12.5 Equalization — Measurement-Driven Correction

🔰 BEGINNER LEVEL: What EQ Can and Cannot Fix

The Purpose of In-Car EQ

The frequency response measured at your ear in a car will never be flat. Unavoidable factors:

• Cabin resonances: The rectangular interior has standing wave modes that boost certain bass frequencies 6–15 dB
• Seat and glass reflections: Create comb-filtering peaks and dips in the midrange
• Driver placement asymmetry: Left and right paths differ, creating response differences
• Speaker roll-offs: Natural frequency limits of drivers

EQ corrects the response you measured, not the ideal response.

Important: EQ applied to a problem that isn't acoustic (a broken connection, a phase issue, a fundamentally incompatible crossover point) doesn't fix the problem — it compensates partially at the measurement position and may make other positions worse.

Fix problems first, EQ last.

Target Curves

A flat frequency response in-room is not the correct target for car audio. Research consistently shows that listeners prefer a specific curve with bass lift.

Harman Automotive Target Curve:

• +6 dB at 20 Hz relative to 300 Hz
• Flat from 300 Hz to 2 kHz
• Gentle −4 dB shelf above 10 kHz

Illustration note: Target curves comparison — Harman, flat, and measured system overlaid

Apply this as your EQ target. When REW shows your measured response, compare it to this target and add EQ filters to bring the measurement closer to the target.

🔧 INSTALLER LEVEL: EQ Workflow and Filter Strategy

The Correct EQ Approach

Always cut before boosting. Every boost increases the signal level at that frequency, risking clipping at some point in the signal chain and requiring increased gain elsewhere to compensate. Cuts don't cause these problems.

Maximum boost: 6 dB. If a frequency needs more than 6 dB of boost, there's a deeper problem — driver capability, enclosure, or crossover.

Illustration note: Before and after EQ response graphs showing measured, corrected, and target curves

Systematic workflow:

Step 1 — Measure:

In REW, measure the frequency response at the listening position. Use a calibrated microphone (UMIK-1 recommended). Average 4 positions within 12 inches of the listening position.

Step 2 — Identify problems:

Look for: - Peaks > 6 dB above target curve: identify center frequency and width - Dips > 6 dB below target: check if they're caused by phase cancellation first (EQ cannot fix phase cancellation — only phase adjustment can)

Step 3 — Apply filters:

Work from low to high frequency:

Start below 200 Hz. These are typically broad resonances (low Q ≈ 0.5–1.0). Use wide-band filters.

Proceed to 200 Hz–1 kHz. Seat and body reflection zone. Medium Q filters (1.0–2.0).

Address above 1 kHz last. Narrow resonances possible (Q 2–5) but avoid over-correcting — small positional changes affect HF significantly.

Step 4 — Re-measure, compare, iterate:

After each batch of EQ filters, re-measure. Compare to target. Do not adjust more than 4–6 filters at a time before measuring again.

Identifying and Handling Phase Cancellation

Symptom: A sharp dip that stays narrow across all microphone positions. EQ boosts at that frequency produce no audible improvement.

Test: Disconnect one channel (sub only, or one speaker). Does the dip disappear? If yes — phase cancellation between two sources.

Fix: Adjust time alignment or polarity of the canceling driver. Re-measure before continuing EQ.

EQ cannot fix phase cancellation — boosting at a cancellation null increases level from both sources, and they still cancel. Result: more distortion from both drivers working harder, same acoustic null.

⚙️ ENGINEER LEVEL: Minimum-Phase vs Non-Minimum-Phase Responses

Classification of Acoustic Measurements

Minimum-phase response: The phase response is uniquely determined by the amplitude response. Every dB of amplitude change has a corresponding predictable phase shift. Minimum-phase anomalies CAN be corrected by EQ — correcting the amplitude also corrects the phase.

Non-minimum-phase response: Phase is not determined by amplitude. Created by: - Acoustic cancellation between two sources arriving at different times - Room reflections (multipath effects) - FIR filtering with symmetric impulse response

Non-minimum-phase anomalies in the acoustic measurement CANNOT be corrected by IIR EQ — correcting the magnitude makes the phase worse, or vice versa.

Distinguishing minimum-phase in REW:

1. Measure impulse response
2. Use REW's "Minimum Phase Response" feature to compute the minimum-phase version of your measurement
3. Compare to actual measurement
4. Regions where actual phase ≠ minimum-phase phase: non-minimum-phase anomalies
5. In those regions, EQ amplitude correction will not fix the phase — accept the limitation or address with time alignment

Practical implication: In a car, below approximately 300 Hz, the response is largely minimum-phase (dominated by single direct path and room modes). Above 300 Hz, multipath becomes significant — expect non-minimum-phase regions. Aggressive EQ above 1 kHz risks making the phase response worse.