--> Lesson 1 — Audio I/O and Calibration on macOS | REW Pro Reference

Lesson 01 · Signal Integrity & Platform Architecture

Lesson 1 — Audio I/O and Calibration on macOS

Establish a bit-transparent, sample-rate-locked, SPL-traceable measurement chain on macOS before firing a single sweep. Covers CoreAudio-to-Java Sound bridging, Audio MIDI Setup discipline, loopback and microphone calibration mechanics, headroom engineering, tccutil permission recovery, and exact Windows-only feature boundaries.

Platform: macOS 10.15+ (Bundled JRE) Driver Stack: Java Sound → CoreAudio Target Headroom: 6–40 dB (Green) Scope: Acoustic & Electronic I/O

1. What This Is For

Every downstream calculation in REW—impulse deconvolution, minimum/excess phase extraction, distortion separation, RT60 Schroeder integration, and target-matched biquad synthesis—is mathematically conditioned on the integrity of the raw digital stream entering REW’s Java audio layer. If the macOS audio stack silently resamples your interface, drops microphone permissions after a point update, runs mismatched calibration files, or operates too close to either 0 dBFS or the ADC noise floor, REW will still render authoritative-looking graphs that are systematically wrong.

For an experienced operator migrating between Windows rigs and macOS laptops—or commissioning a repeatable Mac-based studio/field rig—this lesson locks down the four failure domains unique to macOS I/O and calibration:

  • OS-level clock and bit-depth lock: Unlike Windows ASIO or WASAPI Exclusive (EXCL:) streams that seize hardware parameters directly, REW on macOS relies strictly on Java audio drivers layered over CoreAudio. If Audio MIDI Setup is left at 16-bit or a mismatched sample rate prior to launching REW, you sacrifice dynamic range or induce silent CoreAudio sample-rate conversion.
  • Repeatable gain staging without in-app faders: REW’s internal volume controls are intentionally disabled on macOS because CoreAudio exposes discrete, device-dependent dB steps. Without disciplined external gain control (interface preamp/mixer or Audio MIDI Setup) and strict recalibration discipline, absolute SPL values drift between sessions.
  • Clean calibration boundaries: Analog XLR measurement paths require sample-rate-specific loopback calibration plus an external SPL reference (or calibrated sensitivity header); USB measurement microphones carry their own ADC and embedded sensitivity factor (Sens Factor = ...) and must never be paired with a soundcard loopback calibration file.
  • Zero-surprise macOS TCC and driver limits: Knowing how macOS Transparency, Consent, and Control (TCC) breaks mic capture silently after updates—and knowing upfront which Windows-only features (ASIO, WASAPI exclusive, RS232 serial, safe-by-default MIDI) do not exist on macOS—eliminates lost billable time on site.
macOS Runtime Architecture Baseline

REW requires macOS 10.15 or later, minimum 1024 × 768 display, and 2 GB RAM minimum (4 GB+ recommended). The macOS installer bundles a private Java runtime inside REW.app; do not install a separate system JRE unless a specific workflow demands it, and never point REW at a headless Java runtime because audio I/O subsystems are stripped from headless builds.

2. Concrete Operational Steps

Execute these phases in strict sequence. Hardware enumeration, TCC permission checks, and Audio MIDI Setup sample-rate locks must all occur before REW initializes its Java audio subsystem.

Phase A — Hardware Enumeration & macOS Mic Permission Verification

  1. Connect all measurement hardware before launching REW. Attach your USB measurement microphone or audio interface directly (avoiding unpowered hubs where possible) and power on external preamps. REW enumerates Java audio devices at startup; hot-plugging after launch frequently yields stale device lists or silent input streams.
  2. Verify or reset macOS TCC microphone permissions if input is silent. Microphone entitlement is embedded in REW’s code signature and should prompt automatically on first run. However, macOS OS updates, REW version upgrades, or a previously dismissed dialog can leave REW capturing zeros (flat digital silence) while meters appear active. Quit REW completely, open Terminal, and force a fresh TCC prompt:
    Terminal (run BEFORE starting REW)
    tccutil reset Microphone
  3. Confirm System Settings override if capture still fails. Navigate to System Settings → Privacy & Security → Microphone, verify the toggle for REW is enabled, then fully quit (⌘+Q) and relaunch REW. Toggling the permission while REW is running does not rebind an already-denied audio unit.

