--> Lesson 7 — Impedance, Thiele-Small, Room Sim and the API | REW Professional Training

Lesson 07 · Electro-Acoustics, Spatial Synthesis & Automation

Lesson 7 — Impedance, Thiele-Small, Room Sim and the API

Move beyond single-mic acoustic sweeps into transducer characterization, complex load diagnostics, modal pre-modeling, multi-input spatial capture, deterministic trace math, cross-tool data exchange, and headless HTTP API orchestration on macOS.

Dual-Channel Z(f) & EPDR Dual Added-Mass T/S Up to 32-Ch Pro Capture REST API · localhost:4735

01 · Engineering Scope

What This Is For

Standard room sweeps tell you what arrived at a microphone, not how the loudspeaker behaves as an electro-mechanical load, how multiple drivers and seats combine mathematically, or how to scale measurements across production QA benches and multi-room commissioning workflows. The tools in Lesson 7 turn REW from a manual acoustic analyzer into an integrated electro-acoustic lab bench and scriptable measurement engine.

Subsystem Professional Problem Solved Primary Engineering Deliverable
Impedance & EPDR Verify crossover networks, port tuning ($F_b$), enclosure air leaks, batch driver tolerances, and worst-case amplifier current/dissipation stress. Complex $|Z(f)|$, phase $\theta(f)$, real resistance $R(f)$, and EPDR curve (exportable as .zma).
Thiele-Small Extraction Measure actual post-break-in electromechanical parameters rather than trusting nominal OEM datasheets before cabinet or DSP alignment. $F_s$, $Q_{ms}$, $Q_{es}$, $Q_{ts}$, $V_{as}$, $M_{ms}$, $C_{ms}$, $R_{ms}$, $Bl$, $\eta_0$, plus fitted impedance overlay.
Room Simulator Screen subwoofer topologies, crossover polarity/delay, and seat-to-seat modal variance before rigging heavy hardware or building soffits. Lossy-boundary modal prediction curves and direct simulated traces pushed into REW’s main graphs.
Multi-Input Capture Eliminate sequential mic-stand repositioning during multi-seat room averaging or multi-point polar/array diagnostics. Simultaneous capture of up to 32 calibrated input channels (Pro upgrade) with unified timing reference.
Trace Math & Averaging Synthesize valid multi-seat EQ targets, splice nearfield bass to gated farfield mids/highs, extract differential transfer functions, and align sub/main phase. RMS/dB/Vector averages, regularised A/B transfer divisions, spliced Merge B to A traces, Aligned Sums, and IACC.
Import / Export & REST API Integrate REW with VituixCAD, miniDSP, CamillaDSP, rePhase, SigmaStudio, Python/Node QA scripts, and CI regression pipelines. .frd/.zma/.wav/.pir/biquad files and automated JSON/HTTP orchestration over 127.0.0.1:4735.
macOS Architecture Constraints to Keep in Mind
  • Audio I/O: macOS uses Java audio drivers exclusively (CoreAudio bridge). ASIO and WASAPI Exclusive (EXCL:) are Windows-only. Configure sample rate, 24-bit/32-bit word length, and channel levels in Audio MIDI Setup before launching REW; REW’s internal volume controls are disabled on macOS.
  • Permissions & Logs: If macOS silently blocks multichannel or reference inputs after an update, run tccutil reset Microphone in Terminal before starting REW. Diagnostic logs and 7-day automatic temporary measurement backups live in ~/Library/Logs/REW (and /temp).
  • Control Interfaces: RS232 serial control (e.g., TAG McLaren) is Windows 32-bit only, and MIDI port selection is disabled by default on macOS due to JRE crash risk. For automation on macOS, the localhost HTTP REST API is the canonical path.

