Phase & Polarity · Foundation Track · Free The Music

Module 2.5 · Foundation Track · How You Hear

Phase & Polarity

Why two identical signals can become silence — or hollow comb filtering — depending on how they meet.

Headphones recommended. The polarity-flip cancellation is most striking with focused listening.

Picture this: you've spent two hours setting up mics on a snare drum. Top mic up close to capture the crack, bottom mic underneath to capture the wires. You record a few hits, pull both faders up in your DAW, and… the snare sounds thin. Hollow. Like the body's been scooped out. You compare to the top mic alone — fuller. You add the bottom — thinner again. Confused, you click a small button on one of the channels labeled "polarity invert." Suddenly the snare is full, present, alive. What just happened?

Two scenarios: identical waves in phase reinforce each other; identical waves out of phase cancel each other IN PHASE OUT OF PHASE A B SUM A + B = +6 dB louder A B ↕ INVERTED SUM A + B = silence (cancellation)

Two identical waves can either reinforce each other (left) or cancel each other (right) depending on their polarity relationship.

What you saw above is the entire mechanism. Polarity is whether a wave is right-side-up or upside-down. When two identical signals are right-side-up together, every push of one matches a push of the other — they add up to twice the amplitude (+6 dB louder). When one is flipped upside-down, every push of one is met by a pull of the other — they cancel exactly, summing to silence. The bottom mic on a snare naturally captures the drum head's motion in opposite polarity to the top mic (the head is moving toward the bottom mic when it's moving away from the top, and vice versa). Without a polarity flip, the two mics partially cancel each other. With the flip, they reinforce.

That's polarity — a binary, instantaneous flip. Phase is more nuanced: when one signal arrives slightly later in time than the other (a few milliseconds, say from a mic placed further from the source), the result isn't full cancellation or full reinforcement — it's a frequency-dependent mix of both, called comb filtering. Some frequencies reinforce, others cancel, and the spectrum takes on a hollow, swooshy character. The widget below lets you hear both phenomena in your own ears — pure tone with polarity flip, then pink noise with delay. The lesson lands instantly.

Foundation · Module 2.5 · How You Hear

Phase & Polarity

Why two identical signals can become silence — or hollow comb filtering — depending on how they meet.

Combined waveform (time domain)
Frequency spectrum (comb visible on noise)
B Path B

Two identical signals, perfectly aligned

Press Play. Both paths are identical 1 kHz sine waves with no delay, no polarity flip — they reinforce each other. The combined output is exactly twice as loud as one path alone (+6 dB). Now click Flip Path B polarity — the two waves cancel completely. Silence. That's the polarity flip in its purest form. The two waves are physically opposite, and they sum to zero everywhere.

Master

Going deeper

Polarity vs. phase — the semantic distinction

Engineers often use the words "polarity" and "phase" interchangeably in casual speech, but they describe different things:

  • Polarity — binary, instantaneous, frequency-independent. Either a signal is right-side-up (positive polarity) or upside-down (inverted polarity). The polarity-flip button just multiplies the entire signal by −1. It's the same effect at every frequency, simultaneously.
  • Phase — continuous, time-related, frequency-dependent. Phase shift refers to where in a wave's cycle a signal sits at any given moment. A 1 ms delay creates 360° of phase shift on a 1 kHz tone (a full cycle), but 720° (two cycles) on a 2 kHz tone. So a single time delay creates different phase shifts at every frequency.

The practical consequence: polarity issues are fixable with a polarity flip (one click). Phase issues caused by time delay are not — flipping polarity on a delayed signal just shifts the comb filter's notch positions; it doesn't eliminate the comb. Time alignment fixes phase; polarity flip fixes polarity.

Where phase issues come from in real recordings

Phase relationships go wrong almost any time more than one mic captures the same source. Common culprits:

  • Snare top + bottom — the canonical case. The drum head moves toward the top mic and away from the bottom mic simultaneously, so the bottom mic's signal is naturally inverted relative to the top. Polarity flip is the fix (no time delay involved — they're equidistant from the head).
  • Kick in + kick out — inside-the-kick mic is just inside the shell; outside (sub-kick) mic is several feet away. Different distances = time delay = comb filtering even when polarity is correct. Solution: time-align the outside mic by a sample-delay equal to the distance/sound-speed.
  • Guitar amp + room mic — close mic 1 inch from cone, room mic 6 ft away. The 6 ft = 5 ms delay. Comb filtering. Time-align the room mic, or use it as a separate "wet" send rather than blending directly.
  • DI bass + amp bass — DI is electronically instant; amp signal travels through cabinet, mic, cable. ~5 ms delay creates comb filter. Time-align the DI by sample-delay matching the amp signal's lag.
  • Stereo mic pair (XY, ORTF, spaced) — both pick up the same source from slightly different angles. XY is co-incident (no time difference); ORTF has a controlled small offset; spaced pair has bigger offsets and is most prone to comb filtering when summed to mono. Each technique trades off stereo width vs. mono compatibility.
  • Drum overheads + close mics — the kick and snare appear in the overheads and their close mics, with the close mic getting the signal first. Without time alignment or careful polarity work, the kick and snare lose punch in the mix.

