Dont FearThe Phase, Phix it....Phast!

Phases of the moon. Going through a phase. And of course, the reason why we’re all here: phase in the studio. Can you really hear the difference if a signal is out of phase? Or can you do so only if it’s referenced to other signals? Besides, how do you check if your studio’s phase relationships are correct? Thought you’d never ask . . . but you did, so let’s GO.
Publish date:
Social count:

When most people refer to an “out-of-phase” condition, they really mean a reversal in polarity (i.e., positive-going waveforms go negative and vice-versa), which is independent of frequency. True phase shift can be frequency-dependent, where the amount of shift varies at different frequencies. But we’ll use the term “out of phase” because it has a common, understood meaning: Flipping a signal’s polarity.

Phase problems are, unfortunately, not uncommon. Balanced cables can be miswired, and some older gear had XLR connectors wired with pin 3 hot instead of pin 2 (the existing standard). Furthermore, some “vintage” effects weren’t careful about phase (e.g., having the delayed signal in a digital delay out-of-phase compared to the straight signal), and I’ve even tested some recently designed processors that flip polarity. The rationale is that if it doesn’t mix this out of phase signal with a dry signal, no one will hear the difference anyway, and adding another stage to correct the phase would add noise. But does it matter?

Try this experiment: Put a signal through a single mixer channel (hardware or virtual), then flip the channel’s phase switch. Hear a difference?

Maybe, maybe not. For example, with a kick drum, the first rush of air pushes out at you. If this signal goes through a properly phased system, the speaker will push air out to re-create the kick sound. But if the system flips phase, then the speaker will suck in to move the required amount of air. The result will still sound like a kick drum, but some people claim the one with the proper phase “feels” better, and that the positive effects of proper phase are cumulative with multiple tracks.

My take? I think people can hear a difference with some sounds; on the other hand, trying to detect a phase flip on a vocal seems impossible.

Some engineers make sure that absolute phase is preserved throughout the entire signal chain. You can test for this with phase meters, but a less expensive option is to fire up a two-track digital audio editor and probe around.

Figure 1 shows a “test setup” for checking phase. Split the input signal, and send it to the input of the device being tested and a digital audio editor’s left channel via its audio interface. This is your reference. Then, feed the output signal of the device (or chain of devices) being tested to the digital audio editor’s right channel. This setup will also show up if there’s a time delay between the two signals.

Use a test sound source with an asymmetrical waveform (e.g., kick drum), then record a bit of sound into the editor. Vintage guitar effects are notorious for phase problems, and are well worth testing (Figure 2). Also check all hardware mixer outs — master out, submaster out, monitor out, sends out, direct out, and so on. Sometimes you’ll uncover something like an out-of-phase headphone or aux out.

Also, test the entire input-to-speaker chain to make sure nothing’s amiss, and if you use XLR cables, check those to make sure they’re wired correctly.

Once you’ve identified any problems, fix them. Some fixes are easy: If a piece of gear’s balanced output is out of phase, just reverse the “hot” and “neutral” leads going to the out. With cables, mark them and repair them when you get a chance. For old effects boxes, send the output into a mixer channel, and flip the channel-phase to compensate. As to how that process works . . . keep reading.

While polarity issues are pretty cut-and-dried — either the polarity is correct, or it’s not — phase, as it can change at different frequencies, is a whole other matter.

The classic example of a phase difference occurs when using multiple mics on a source. For example, let’s assume you’re miking an acoustic guitar with two mics: One pointed toward the sound hole to emphasize the bass, and one capturing more of the overall sound. When you combine these in mono, some frequencies will reinforce each other, while others will cancel. That’s because the mics are a fixed distance from the instrument, so different audio wavelengths will hit the mics at different points along their curves.

Does mono really matter any more? Yes. We’re not just talking sounds coming over your TV; the only way you’re going to get true stereo is with headphones. Sound waves coming from two speakers will tend to mingle on the way to your ears in an acoustic space, thus “monoizing” the sound at least a little bit.

Getting back to our acoustic guitar example, as you set up your mics, periodically check the resulting sound in mono. You will likely find that certain mic positions will provide a more realistic sound than others when combined in mono. However, also try flipping a mixer’s phase switch (also called reverse, invert, or identified with a Ø symbol) before you change positions because that may give better results — and it’s a lot less hassle than moving mics.

Usually, one switch position will sound clearly better than the other. However, sometimes neither will be quite satisfactory. At that point, it’s time to experiment again with mic positioning.

When it’s time to flip phase with a hardware mixer, look for a switch in each channel strip to change phase, usually located near the preamp section. Software hosts handle phase switching in various ways. For example, with Pro Tools, you insert a channel gain plug-in that includes a phase switch. Steinberg Cubase SX’s mixer (Figure 3), which models a conventional hardware mixer, includes phase switches in both the mixer view and individual channel strips; Cakewalk Sonar also follows the hardware paradigm, including phase switches in the mixer view, inspector, and track view. Sony’s video-oriented Vegas nonetheless recognizes the phase switch’s importance by including one within each audio track.

Using a phase switch is a non-destructive process, so you can revert to the opposite setting at any time. But if you want to lock in a phase change, most hosts also allow for destructive, DSP-based changes. For example, in Apple Logic’s Sample Editor (Figure 4), an Invert process located under the Functions menu can invert the polarity of what’s being edited.

Finally, note that sometimes the point of flipping phase is cancellation. In one mix I encountered, the stereo drum and percussion tracks were premixed. However, the kick and snare were annoyingly loud compared to the delicate, and rather well recorded, percussion. Fortunately, the kick and snare were panned dead center, and the percussion was spread to the left and right of the stereo image. I patched the stereo track into two more mixer channels, flipped their phase, and brought up the channel levels just enough to provide a bit of center cancellation. Bingo — the kick and snare went down a few dB, which made the percussion seem more prominent.


And that pretty much does it for our discussion of phase. Some day, when you have a few minutes, take the time to boot up your digital audio editor of choice, and check out your system’s phase integrity — you never know exactly what kinds of ghosts are lurking in your machines. Then, once you know everything is properly phased, explore your mixer or software host for its phase-switching options. You may find they have more uses than you originally thought.