02 · Concrete Procedures

Steps: Lab Setup Through API Automation

A. Dual-Channel Impedance Rig & Three-Point Calibration

  1. Build and verify the sense-resistor jig. Wire a precision non-inductive sense resistor ($R_s$) in series between the interface output (or low-output-impedance power/headphone amplifier) and the Device Under Test (DUT). Connect the Reference input across the drive output (before $R_s$) and the Measurement input across the DUT junction (after $R_s$). Verify 48 V phantom power is OFF, hardware direct monitoring is muted, and both input preamps are set to identical fixed gain (line-level inputs preferred).
  2. Lock macOS Audio MIDI Setup and REW Soundcard Preferences. Set the interface to 48 kHz / 24-bit in Audio MIDI Setup. Launch REW, open Preferences → Soundcard, explicitly choose the named hardware device (never Default Device), and assign the Measurement Input, Reference Input, and Output channels. Verify sample rate and bit depth in the bottom status bar.
  3. Configure the Impedance Measurement dialog. Press ⌘+M (Measure) and select Impedance mode. Measure your sense resistor with an accurate DMM (4-wire Kelvin if available) and enter the exact measured resistance—not the nominal printed value—into Sense Resistor.
  4. Execute the 3-step rig calibration in strict sequence.
    • Step 4a — Open Circuit Cal: Leave the DUT test clips open (disconnected, not touching) and run Open Circuit Calibration. This characterizes channel gain/phase mismatch and interface input impedance.
    • Step 4b — Short Circuit Cal: Clamp the DUT test clips firmly together and run Short Circuit Calibration. This nulls test-lead series resistance and parasitic lead inductance.
    • Step 4c — Reference Resistor Cal: Connect a known, precision non-inductive reference resistor (close to expected DUT impedance, e.g. $8\text{–}100\,\Omega$) across the DUT clips, enter its exact measured value, and run Reference Calibration.
  5. Set sweep length, band, and small-signal drive level. Set Start Freq to 20 Hz (or 5–10 Hz for subwoofers) and End Freq to 20 kHz. Select a sweep duration of 10 seconds or longer (e.g. 512k or 1M samples at 48 kHz) to resolve high-$Q$ free-air resonance peaks without window smearing. Adjust output level so voltage across the driver terminals sits around 100–200 mV RMS—high enough for clean SNR while keeping cone excursion inside the small-signal linear region.
  6. Capture the free-air baseline. Clamp the driver rigidly by its magnet or basket rim in free air (pole vent completely unobstructed, cone axis clear of nearby desks/walls that would add acoustic radiation mass), name the trace clearly (e.g. Woofer_FreeAir_150mV), and press Start. Confirm headroom stays green (6–40 dB) and verify phase polarity immediately.

B. Thiele-Small Parameter Extraction

  1. Capture perturbation measurement(s) without disturbing calibration. Without changing gain, sample rate, or lead wiring, perform one of REW’s three perturbation protocols:
    • Dual Added Mass (preferred for highest accuracy): Weigh two distinct non-magnetic masses ($m_1$ and $m_2$, typically shifting $F_s$ downward by roughly 20–30% and 35–50% respectively) on a 0.01 g scale. Distribute $m_1$ symmetrically around the cone neck/dust-cap perimeter using tack/putty, sweep and save; then add or swap to $m_2$ and sweep again.
    • Single Added Mass: Same procedure with one known mass shifting $F_s$ by ~20–35%.
    • Sealed Box: Mount the driver airtight into a rigid, unstuffed (or accurately accounted) sealed test enclosure of precisely known internal net volume $V_{\text{test}}$ (subtracting driver basket/magnet displacement) and sweep.
  2. Measure true voice-coil DC resistance ($R_{dc}$). Disconnect the driver from the rig, allow the coil to stabilize at ambient temperature, and measure $R_{dc}$ across the terminals (subtracting shorted-probe resistance on a 2-wire meter, or using a 4-wire Kelvin bridge).
  3. Compute parameters in the Thiele-Small window. Open Tools → Thiele Small Parameters. Select your free-air measurement and added-mass (or dual-added-mass / sealed-box) trace(s). Enter measured $R_{dc}$, effective piston area $S_d$ (measured from mid-surround to mid-surround: $S_d = \pi r_{\text{eff}}^2$), exact added mass(es) or box volume, ambient temperature, and barometric pressure.
  4. Validate fit and export. Click Calculate Parameters. Inspect the automatically generated modelled impedance overlay against the measured curves across both $F_s$ and the inductive voice-coil rise. Copy/export the parameter report and save the `.mdat` bundle.