"Polarity is binary — either identical or opposite. Phase is a spectrum. Both matter; learning the difference between them is half the work." — FTM, on the polarity / phase distinction every working engineer eventually internalizes

The 3-to-1 rule

When using multiple mics, the classic guideline is the 3-to-1 rule: the distance between mics should be at least 3× the distance from each mic to the source. So if a vocal mic is 12 inches from the singer, a backing-vocal mic on someone behind them should be at least 36 inches from the lead vocal mic.

Why 3-to-1? Because at that ratio, the bleed of the second source into the first mic is about 9 dB quieter than the direct signal of the first source — quiet enough that comb filtering when the two are mixed becomes minimally audible. It's not a hard physical law; it's a practical guideline that prevents the worst phase-cancellation problems in multi-mic situations.

The fix: polarity flip vs. time alignment

Once you've identified a phase issue, two fixes are available:

  1. Polarity flip — one click, frequency-independent. Use when the issue is pure polarity (snare top/bottom, push/pull mics). Doesn't help with time-delay issues; in fact may make them worse by shifting comb filter positions to less-musical frequencies.
  2. Time alignment (sample delay) — slip the closer mic backward by an amount equal to the time difference between mics. Use when the issue is comb filtering from spaced mics (kick in + sub-kick out, DI + amp, room mic blends).

How to time-align: identify a transient (kick hit, snare crack) that appears in both signals. Zoom in on the audio waveforms. Slide one until the transient peaks line up. That's the proper alignment. Modern DAWs make this trivial:

  • Pro Tools has Slip mode and the Time Adjuster plugin (sample-accurate)
  • Logic Pro has the Sample Delay plugin (sample-accurate)
  • Ableton Live has Sample Delay (Audio Effect, 1-sample increments)
  • Reaper has clip-level offset adjustments and JS plugins for sample delay
  • FL Studio: use Patcher with a sample delay or shift the audio clip by samples

Specialized auto-alignment tools: Sound Radix Auto-Align (paid, automatic phase alignment for drum sessions) and iZotope RX (paid, includes phase issue diagnostics). Worth the investment if you record drums regularly.

Phase correlation meters — the diagnostic tool

A phase correlation meter shows the phase relationship between left and right channels of a stereo signal. The reading is a number between −1 and +1:

  • +1.0 — L and R are identical (phase-locked, mono signal essentially). Perfectly mono-compatible.
  • +0.5 to +1 — Mostly correlated. Normal for centered material.
  • 0 — L and R are independent (typical for natural stereo recordings). Mono-summing reduces level slightly but no cancellation.
  • −0.5 to 0 — Some out-of-phase content. Watch for thinning when summed to mono.
  • −1.0 — L and R are exact opposites. Perfect cancellation in mono — total silence. This is the danger zone.

Anything below 0 sustained for any length of time means something stereo is out of phase. Investigate. The meter doesn't tell you what's wrong, but it tells you that something is wrong, which is the first step.

In your DAW — polarity flips and correlation meters

DAWPolarity flipPhase correlation meter
GarageBand Limited — insert Gain plugin, check Phase Invert. Some channels may not have it. None stock. Use third-party (free Voxengo Correlometer).
Logic Pro Gain plugin → "Phase Invert Left" / "Phase Invert Right" checkboxes (independent for stereo channels) Correlation Meter stock plugin. Drop on master bus or any stereo channel.
Ableton Live Utility plugin → Phase L / Phase R buttons Stock Spectrum Audio Effect has correlation, or use Voxengo Correlometer.
Pro Tools Trim plugin → polarity invert button (Ø) PhaseScope stock metering plugin.
Reaper Right-click channel volume → "Polarity invert" (or channel inspector toggle) JS: Phase Correlator (stock JS plugin). Free, instant.
FL Studio Mixer channel → polarity invert button (Ø) Stock Wave Candy has correlation modes.

Free cross-DAW tool: Voxengo Correlometer (free) is a frequency-aware phase correlation meter — shows correlation by frequency band, so you can identify exactly which range has the phase issue (e.g., low-end is out of phase, mids are fine). Goes deeper than simple full-range correlation. Drop on any master bus.

Common mistakes

  • Confusing polarity and phase. Polarity flip doesn't fix time-delay phase issues. If the problem is "two mics at different distances," polarity won't help.
  • Not polarity-checking mic pairs. Should be a default move on any session with multi-mic'd sources. Even one click can transform the sound.
  • Trying to fix phase with EQ. EQ can mask phase issues (cut the cancelled frequencies' harmonics, etc.) but doesn't solve them. Fix phase at the source — polarity, time alignment, or by removing one of the mics if they can't be reconciled.
  • Polarity flip as a magic fix. Sometimes polarity flip makes things worse. Always trust your ears: whichever sounds fuller is the right answer for that situation, regardless of which is "technically correct."
  • Not checking mono. Phase issues on stereo content often only become audible when summed to mono. 50% of listening is mono. Always check.
  • Ignoring correlation meters. They reveal phase issues you can't hear yet. Drop one on the master bus permanently.
🎉 Module 2 Complete. You've learned how your ear works. The rest of the curriculum builds on this.

Next up · Module 3

EQ — Shaping Tone

Continue