Phase B — Lock Sample Rate, Bit Depth, and Channels in Audio MIDI Setup

  1. Launch macOS Audio MIDI Setup (/Applications/Utilities/Audio MIDI Setup.app) before starting REW.
  2. Set the exact sample rate on both input and output devices. Select 48,000 Hz (or 44,100 Hz) for standard acoustic measurements unless you explicitly require bandwidth beyond 20 kHz (e.g., ultrasonic transducer or electronics testing). Higher rates (up to 192 kHz historically; up to 1536 kHz selectable in REW V5.30+ where hardware permits) inflate FFT memory footprint and processing overhead without automatically improving acoustic accuracy.
  3. Set format bit depth to 24-bit (or 32-bit float/integer where offered). macOS Java drivers natively pass 24-bit or 32-bit audio when the interface is configured for it in Audio MIDI Setup. Leaving an interface at 16-bit wastes ~48 dB of theoretical quantization headroom and degrades low-level tail decay and distortion noise-floor visibility.
  4. Configure multichannel speaker layouts if measuring surround/immersive systems. Click Configure Speakers… in Audio MIDI Setup and set the interface to its multichannel topology so CoreAudio exposes all discrete output channels to REW’s Java driver.
  5. Establish initial hardware gain in Audio MIDI Setup or the interface’s native DSP mixer. Because REW’s own volume sliders are disabled on macOS, disable any direct hardware input monitoring (zero-latency pass-through) on your interface mixer right now to prevent acoustic or loopback feedback.

Phase C — Configure REW Preferences → Soundcard

  1. Launch REW and open Preferences (⌘+,) → Soundcard. Verify the Driver dropdown shows Java (on macOS, Java drivers are the sole driver model).
  2. Explicitly select physical Output Device and Input Device—never leave either on Default Device. Leaving either set to Default Device exposes your session to macOS system sound routing changes (Bluetooth handoffs, HDMI display wakes, or AirPlay switches) and obscures channel enumeration. Note that Java drivers permit input and output to reside on different devices (e.g., HDMI/interface output + USB mic input), though this asynchronous split carries specific timing/averaging constraints detailed below.
  3. Select the exact Sample Rate matching Audio MIDI Setup (e.g., 48 kHz) and assign the physical Output channel, Input channel, and (if using a wired loopback timing reference) the dedicated Ref Input / timing reference channel.
  4. Map multichannel outputs if physical jacks do not match standard labels. CoreAudio multichannel interfaces often expose channels in hardware-specific indices that diverge from REW’s standard L/R/C/LFE/SL/SR naming. Open Output Channel Mapping in Preferences → Soundcard and verify each logical channel against physical speaker output with pink noise or channel ID tones.
  5. Tune input/output buffer sizes only if dropouts occur. If sweeps exhibit audible ticks, pops, or discontinuities in the Captured graph, increase the Input/Output buffer size in Soundcard Preferences until glitch-free capture is verified.

Phase D — Soundcard Loopback Calibration (Analog Interface / Analog Mic Path Only)

Hardware & Feedback Safety Gate — Check Before Patching Loopback

1. Turn OFF +48V phantom power on the measurement input channel before connecting a line output directly to a mic/line input. Applying +48V into an unprotected output stage can permanently damage DAC output buffers.
2. Mute all direct/software monitoring so Line In is not routed back to the interface Line Out. An open monitoring path creates an immediate full-scale digital/analog oscillation loop.
3. Skip Phase D completely if using a USB measurement microphone. A USB mic has its own integrated ADC; a soundcard loopback calibration characterizes the interface’s ADC, not the USB mic’s ADC, and would inject invalid correction into acoustic captures.

  1. Patch a direct electrical loopback from the interface’s chosen measurement Line Output to the chosen measurement Line Input (using a clean balanced/unbalanced patch cable, matching levels appropriately).
  2. In Preferences → Soundcard, click Calibrate Soundcard… Follow REW’s level-check prompt to verify strong, unclipped loopback signal (no feedback, healthy green headroom), then run the loopback calibration sweep.
  3. Inspect the generated Soundcard Cal curve and save it with the sample rate in the filename (e.g., InterfaceName_48kHz_Loopback.cal). Soundcard calibration is strictly sample-rate-specific because ADC/DAC anti-aliasing and reconstruction filter roll-offs shift with fs. If you later switch from 48 kHz to 96 kHz, you must generate and load a 96 kHz soundcard cal file.
  4. Remove the loopback cable before re-enabling phantom power and connecting your analog measurement microphone.