C. Room Simulator Pre-Modeling

  1. Define room geometry and boundary loss. Open Room Simulator from the toolbar. Enter exact interior Length, Width, and Height. Set per-surface absorption coefficients (floor, ceiling, front/rear/side walls) to approximate real room construction (e.g., drywall/stud cavities absorb far more LF energy than poured concrete), and toggle sealed-volume modeling where appropriate.
  2. Configure sources, crossovers, and listening array. Enable main speakers and one to four subwoofers. Assign crossover frequencies, individual gains, delays, and polarities. Add secondary microphone positions around the main listening position to evaluate seat-to-seat spatial consistency in real time as you drag subs in plan/elevation views.
  3. Export simulated traces into the workspace. Click the trace-generation controls to push combined or individual simulated source responses (at main or selected mic positions) directly into REW’s Measurements Panel. Overlay them against physical room sweeps in All SPL to diagnose where real structural leaks or non-rectangular openings diverge from the idealized modal model.

D. Multi-Input Capture, Trace Arithmetic & Alignment

  1. Configure simultaneous multi-input capture (Pro upgrade). On a multichannel interface (up to 32 inputs supported with the Pro upgrade), verify all input channels are active at matching sample rate and 24-bit depth in macOS Audio MIDI Setup. Open Level Meters in multi-input view to verify signal presence and headroom across every mic channel. If multichannel output connector order on macOS differs from standard channel labeling, remap assignments in Preferences → Soundcard → Output Channel Mapping. Assign per-channel microphone calibration files before capturing.
  2. Pre-align traces in All SPL / Overlays before math. Open All SPL, check the target measurement subset, and open the Actions panel. Run Align SPL (over a clean midband window) to normalize absolute level offsets. Before any phase-dependent operation, apply Time align, Align IR start, or Cross corr align—confirming all traces share the same sample rate and were captured with a valid timing reference (loopback or acoustic timing reference).
  3. Select the correct averaging operator.
    • Use RMS average for multi-position room EQ targets (treats seats as power/incoherent sums without introducing artificial comb-filter nulls).
    • Use dB average when equal logarithmic weight per trace is preferred without high-peak dominance.
    • Use Vector average strictly for time- and phase-aligned coherent captures (e.g., repeated fixed-position sweeps or synchronous array summing).
    • Use RMS + phase or dB + phase when you need an incoherent magnitude average paired with an averaged phase curve for subsequent minimum-phase/filter work.
  4. Run Trace Arithmetic, Alignment Tool, and IACC. Open All SPL → Actions → Trace Arithmetic to execute A+B, A-B, A*B, A/B (with regularisation to prevent extreme gain spikes at deep notches), Invert, or Merge B to A (which automatically level- and phase-aligns B to A at the merge frequency—ideal for splicing nearfield bass to gated farfield measurements). For sub/main integration, open Alignment Tool, choose Phase or Impulse alignment mode, adjust delay/gain/polarity for Trace B, and generate an Aligned sum or Aligned copy. For binaural pairs, run IACC to inspect early, late, and full inter-aural cross-correlation.

E. Data Import/Export & REW REST API Orchestration

  1. Exchange measurement and DSP files. Use File → Import to ingest frequency responses (.txt, .frd, .zma), impulse responses (.wav, .aiff, ARTA .pir, .pcm, ETF), or offline recorded sweeps (for devices that cannot connect live to REW). Use File → Export to write text magnitude/phase, windowed impulse WAVs, EQ filter text, or sample-rate-specific biquad coefficients for miniDSP, CamillaDSP, SigmaStudio, StormAudio, Hypex, or rePhase.
  2. Start REW with the HTTP API server on macOS. Launch REW from Terminal with the -api flag (binds strictly to localhost / 127.0.0.1 on default port 4735, or specify a custom port via -port):
macOS Terminal — Launch REW API Server (Default & Custom Port)
# Launch REW with API enabled on default port 4735 (localhost / 127.0.0.1 only)
open -a REW.app --args -api

# Or bind to a custom localhost port (e.g. parallel QA instances / port conflict)
open -a REW.app --args -api -port 4567

# Verify API readiness and list active endpoints / application status
curl -s http://127.0.0.1:4735/application/commands | jq .
  1. Query measurement state via GET endpoints. Standard REW permits read-only GET queries across endpoints such as /measurements, /audio, /eq, /roomsim, and /spl-meter to pull measurement lists, frequency responses, impulse data, and calibration state into Python/MATLAB/Node post-processors.
  2. Automate sweeps, EQ, and exports via PUT/POST endpoints (Pro upgrade). With the REW Pro upgrade, issue PUT/POST commands to /measure, /generator, /rta, /stepped-measurement, /alignment-tool, /eq, /import, and /application to run unattended sweep matrices, update target curves, compute filters, and archive results.

