seems to be in line with what Studer show with their equation, but has no apparent relation to the Soundcraft and SSL, and clearly contradicts the Midas and Steinberg descriptions.
So what is the answer here?
Different kinds of phases: The Symbol Ø
Since the only common parameter in all of the examples above is the symbol which looks like from an Ikea product name, we will now follow that lead.
This symbol has, since the days of old, been used for what used to be called the phase in electrical alternating current (AC) systems, including polyphase systems like three-phase (sic!) electric power. As this term is somewhat ambiguous, today the term line (German: Außenleiter, formerly Phase) is used, and also results in the symbol L, which also is in line with the alternate name live wire. (Sadly, line would also lead to confusion in our audio engineering context).
Some of that terminology swapped over by the use of differential (also called symmetrical or balanced) signalling in audio applications – best known from (three-pin) XLR microphone connectors.
On the three pins on an XLR connector, you have:
- The Ground or Neutral pin, which is “always zero”,
- The Hot or Live pin, which carries your signal,
- The Cold or Return pin, which carries the polarity-inversed version of your signal.
That means that if your input signal is I, then you get for the N, L and R signals:
L = I
R = -I
N = 0
Putting those together at the input of your mixer, you normally get for your signal O
O = L-R = 2I
Or, if you swap both signals (or phases):
O’ = R – L = -2I = -O
In other words:
If you reverse both phases (the hot and cold lines from your XLR line) in your mixer, then the resulting signal ends up having reverse polarity.
And the Result
The Ø switch on our mixer or in our DAW reverses the polarity of our signal, in other words it multiplies the signal by -1. This doesn’t have anything to do with the phase of any periodic content in our signal.
So why the confusion?
I can only make assumptions on why things like in the Midas and Steinberg manuals (both well-respected companies, and I happen to use Steinberg DAWs exclusively for many years) happened.
Obviously, the phase terminology found its way into the products of old back when there wasn’t a small specialization niche, together with a specialized language, for every engineering sub-field in each sub-discipline: audio engineering people knew their stuff and language also from electrical power systems and vice versa, and so it was only logical to give the pins in a balanced audio transmission the same (now outdated) names than on the polyphase AC systems.
The official name changed, but audio engineering (with their pendant for vintage/we always did it like this thinking) stuck with the old name, which wasn’t a problem because the engineers knew what it meant.
Then digital audio and DAWs came along, and with them a breed of developers who had never seen a balanced signal (because they didn’t have them in their code). And as in most companies authors for tech documentation and user manuals are not Thomas Pynchon but the software engineers who are outperformed by their colleagues – we got this misunderstanding.