Phase E — Microphone Calibration File & Absolute SPL Calibration

  1. Load the microphone calibration file in Preferences → Cal Files (under Mic/Meter Calibration). Verify you load the file corresponding to your physical mic orientation (e.g., 90° vertical grazing incidence is standard for room/subwoofer omni work; 0° on-axis for direct nearfield/anechoic speaker work).
  2. Enable Limit cal data boost to 20 dB (unless doing specialized known-bandwidth transducer characterization) so aggressive low- or high-frequency inverse corrections at the edge of the mic’s passband do not artificially lift the ADC noise floor by 30+ dB.
  3. Verify sensitivity handling (USB vs Analog):
    • USB mic with calibration header: If the file contains a valid Sens Factor = ... line, REW automatically reads the mic’s calibrated sensitivity and sets SPL scaling—external SPL meter calibration is normally unnecessary.
    • Analog mic / unreferenced USB mic: Either supply an analog sensitivity header (Sensitivity -12.34 dBFS) if rigorously characterized at a fixed preamp detent, or open Preferences → Soundcard → Calibrate SPL… (or the SPL Meter → Calibrate button).
  4. Run external SPL calibration (when required): Select Use REW speaker cal signal (band-limited pink noise) or an external calibrator tone (e.g., 94 dB / 114 dB 1 kHz pistonphone). Position the reference SPL meter (typically set to C-weighted, Slow; check REW’s C-weighted SPL meter option if using a C-weighted meter whose weighting curve needs accounting) directly adjacent to the measurement capsule, enter the meter’s exact reading into REW, and confirm.
  5. For electronic-only distortion/response measurements (no microphone): Enter your interface’s exact full-scale input voltage in FS sine Vrms (in Preferences → Soundcard / Cal) so RTA, Scope, and Distortion panels read accurate dBV / dBu / Vrms values.

Phase F — Gain Staging & Pre-Sweep Level Verification

  1. Open Measure (⌘+M) and click Check Levels (or open REW’s Level Meters + Signal Generator with pink noise).
  2. Set playback SPL to approximately 75 dB SPL at the listening position using your AV processor / monitor controller volume and/or REW’s sweep level (-12 dBFS is a reliable starting digital output level). Maintain consistent sweep output level across all measurements you intend to compare directly.
  3. Trim interface preamp / Audio MIDI Setup input gain so the analog input sits squarely in the -30 dBFS to -12 dBFS RMS zone during pink-noise playback.
  4. Confirm Measurement Headroom reads green—ideally between 6 dB and 40 dB. Once locked, do not touch the input gain knob or Audio MIDI Setup input slider; any change invalidates your SPL calibration until you recalibrate.
Calibration Invalidation Checklist

You must recalibrate SPL (and/or regenerate soundcard cal) whenever any of the following changes: (1) analog mic preamp gain or Audio MIDI Setup input volume, (2) interface sample rate, (3) physical input channel or pad switch, (4) switching between a USB mic and an analog XLR mic. A REW measurement is strictly valid only for the exact sample rate, bit depth, gain state, and calibration files active at capture time.

3. Reading the Output

Do not trust a green "OK" passively. Before recording a full multi-position room survey or tuning an active crossover, audit these six REW telemetry surfaces to verify bit depth, headroom margins, calibration integrity, and hardware clock stability.

Signal Chain Gain-Staging & Headroom Architecture (macOS CoreAudio → REW) Level Meters & Check Levels
HIGH HEADROOM (>40 dB) Poor S/N · ADC & room noise intrusion OPTIMAL HEADROOM WINDOW (6–40 dB) Analog RMS Target: -30 to -12 dBFS <6 dB Peak Clip Risk CLIP -90 dBFS -40 dBFS -30 dBFS -12 dBFS -6 dBFS 0 dBFS Pink Noise RMS (-22 dBFS @ 75 dB SPL) Modal Peak (+12 dB crest → -10 dBFS) Note: Room modes can add +10 to +18 dB above broadband pink-noise RMS during a log sine sweep. Reserve ≥ 12–18 dB peak headroom during Check Levels so narrowband modal resonances never clip 0 dBFS mid-sweep.