03 · Data Interpretation & Diagnostics

Reading the Output

Impedance View — Complex |Z(f)|, Phase θ(f) & EPDR Load Analysis
|Z(f)| Magnitude (Ω) Phase θ(f) (°) EPDR (Ω)
Vented 2-Way / Woofer Diagnostic Anatomy
40 Ω 30 Ω 20 Ω 10 Ω 4 Ω 0 Ω +90° +45° 0° -45° -90° 20 50 100 200 500 1k 2k 5k 10k 20k Hz Fb Saddle (θ=0°) EPDR min (~2.1 Ω): low |Z| + steep -θ Xover Z-peak
1. Impedance, Phase & EPDR Impedance Graph

In free air, a raw moving-coil driver exhibits a single symmetrical resonance peak at $F_s$ where phase crosses $0^\circ$ (inductive below $F_s$, capacitive immediately above $F_s$, then climbing inductively at HF due to voice-coil inductance $L_e$). In a vented enclosure, look for twin peaks ($F_L, F_H$) bracketing a saddle minimum at port tuning $F_b$ where phase also crosses $0^\circ$.

Always enable EPDR (Equivalent Peak Dissipation Resistance) when evaluating complete loudspeakers. EPDR translates reactive phase lag/lead plus magnitude into the equivalent purely resistive load that would produce the same peak output-stage dissipation in a Class-AB amplifier. A speaker whose $|Z|_{\min}$ is $4.0\,\Omega$ can drop below $2.0\,\Omega$ EPDR on the capacitive shoulder right above an upper resonance peak.

2. Thiele-Small Parameters & Model Overlay T/S Window

Check the physical plausibility of extracted parameters before exporting: $Q_{ts}$ must satisfy $Q_{ts} = (Q_{ms} Q_{es}) / (Q_{ms} + Q_{es})$, and $V_{as} = \rho_0 c^2 S_d^2 C_{ms}$ scales with the fourth power of effective cone radius—making $S_d$ measurement error the most common cause of $V_{as}$ drift.

Judge extraction fidelity primarily by the modelled impedance overlay: tight tracking across both resonance peaks and skirts validates the lumped model. Divergence above $2\text{–}3\times F_s$ reflects eddy-current semi-inductance ($L_e$ loss); asymmetry or glitches right on the resonance slope indicate buzzing dust caps, rubbing coils, or mass-putty rocking modes.

3. Room Simulator & Modal Resonance Lines Room Simulator

Vertical modal markers identify axial, tangential, and oblique resonances derived from REW’s lossy-boundary modified rigid-boundary solution. Look for where a single subwoofer lands on a modal pressure node (creating an un-EQable deep dip) versus where a second sub placed asymmetrically or delayed/inverted fills that null across all enabled mic seats.

Treat the Room Simulator as a low-frequency modal topology optimizer, not an exact SPL oracle. Wall flex, doorways, open stairwells, and non-uniform absorption shift actual modal Q and frequencies—always verify with physical measurements once hardware is positioned.

4. Averaging, Arithmetic & Alignment Tool All SPL / Alignment

Comparing an RMS average against a Vector average of the same multi-seat set immediately exposes phase incoherence: wherever the Vector average drops sharply below the RMS average, path-length differences exceed a quarter-wavelength and coherent cancellation dominates. Never EQ based on a Vector average of unaligned seats.

In the Alignment Tool, confirm both crossover magnitude summation and phase-slope alignment across the transition band (not just a single cursor frequency), or use Impulse alignment mode (which cross-correlates 1/3-octave filtered IRs) when phase wrapping obscures the true integer-cycle delay.

