How to Conduct an Acoustic Impulse Response Survey

By Priya Nair · February 16, 2026

How to Conduct an Acoustic Impulse Response Survey

1) What you’ll learn (and why it matters)

An acoustic impulse response (IR) survey is a structured way to measure how a room or space behaves over time and frequency: reflections, reverberation, clarity, and intelligibility. By the end of this tutorial you’ll be able to capture repeatable impulse responses, derive practical metrics (RT60/EDT, C50/C80, frequency response, early reflection timing), and interpret them in a way that helps you make real decisions—speaker placement, treatment priorities, mic choice/position, and system tuning.

Real-world use cases include: tightening dialog intelligibility in a small video studio, verifying reverb time in a church before a choir recording, comparing different audience areas in a venue, or diagnosing “why the PA sounds harsh in the balcony.” The goal isn’t just to take measurements—it’s to build an accurate acoustic picture you can act on.

2) Prerequisites / setup requirements

Measurement microphone: omni, calibration file preferred (e.g., Earthworks M30, iSEMcon EMX-7150, miniDSP UMIK-1). A calibrated mic reduces bias above 5 kHz where many rooms and speakers get tricky.
Audio interface: 2 in / 2 out is sufficient. Use stable drivers and disable system audio enhancements.
Reference loudspeaker: ideally a single full-range speaker with predictable directivity. For venue surveys, you can measure the actual PA, but a portable reference speaker makes comparisons easier.
Software: Room EQ Wizard (REW), ARTA, Smaart, or similar. REW is a common choice for IR capture via log sweep.
Stands and accessories: mic stand with boom, speaker stand, tape measure or laser distance tool, gaffer tape, headphones, and a notebook (or measurement log template).
Quiet time: you need stable noise conditions. HVAC rumble and exterior traffic will contaminate decay measurements.

Target baseline settings: 48 kHz sample rate, 24-bit recording, interface buffer 128–256 samples (not critical for sweep measurement, but helps stability). Disable automatic gain control anywhere in the chain.

3) Step-by-step: Conducting the survey

Step 1: Define the survey goal and choose measurement positions

Action: Write down what you’re trying to learn and select positions that represent real listening or recording locations.

Why: An IR is only meaningful when tied to a position. A theater doesn’t have one “sound”—front row, mix position, and balcony behave differently. Your goal determines how many points you need and what you’ll analyze (intelligibility vs. music clarity vs. low-frequency modal behavior).

Technique: Start with 6–12 mic positions for small/medium spaces. For a studio control room, 3–5 around the listening area may be enough. For venues, cover: FOH/mix position, front center, under balcony, balcony, and off-axis seats. Mark positions on a simple room sketch and label them (M1, M2, etc.).

Pitfalls: Measuring only one “good” spot and assuming it represents the room; choosing positions too close to walls (within 0.5 m) unless boundary behavior is specifically what you’re investigating.
Step 2: Set up a stable source and verify the signal chain

Action: Place the speaker and connect your interface output to it. Connect the measurement mic to an interface input and enable phantom power (48 V) if required.

Why: IR surveys depend on repeatability. If the speaker moves, rotates, or changes level between measurements, you’ll end up comparing your setup inconsistencies rather than the room.

Specific setup: Put the speaker on a stand with the acoustic center roughly 1.2–1.5 m high (typical ear height), aimed at the primary listening area. Keep it at least 1 m from large boundaries if possible to reduce early boundary domination (unless you’re measuring actual installed placement).

Pitfalls: Speaker on a chair or table (causes strong early reflections); mic cable rubbing the stand; phantom power off; unbalanced/ground noise contaminating the measurement.

Troubleshooting: If you see a strong 50/60 Hz hum in the spectrum, lift the audio ground safely via proper DI/isolators, or run balanced connections. Don’t “fix it later” with EQ—noise ruins decay metrics.
Step 3: Calibrate levels and set measurement SPL

Action: Set playback level so the sweep has good SNR without clipping the interface, speaker, or mic preamp.

Why: Too low and your decay tail disappears into noise (RT/EDT becomes unreliable). Too high and you’ll distort the speaker or overload the mic preamp, creating false harmonics and incorrect IR shape.

Recommended targets: Aim for 75–85 dB SPL (C-weighted) at the mic position for typical rooms. For large, noisy venues you may need 90 dB SPL briefly, but keep hearing safety in mind. In REW, adjust so the recorded sweep peaks around -12 to -6 dBFS and never hits 0 dBFS.

Pitfalls: Relying on DAW meters only; you can clip the mic preamp before the converter clips. Also, changing level between measurement positions destroys comparability.

