Repeatable Audio Measurement

What You Will Be Able To Do

• Choose the correct measurement for the question instead of opening a graph at random.
• Map and document the complete source-to-analysis chain.
• Separate relative response correction from absolute SPL calibration.
• Capture a repeatable frequency-response baseline.
• Use smoothing, windowing, and spatial averaging without hiding the problem being diagnosed.
• Measure distortion and SPL with stated level, bandwidth, detector, and safety conditions.
• Build a record another person can repeat.

Start With The Question

Measurement Methods At A Glance

1Define The Device, Position, And Goal

Write one sentence that defines the test:

Measure the driver's-seat average frequency response of the front left and right system from 50 Hz to 20 kHz at a safe tuning level, with the current crossovers and no EQ.

Name the listening position or sensor position, the channels included, the frequency region, the system state, and the decision the result will support.

2Map The Measurement Chain

signal file or generator
→ playback app and source mode
→ interface / head unit / factory integration
→ DSP preset and processing
→ amplifier and load
→ loudspeaker and cabin
→ microphone / probe
→ interface / ADC
→ calibration and timing reference
→ software analysis

Disable or document loudness, volume normalisation, ANC, factory EQ, automatic gain control, bass restoration, dynamic processing, and hidden presets.

3Build A Safe Level Baseline

1. Confirm routing, mutes, crossover protection, and driver polarity.
2. Begin at a conservative level.
3. Check source, DSP, amplifier, microphone, and ADC headroom.
4. Raise the signal only enough to sit clearly above ambient noise.
5. Stop if the driver, amplifier, electrical system, microphone, or input stage shows distress or overload.

REW commonly suggests a moderate measurement level around normal listening level rather than maximum output. The correct level is the lowest one that gives adequate signal-to-noise ratio without stressing the system.

4Load The Correct Calibration

• match microphone serial number
• match 0-degree or 90-degree orientation
• record whether sensitivity data is included
• keep interface and operating-system gain unchanged after SPL calibration
• label the result relative if no absolute sensitivity reference exists

5Fix The Microphone Position

Use a stable stand and a documented reference point. Record height, lateral position, fore-aft position, orientation, seat location, and whether the microphone represents one ear, head centre, competition sensor location, or an average region.

A few centimetres can move a narrow comb-filter dip dramatically. Repeatability matters more than chasing a mythical perfect inch.

6Add A Timing Reference When Needed

Magnitude-only measurements can be useful without absolute timing. Use a wired loopback, acoustic timing reference, or documented equivalent before comparing phase, impulse arrival, delay, vector averages, or trace arithmetic. Keep the reference path unchanged.

7Capture And Repeat The Baseline

1. Run the first measurement and save it immediately.
2. Do not change the system.
3. Run the same measurement again.
4. Overlay the traces.
5. If they do not agree closely enough for the decision, investigate noise, movement, timing, gain, clock, or hidden processing.

A repeatable ugly graph is useful. A beautiful graph that cannot be repeated is not.

8Choose The Display Intentionally

• Use light smoothing for detailed crossover and resonance work.
• Use broader or psychoacoustic smoothing for tonal overview, but retain the raw trace.
• Keep before-and-after axes identical.
• Record normalisation and vertical scale.
• Do not interpret gated data below the valid frequency limit.

9Use Windowing Only For A Defined Reason

A shorter impulse window can reduce later reflections while removing low-frequency information. Choose whether you are measuring direct sound, installed cabin response, reflection behaviour, nearfield response, or a merged result. Save the raw impulse and document the window.

10Add Spatial Information

After the single-point baseline is stable, measure nearby positions. Keep the individual traces. Create a documented RMS or dB average for tonal work, or use vector averaging only when stable relative timing is preserved. A moving-microphone average is useful for broad tonal balance but not for impulse timing or crossover phase.

11Run Specialised Tests

Distortion

Record fundamental level, sweep or stepped-sine method, harmonics included, bandwidth, noise floor, load, distance, and system temperature. Repeat at multiple levels.

SPL and exposure

Record A, C, or Z weighting; fast, slow, peak, Leq, or SEL; sensor placement; signal; duration; calibration; and input overload state. Use A-weighted level plus duration for hearing-exposure context.

12Change One Variable And Verify

1. Duplicate or save the baseline preset.
2. Change one filter, delay, level, position, or component.
3. Repeat the same measurement chain and display settings.
4. Compare the intended band and check unintended changes elsewhere.
5. Repeat the final result and inspect nearby positions, headroom, distortion, and safety.

Troubleshooting Matrix

Continue

• Print the Repeatable Measurement Workflow
• Measurement Chain
• Frequency Response
• Distortion Measurement
• SPL Measurement