Diagnostic Signatures: Good Capture vs. Faulty Setup

04 · Advanced Engineering Depth

Deeper Detail

Electro-Mechanics Sense-Resistor Optimization, Rig Error Budgets & Dual Added-Mass T/S Math

In REW’s dual-channel impedance bridge, complex impedance is solved from the complex voltage ratio between the reference channel (drive side of $R_s$) and the measurement channel (DUT terminal):

Z_DUT(f) = R_s · [ H(f) / (1 - H(f)) ]  where  H(f) = V_meas(f) / V_ref(f).

Because the denominator approaches zero as $|Z_{DUT}| \gg R_s$ and the numerator shrinks as $|Z_{DUT}| \ll R_s$, sense-resistor selection directly bounds dynamic accuracy:

  • Too large $R_s$ (e.g., $1\,\text{k}\Omega$ on a $4\,\Omega$ subwoofer): Voltage across the DUT drops by ~48 dB relative to the drive rail, degrading SNR near $R_{dc}$ and exaggerating cable capacitance / preamp input-impedance shunt errors.
  • Too small $R_s$ (e.g., $1\,\Omega$ directly on a line output): Loads the audio interface output stage into distortion/current-limiting and shrinks the voltage drop across $R_s$ at the $50\text{–}100\,\Omega$ resonance peak, amplifying differential gain-mismatch noise.
  • Practical sweet spot: Use ~$100\,\Omega$ (1% or 0.1% metal-film / non-inductive) when driving directly from a headphone/line output with modest output impedance, or ~$10\text{–}27\,\Omega$ (high-wattage non-inductive) when buffering the drive through a low-output-impedance power amplifier.

Why Dual Added Mass outperforms Sealed Box and Single Added Mass: A sealed test box introduces adiabatic-to-isothermal compliance uncertainty, microscopic enclosure leakage (shifting effective $Q_{ms}$), and internal air-load perturbation. Single added mass eliminates box leaks but relies on one compliance-shift point. Dual Added Mass solves $M_{ms}$ and $C_{ms}$ across two independent known perturbations ($m_1, m_2$), giving REW an overdetermined cross-check that reduces sensitivity to single-measurement noise—provided both masses are weighed accurately, distributed symmetrically near the dust cap without stiffening the cone, and measured at identical ambient temperature and small-signal terminal voltage.

DSP & Signal Processing Trace Arithmetic Numerics, Window Coupling, Regularisation & 32-Ch Pro Capture

Trace Arithmetic operates on the currently windowed frequency responses of operands A and B. Two numerical traps routinely trip up even seasoned engineers:

  1. Hidden IR-window mismatch: Performing A + B or A / B between a 4 ms gated farfield trace (valid only down to $\sim 250\,\text{Hz}$) and a 500 ms room trace contaminates the result below the shorter window’s resolution limit. When splicing nearfield low frequencies to gated farfield mids/highs, use Merge B to A rather than manual addition: REW aligns both magnitude and phase seamlessly at the splice frequency while preserving the high-resolution LF data below and the reflection-free gated data above.
  2. Singularity blow-up in A / B and Invert: Dividing by a measurement containing a deep non-minimum-phase acoustic cancellation creates massive gain spikes and pre-ringing if exported as an FIR/convolver filter. Always configure REW’s regularisation parameter to bound maximum boost at notches, and restrict inversion to minimum-phase regions confirmed by a flat Excess Group Delay trace.
  3. Averaging semantics & clock constraints: All operands in Trace Arithmetic, Averaging, and the Alignment Tool must share the same sample rate. Remember that multiple-sweep synchronous averaging in the Measure dialog is invalid when using a USB mic or separate input/output clock domains because sample-clock drift smears the high-frequency IR; for multi-mic arrays (up to 32 inputs with the Pro upgrade), run all microphones into a single word-clock-locked multichannel interface over macOS CoreAudio/Java.
Automation & Integration REW HTTP REST API Architecture, Endpoint Map & Licensing Boundaries

REW’s embedded HTTP server listens exclusively on localhost / 127.0.0.1 (default port 4735) to prevent accidental LAN exposure of audio output hardware. Read-only GET endpoints are available in standard REW; active control via PUT/POST—including automated sweep triggering and stepped-sine orchestration—requires the REW Pro upgrade.