Troubleshooting: If your IR shows a “fuzzy” or smeared direct arrival, reduce playback level and check for limiting in the speaker/amp DSP. Turn off any system EQ, compression, or “enhancers” during measurement unless they’re part of the system you’re evaluating.
Step 4: Choose the excitation signal (log sweep) and measurement settings

Action: Configure a log sine sweep and consistent capture settings across all positions.

Why: Log sweeps provide high SNR and allow harmonic distortion rejection in deconvolution, producing cleaner IRs than claps or starter pistols in most practical environments.

Suggested settings (typical):
- Sweep range: 20 Hz to 20 kHz (or 30 Hz to 18 kHz if your speaker can’t do extremes cleanly).
- Sweep length: 5–10 s for quiet rooms; 10–20 s for noisy venues. Longer sweeps improve SNR in the decay tail.
- Silence after sweep: 2–4 s minimum so the reverb tail is fully captured (longer for reverberant spaces).
- Sample rate: 48 kHz; bit depth: 24-bit.
Pitfalls: Using too short a sweep in a noisy environment (RT becomes noise-limited); using a sweep that drives the speaker into distortion (the IR will show pre-ringing-like artifacts or inflated HF energy).

Troubleshooting: If the software reports poor correlation or unstable timing, ensure the input channel is the measurement mic (not loopback), and disable acoustic echo cancellation on laptops/OS settings.
Step 5: Place the mic correctly and document geometry

Action: Place the mic at the chosen position, set height and orientation, and measure distance to the source.

Why: Small positional changes can shift comb filtering and early reflection timing. Good documentation lets you repeat measurements after treatment or system changes and compare meaningfully.

Specific technique: Use a mic height of 1.2 m seated or 1.5 m standing unless your use case demands otherwise. For room surveys, point the mic straight up (common measurement practice for omni mics) to reduce directional bias; for system voicing or PA alignment, point toward the source to emphasize on-axis response. Record: mic height, distance, and any nearby reflective surfaces (desk, balcony edge, side wall distance).

Pitfalls: Handholding the mic (movement during sweep); measuring too close to a seat back or tabletop; placing the mic in a pressure maximum corner unless investigating low-frequency buildup.
Step 6: Capture multiple takes and check IR quality immediately

Action: Record at least 2 sweeps per position and validate the impulse response before moving on.

Why: A single take can be ruined by a door slam, HVAC cycling, or someone walking through. Two consistent takes confirm repeatability.

What to look for: In the IR view, you should see a strong, clean direct arrival spike, followed by discrete early reflections, then a smooth decay. The time of the direct arrival should match distance roughly: ~2.9 ms per meter (speed of sound ≈ 343 m/s). Example: 5 m ≈ 14.6 ms.

Pitfalls: Ignoring a delayed direct arrival (often indicates you measured the wrong source, or the speaker DSP introduced large latency); clipping in the recorded sweep; inconsistent IRs between takes.

Troubleshooting: If the decay “floor” is jagged and stops early, increase sweep length from 5 s to 10–20 s, or raise SPL slightly (within safe limits). If you see multiple similar “direct” spikes, you may be too close to a strong reflector (glass, side wall); move the mic 0.5–1 m and re-measure.
Step 7: Window the IR and extract meaningful metrics

Action: Apply time windows to isolate direct sound vs. room behavior, then generate frequency response and decay metrics.

Why: A raw IR contains everything. Windowing helps you answer specific questions: “Is the tonality problem from the speaker or early reflections?” vs. “Is the room too live?”

Practical windows:
- Direct/early window (for speaker + early reflection assessment): start at the direct arrival, end at 5–20 ms depending on room size. In small studios, 5–10 ms can reveal desk/console reflections; in larger rooms, 10–20 ms is more realistic.
- Full window (for reverberation/decay): long enough to include the entire tail; don’t truncate if you want reliable RT estimates.
Key metrics to compute:
- RT60 / T20 / T30: use T20 or T30 depending on available decay range. If the noise floor limits you, T20 is often more stable.
- EDT (Early Decay Time): correlates strongly with perceived liveliness, often more relevant than RT60 in small rooms.
- C50 (speech) and C80 (music): clarity ratios. As rough guidance, many rooms benefit from higher C50 for intelligibility; very low C50 often matches “washy” dialog.
- Frequency response: apply smoothing appropriate to the task (e.g., 1/12-oct for detailed work, 1/6-oct for broader trends).
Pitfalls: Comparing RT values across positions without checking noise floor; using heavy smoothing (1/3-oct) and missing narrow resonances; treating RT60 as a single number instead of a frequency-dependent curve.

