How to read correlation & phase meters (without guessing)
Your DAW shows a correlation needle dancing between +1 and −1. Headphones still sound huge. Flip to mono and the bass vanishes. The meter was warning you — if you knew what the numbers meant. This guide is the plain-English map from meter reading to an honest fix order.
Download PHIXER free PHIXER hub
PHIXER is free stereo phase-repair (VST3/CLAP/AU · Win/Mac/Linux) from Riverflux Audio. Frequency-aware phase rotation for leftover L/R fights — not a polarity button, not a delay compensator, not a widener. Use meters to name the problem; use PHIXER only after polarity and timing are honest.
What a correlation meter is measuring
A stereo correlation meter (sometimes labeled “phase meter”) estimates how similar Left and Right are over a short window. It does not tell you “good mix” or “bad mix.” It reports similarity of the two channels — which strongly predicts what happens when you sum to mono.
- +1 — L and R are nearly identical (highly correlated). Mono sum is strong; stereo image is narrow or centered.
- Around 0 — L and R share little correlation. Wide / diffuse stereo. Mono may thin some bands; not automatically “broken.”
- Toward −1 — L and R are largely opposite (anti-correlated). Classic mono-killer territory: sum cancels energy hard.
Different meters average differently (fast vs slow, weighted bands, goniometer + number). Treat the number as a hint, then confirm with your ears on a real mono fold — workflow: mono-check before/after.
Correlation vs “phase” vs polarity vs timing
People overload the word phase. Separate the tools:
| Thing | What it is | Typical first fix |
|---|---|---|
| Polarity (Ø) | Whole channel flipped 180° | Invert one mic / layer |
| Time offset | Mic distance / track delay / Haas | Nudge timing / time-align |
| Spectral phase fight | Different bands cancel differently L vs R | PHIXER after Ø + timing |
| Creative width | Intentional out-of-phase energy | Own the mono tradeoff |
Correlation meters react to all of the above. A −0.8 reading might be a flipped OH mic, a 2 ms cab delay, a Haas widener, or a stacked sample. The meter does not pick the tool — you do. See polarity invert vs phase rotation and phase repair vs time-align.
Vectorscope / goniometer: the picture version
A vectorscope (Lissajous / goniometer) plots L vs R as a 2D cloud:
- Vertical line — mono / highly correlated (energy on the mid).
- Fat vertical oval — healthy stereo with shared mid energy.
- Wide horizontal smear — lots of side / anti-correlated energy. Mono fold will thin or cancel.
- Diagonal slash the “wrong” way — often a polarity issue on one channel.
Use the picture to see width shape; use correlation numbers for a quick trend; use mono fold to decide if the tradeoff is acceptable. None of these replace listening on phone speakers and a true mono bus.
What “good” numbers look like in practice
There is no sacred target for every genre. Rough studio heuristics:
Most of the mix hovering +0.3 to +1
Common for pop / rock / podcasts with a strong mono-compatible center. Side energy exists but mid is solid.
Dips toward 0 on wide sections
Pads, stereo FX, ambient beds — fine if low end stays correlated and mono fold still has body.
Sustained negatives on the full mix or bass bus
Investigate. Bass that lives out of phase rarely survives phones, club mono modes, or vinyl cutting. Trace which bus/group causes the dive.
Intentional anti-phase FX
Some wideners, stereoizers, and creative FX purposely push correlation negative. That can be the sound — just don’t ship surprise mono collapse on the kick/sub. See PHIXER vs stereo imagers.
Why headphones lie and meters still help
Headphones deliver L and R to separate ears. Anti-correlated energy still “sounds wide.” Mono sum on a single speaker does the physics homework the meter was approximating. So: meters for early warning, mono fold for the verdict, phones/club for the audience. Related mono physics: why bass disappears in mono.
Honest order when the meter looks scary
1. Identify the bus that dives
Solo groups with the meter visible. Find whether OH drums, stereo synth, vocal doubles, or a master widener is the culprit. Don’t “fix the master” first.
2. Polarity tests
Flip Ø on one mic / one layer. If correlation jumps toward +1 and mono body returns, stop. Done.
3. Timing / distance
Nudge delayed tracks, check buffer / PDC, align multi-mics. Time-align plugins for mic distance; not for spectral L/R fights on a printed stereo file.
4. Bypass wideners / Haas
If correlation recovers with the widener off, you found the tradeoff. Re-introduce width only above the problem band.
5. Frequency-aware phase rotation (PHIXER)
When polarity and timing are honest but some bands still cancel in mono on a stereo bus, spectral phase rotation can recover more summed energy without inventing missing fundamentals. Skip when the meter dive is intentional creative width you want to keep — when not to use PHIXER.
Decision table
| Meter / ear cue | First try | Why |
|---|---|---|
| Ø flip restores correlation + mono body | Keep polarity fix | Simple 180° mismatch |
| Small delay nudge restores mono | Time-align / mic distance | Arrival-time problem |
| Widener off → correlation healthy | Tame width | Haas / M/S caused anti-phase |
| Some bands still die in mono after Ø/timing/width | PHIXER on that stereo bus | Frequency-dependent residual fight |
| Negative correlation is the artistic point | Don’t “repair” — document the tradeoff | Intentional stereo smear |
Meter myths worth dropping
- “Always stay above +0.5.” Wide electronic / ambient material dips lower by design. Protect the low end; don’t flatten every side.
- “Negative means broken.” Negative means anti-correlated. Broken only if mono (or your delivery path) needs that energy.
- “A widener will fix a bad meter.” Usually the opposite. Width tools often create the dive. See imagers vs repair.
- “PHIXER replaces mono checks.” No. Ears + mono fold remain the judge; PHIXER is one repair after diagnosis.
Short practical checklist
- Put a correlation meter on the suspect bus and the master.
- Note typical range, then flip mono and name what disappears (sub, body, crack, presence).
- Solo groups until you find the diver.
- Try polarity, then timing, then bypass wideners.
- If a band still cancels on a stereo bus, try PHIXER — mono-check again.
- Only then add gentle high-band width if you still want image.
Order of operations: meter as warning → mono fold as verdict → polarity → timing → width hygiene → frequency-aware phase repair → creative width last. Skipping to “more width because the meter looks boring” is how hollow mono masters ship to phones.
Related reading
- Mono-check workflow — name what disappears
- Why bass disappears in mono
- Polarity invert vs phase rotation
- Phase repair vs time-align
- PHIXER vs stereo imagers
- Mid/Side vs phase repair
- When not to use PHIXER
- Product · Hub
Get the free fixer
PHIXER — free stereo phase-repair from Riverflux Audio (VST3/CLAP/AU · Win/Mac/Linux). Use it after the meter and mono fold agree a residual spectral fight remains.