Endpoint Prefix Capabilities Exposed Typical Automation Use Case
/application Application commands, errors, warnings, graceful shutdown. Watchdog health polling, error capture, clean teardown in CI benches.
/audio Driver info, sample rate, Java/ASIO settings, input/output calibration. Pre-flight assertion that macOS Audio MIDI Setup rate matches 48 kHz and cal files are loaded.
/measurements Measurement list, frequency/impulse data, IR windows, EQ, target, filters. Batch-pull magnitude/phase JSON into Python/SciPy, apply uniform IR windows, archive results.
/measure Sweep configuration, naming, notes, level, timing ref, playback mode, trigger. Automated multi-channel or multi-angle production QC sweep loops (Pro upgrade).
/alignment-tool Programmatic delay, gain, polarity, and phase/impulse alignment control. Automated sub/main integration optimization across preset crossovers.
/eq EQ defaults, equaliser model selection, target settings, house curve, auto-match. Headless target-matching and biquad generation for DSP fleet provisioning.
/generator Signal generator selection (sine, PN, multitone, bursts) and level/state control. Scripted distortion, CEA-2010, or calibration tone playback.
/spl-meter SPL meter start/stop, weighting, and live SPL/Leq/SEL readings. Automated pre-sweep 75 dB SPL level verification and ambient noise-floor gating.
/rta RTA start/stop, levels, captured spectrum data, and distortion metrics. Live THD/THD+N/IMD/ENOB pass-fail logging for electronics production testing.
/stepped-measurement Stepped-sine distortion sweep configuration, triggering, and progress polling. High-precision low-distortion electronics sweeps where log-sweep noise floor is too high.
/import Import frequency response, impulse response, and offline sweep recordings. Batch-ingesting field-recorded sweep WAVs from remote studios into REW analysis.
/roomsim Room simulator parameters, positions, and simulated response retrieval. External Monte-Carlo or grid-search scripts optimizing multi-sub coordinates against modal flatness.
Remote Control Security & macOS Storage Paths

Because REW binds strictly to 127.0.0.1, control a Mac lab host remotely by forwarding the port over SSH (ssh -L 4735:127.0.0.1:4735 lab-mac) rather than exposing an unauthenticated proxy to the LAN. On macOS, REW preferences persist in ~/Library/Preferences/com.apple.java.util.prefs.plist and logs/temporary backups live in ~/Library/Logs/REW.

Data Interchange Import/Export Format Reference & Third-Party DSP Toolchain Notes
Format / Artifact Direction Key Technical Considerations
.frd / .zma / .txt Import & Export Plain-text frequency/magnitude/phase columns for VituixCAD, Boxsim, and crossover CAD. Verify whether phase is minimum-phase or measured excess-included before exporting.
.wav / .aiff / .pir / .pcm / ETF Import (IR) & Export (WAV) Impulse response exchange with ARTA (.pir) and convolution engines. Exported IRs reflect active windowing and sample rate.
Offline Sweep Recording Import Play REW’s generated measurement sweep file through a standalone streamer/DSP/mobile device, record on a separate recorder, and import into REW to deconvolve the transfer function.
Biquad Coefficients Export (EQ → Filter Tasks) Strictly sample-rate-specific. Confirm the target DSP’s internal rate (e.g., 48 kHz vs 96 kHz) in the Equaliser panel title before saving coefficients for miniDSP, SigmaStudio, or CamillaDSP.
Filter Settings Text Export Human- and parser-readable PEQ tables for Equalizer APO, CamillaDSP, rePhase, StormAudio, Hypex, or commissioning documentation. Verify Q vs bandwidth definition matches the selected equaliser model.

05 · End-to-End Lab Workflow

Worked Example: Dual-Mass T/S Extraction, Room Sim Verification & API Multi-Seat Capture

Engineering Goal: Characterize a raw 8-inch woofer in-house using Dual Added Mass T/S extraction to verify OEM specs and check amplifier EPDR stress once loaded into a vented cabinet, pre-screen dual-subwoofer placement in a 5.4 × 4.1 × 2.6 m control room using the Room Simulator, splice nearfield + farfield sweeps via Trace Arithmetic, and automate a repeatable 4-seat verification capture + data pull on macOS using the REW REST API.