Troubleshooting: If RT curves jump erratically at low frequencies, your speaker may not be exciting LF evenly, or the room’s modal behavior is dominating. Increase measurement positions and consider averaging, or use a subwoofer as the source for LF-focused sweeps (e.g., 20–200 Hz).
Step 8: Repeat systematically and keep variables constant

Action: Move to the next mic position and repeat the exact same sweep settings, level, and source orientation.

Why: The power of a survey is comparative data. Consistent methodology turns your set of IRs into a map of the room rather than a pile of unrelated captures.

Technique: Use a checklist: speaker position fixed, speaker aim fixed, playback level fixed, mic height consistent, and measurement naming consistent (e.g., “Hall_M3_1p5m_48k”). For each position, note anomalies (near pillar, under balcony lip, open door).

Pitfalls: “Small” changes that aren’t small: rotating the speaker 10 degrees, changing mic height by 20 cm, or switching input gain mid-session.
Step 9: Interpret results in real-world terms (not just plots)

Action: Translate the IR and derived metrics into actionable conclusions for the scenario you’re working in.

Why: Clients and collaborators don’t need graphs—they need decisions: move speakers, add absorption, adjust coverage, change mic technique, or accept the room and plan accordingly.

Examples:
- Podcast/video studio: If EDT is long above 500 Hz and early reflections arrive within 5–10 ms from a desk or monitor, you’ll hear “boxy” comb filtering and flutter. Treatment priority: desk reflection control (angled desk mat or reposition), then broadband absorption at first reflection points.
- Church/large hall: If C50 is low at mid/high frequencies in rear seats, intelligibility will suffer. Consider speaker aiming, delayed fills, or distributed reinforcement rather than only adding absorption.
- Control room: If frequency response shows deep notches that change dramatically across small mic movements, that’s modal/comb behavior—treatment and positioning matter more than EQ.
Pitfalls: Over-correcting with EQ when the issue is reflection timing; assuming one position’s response is the “truth” for the whole space.

4) Before/after comparison and expected results

After a proper IR survey, you should have:

A labeled set of IRs for each position, with repeatable direct arrival timing and consistent levels.
RT/EDT curves by frequency that show where the room is too live (often 1–4 kHz in untreated small rooms) or too ringy (often 63–200 Hz in modal spaces).
Clarity indicators (C50/C80) that correlate with what you hear: under-balcony seats often show weaker early energy and smeared clarity; reflective side walls show strong early reflections inside 10–30 ms.
Actionable deltas: You can re-measure after a change (adding panels, moving a speaker, adding a fill) and see objective improvement—shorter EDT, smoother early reflection pattern, or improved clarity in problem zones.

A realistic “after” outcome in a small studio might be: EDT above 1 kHz reduced by 15–30%, early reflection spikes 6–12 ms reduced by 6–10 dB after adding absorption at reflection points, and speech sounding less phasey on camera mics.

5) Pro tips to take it further

Use spatial averaging: For listening areas, measure a cluster of 4–9 points in a 0.5–1 m radius and average frequency response. This reduces seat-to-seat comb artifacts and reveals trends you can actually fix.
Measure at multiple source positions: In a performance space, capture IRs with the source at stage center, stage left, and stage right. Real shows aren’t a single point source.
Track environmental conditions: Temperature and humidity affect speed of sound and HF absorption slightly. Note temperature if you’re doing alignment work or comparing across days.
Check noise floor explicitly: Record 10 seconds of room noise at each position. If the noise floor is within 25–35 dB of your sweep level, RT estimates will be noise-limited, especially at low frequencies.
Look at early reflection timing: A strong reflection at 8–12 ms often reads as coloration; 20–40 ms often reads as “room” or “slap.” The IR tells you what to treat first.
Keep raw data: Save raw IRs and project files. Future-you will want to re-window or re-analyze with new questions.

6) Wrap-up: build skill through repetition

Impulse response surveys reward consistent technique. Run the same process in two different rooms—a small office and a larger hall—and compare how the IR changes: the spacing of early reflections, the slope of the decay, and how clarity metrics line up with what your ears report. The more surveys you conduct, the faster you’ll recognize patterns and the easier it becomes to choose the right fix: placement, treatment, or system design.

If you want to practice efficiently, pick one room, measure it today, change one variable (speaker position by 0.5–1 m, add two broadband panels, or close a curtain), and measure again. The cause-and-effect relationship you’ll see in the IR is one of the quickest ways to level up your acoustic instincts.