Audio Rig (macOS): 48 kHz / 24-bit in Audio MIDI Setup · 4-input interface · 48 V OFF
Impedance Bridge: Measured $R_s = 100.14\,\Omega$ (0.1% non-inductive) · Reference Cal resistor = $10.02\,\Omega$
Driver Inputs: Measured $R_{dc} = 6.12\,\Omega$ · $S_d = 214.0\,\text{cm}^2$ · $m_1 = 22.40\,\text{g}$, $m_2 = 45.10\,\text{g}$
Sweep Config: 10–20,000 Hz · 512k (~10.7 s) · 150 mV RMS at driver terminals
Room Sim Geometry: $5.40 \times 4.10 \times 2.60\,\text{m}$ · Surface absorption 0.15–0.22 · Dual subs + 3 listening seats
Automation Host: macOS Terminal · http://127.0.0.1:4735 (REW Pro upgrade active)

Execution Steps

  1. Calibrate impedance bridge & capture three driver sweeps. In Measure → Impedance, enter 100.14 Ω for $R_s$, execute Open Circuit, Short Circuit, and Reference ($10.02\,\Omega$) calibrations. Clamp the 8-inch woofer rigidly in free air, set 150 mV terminal drive, and capture W8_FreeAir. Attach $m_1 = 22.40\,\text{g}$ symmetrically around the dust cap perimeter and capture W8_Mass1_22p4g; repeat with $m_2 = 45.10\,\text{g}$ for W8_Mass2_45p1g.
  2. Extract Dual Added Mass T/S parameters. Open Tools → Thiele Small Parameters, select Dual Added Mass mode, assign all three traces, enter $R_{dc} = 6.12\,\Omega$, $S_d = 214.0\,\text{cm}^2$, $21.5\,^\circ\text{C}$, $1013\,\text{hPa}$, $22.40\,\text{g}$, and $45.10\,\text{g}$, then click Calculate Parameters. Verify the modelled impedance overlay against all three measured traces and export W8_FreeAir.zma.
  3. Pre-screen cabinet placement in Room Simulator. After loading the woofer/sub into its 42 L vented enclosure ($F_b = 34\,\text{Hz}$), open Room Simulator, input $5.40 \times 4.10 \times 2.60\,\text{m}$, enable two subwoofers plus three mic positions across the console width, compare single front-corner placement against opposing midpoint placement (with delay/polarity tuned), and push the simulated combined response into REW’s main graph panel.
  4. Splice nearfield & gated farfield response. Measure the finished speaker at 1 m on-axis (apply a 5.0 ms right IR window to strip early room reflections, valid above ~200 Hz) and take a nearfield woofer + port measurement. Open All SPL → Actions → Trace Arithmetic, select Merge B to A around 250 Hz to automatically level- and phase-stitch the nearfield LF curve to the gated farfield curve.
  5. Automate multi-seat verification and JSON export via the REW REST API. Launch REW on macOS with -api and run the verification script below to assert audio state, trigger sweeps, pull measurement metadata, and check room-sim/measurement alignment:
macOS Automation — REW REST API Verification Pipeline (bash + curl + jq)
#!/usr/bin/env bash
set -euo pipefail

BASE="http://127.0.0.1:4735"

# 1. Confirm sample rate and audio configuration before sweeping
echo "==> Checking audio interface configuration..."
curl -sf "$BASE/audio/samplerate" | jq .

# 2. Configure sweep metadata and level via /measure (Pro upgrade required for control)
curl -sf -X POST "$BASE/measure/naming" \
  -H "Content-Type: application/json" \
  -d '{"title":"ControlRoom_Seat1_Verify","notes":"48kHz 24-bit DualSub Midpoint"}' | jq .

curl -sf -X POST "$BASE/measure/sweep/configuration" \
  -H "Content-Type: application/json" \
  -d '{"startFreq":15,"endFreq":20000,"length":262144,"sweepLevel":-12.0}' | jq .

# 3. Trigger an SPL measurement sweep via /measure/command
echo "==> Running automated SPL sweep..."
curl -sf -X POST "$BASE/measure/command" \
  -H "Content-Type: application/json" \
  -d '{"command":"SPL"}' | jq .

# 4. List loaded measurements and inspect the newest measurement UUID/index
echo "==> Fetching measurement index..."
curl -sf "$BASE/measurements" | jq 'to_entries | map({id: .key, title: .value.title, sampleRate: .value.sampleRate})'

# 5. Pull Room Simulator configuration/state for regression comparison
echo "==> Reading Room Simulator state..."
curl -sf "$BASE/roomsim" | jq .

Expected Output

  • Dual Added Mass T/S Result: $F_s = 34.2\,\text{Hz}$ (shifting to $26.2\,\text{Hz}$ with $m_1$ and $21.9\,\text{Hz}$ with $m_2$), $Q_{ms} = 5.40$, $Q_{es} = 0.41$, $Q_{ts} = 0.38$, $M_{ms} = 31.6\,\text{g}$, $C_{ms} = 0.686\,\text{mm/N}$, $V_{as} = 46.8\,\text{L}$, $R_{ms} = 1.26\,\text{kg/s}$, $Bl = 10.2\,\text{N/A}$. Modelled impedance tracks all three curves within $\pm 1.5\%$ across $15\text{–}300\,\text{Hz}$.
  • Vented Cabinet Load Check: Twin peaks at $21.5\,\text{Hz}$ and $54.0\,\text{Hz}$ with saddle minimum $6.4\,\Omega$ at $F_b = 34.0\,\text{Hz}$ ($\theta = 0^\circ$). Minimum EPDR reads $3.1\,\Omega$ at $118\,\text{Hz}$ where $|Z| = 6.9\,\Omega$ coincides with $-41^\circ$ capacitive phase—safe for any stable 4 Ω-rated amplifier.
  • Room Sim & Multi-Seat Verification: Moving from a single front corner to opposing wall midpoints eliminates a predicted $14\,\text{dB}$ seat-to-seat null at $63\,\text{Hz}$ (2nd axial length mode). Generating an RMS average of the 4 captured console seats in All SPL confirms $\pm 3.5\,\text{dB}$ seat-to-seat consistency from $30\text{–}120\,\text{Hz}$ before applying any modal EQ.
  • API Telemetry: GET /audio/samplerate confirms 48000, POST /measure/command completes cleanly with zero clipping warnings in /application, and GET /measurements returns structured JSON ready for automated tolerance checks.

Professional Interpretation

Notice how each subsystem cross-validates the next: the Dual Added Mass extraction proved the driver’s real $F_s$ and $V_{as}$ matched the 42 L cabinet design without relying on unverified OEM numbers; the EPDR curve ruled out hidden reactive amplifier stress; the Room Simulator identified the exact physical placement needed to cure the $63\,\text{Hz}$ cancellation (which EQ could never fix); Merge B to A isolated true quasi-anechoic speaker response without an anechoic chamber; and the REST API turned multi-seat verification into a deterministic, auditable script.

06 · Verification

Check Your Understanding

Test your diagnostic judgment on impedance rigs, Thiele-Small/EPDR analysis, trace math, and macOS API automation. Select the most technically accurate answer for each scenario.

Question 1 of 4 · Impedance Rig Diagnostics

Swapped Bridge Channels

You wire an impedance jig using a measured $100.0\,\Omega$ non-inductive sense resistor to test an $8\,\Omega$ nominal woofer ($R_{dc} = 6.2\,\Omega$). Immediately after sweeping, REW plots a baseline impedance sitting around $106\,\Omega$ across the band and the phase curve appears flipped by roughly $180^\circ$. What is the root cause?

Question 2 of 4 · Thiele-Small & Amplifier Load

Dual Added Mass & EPDR Interpretation

Which statement accurately describes both REW’s Thiele-Small extraction accuracy and how EPDR should be interpreted on a completed loudspeaker’s impedance plot?

Question 3 of 4 · Averaging & Trace Arithmetic

Spatial Averaging vs. Quasi-Anechoic Splicing

You have (1) six room measurements captured at different seats across a mixing console for low-frequency modal EQ, and (2) a nearfield woofer sweep plus a 5 ms gated farfield sweep of the same loudspeaker to build a quasi-anechoic curve. Which pair of REW operations is technically correct?

Question 4 of 4 · REW REST API & macOS Deployment

Automating REW on macOS

You are deploying an automated REW test script on a Mac Studio. Which statement accurately reflects REW’s API networking, licensing, and macOS launch behavior?