Telemetry Panels & Trace Interpretation

REW View / Readout What It Actually Measures Healthy Reference State Failure Signature & Root Cause
Main Window Status Bar
(Bottom Edge)
Active Java driver sample rate and negotiated bit depth handed off from CoreAudio. Exact match to Audio MIDI Setup (e.g., 48 kHz 24-bit or 32-bit). Shows 16-bit or unexpected 44.1 kHz when interface is 48 kHz → Audio MIDI Setup was not locked before launching REW, or wrong device selected.
Level Meters
(In / Out RMS & Peak)
Live digital input/output RMS and peak amplitude in dBFS (or calibrated SPL/Vrms). Up to 32 inputs with Pro upgrade. Output at -12 dBFS (nominal sweep); Analog Input RMS at -30 to -12 dBFS; ambient floor ≥ 30–40 dB below test signal. Flatline -inf / zero activity despite preamp LEDs lighting → macOS TCC blocked microphone (run tccutil reset Microphone) or wrong input channel index selected.
Measurement Headroom
(Check Levels & Post-Sweep)
Distance in dB between highest captured sample peak and 0 dBFS digital full scale. Green: 6–40 dB. Sweet spot for room sweeps with strong bass modes is typically 12–25 dB. < 6 dB / Red (0 dB): Hard clipping corrupts FFT harmonic separation and invalidates SPL/distortion.
> 40 dB: Insufficient input gain; decay tails and RT60 T20/T30 hit noise floor prematurely.
Captured Graph
(Main Graph Tab)
Time-domain waveform of the last generated sweep stimulus overlaid with the uncorrected raw captured input. Not live; unavailable for loaded .mdat files. Smoothly tapered envelope following room/speaker response, clean baseline before/after sweep, consistent acoustic/system latency offset. Flat-topped square ceilings = ADC clipping. Sudden zero-value notches = CoreAudio buffer dropout (increase buffer size). Growing runaway amplitude = monitoring feedback loop.
Mic/Meter Cal & Soundcard Cal Traces
(SPL & Phase Graph)
Exact inverse correction curves REW applies to the raw transfer function. Soundcard cal flat within ±0.2–0.5 dB across 20 Hz–20 kHz (rolling off smoothly near DC/Nyquist); Mic cal matches manufacturer datasheet. Massive vertical spike clamped at +20 dB at extreme LF/HF = mic/interface roll-off hitting Limit cal data boost to 20 dB. Wild jagged ripples in Soundcard Cal = loopback ran through an acoustic path or active DSP/EQ!
Measurement Info Panel
(Info Button / ⌘+I)
Immutable audit record attached to each capture: sample rate, timing reference, system delay, loaded cal files, sweep settings, and notes. Lists exact expected mic cal filename, soundcard cal filename (or None for USB mic), consistent sample rate, and repeatable system delay. Missing cal file name, unexpected sample rate, or None for timing reference when phase comparison is needed → discard capture before analysis.
Diagnostic Signatures — Captured Graph & Calibration Trace Overlays Captured & SPL Cal Audit
CAPTURED GRAPH (Time-Domain Raw Input) +1.0 (0 dBFS) Clipped Flat-Top Buffer Dropout Green: Clean sweep capture with latency offset SPL & PHASE — CAL TRACE OVERLAYS +20 dB Cal Boost Limit 0 dB Soundcard Cal (Flat ±0.2 dB in passband) Clamped LF roll-off Verify Mic Cal + Soundcard Cal before interpreting low-end SPL
Valid Session Baseline: Status bar matches Audio MIDI Setup (24-bit / 48 kHz); explicit physical I/O selected (not Default); Headroom 12–28 dB green; Captured trace shows clean tapered sweep without flat-tops or gaps; Measurement Info records exact mic cal and (for analog only) sample-rate-matched soundcard cal.
Compromised Session — Stop & Fix: Input reads digital zero (TCC blocked or wrong channel); Soundcard Cal shows >3 dB ripple across midrange (active EQ/monitoring or acoustic leak during loopback); USB mic loaded simultaneously with a Soundcard Cal file; preamp knob moved after SPL calibration without re-calibrating.

4. Deeper Detail

Expand these technical reference panels for exact macOS vs. Windows driver boundaries, high-sample-rate CoreAudio behavior, calibration file syntax, and CLI/filesystem internals.

Platform Architecture · Compatibility Matrix Unavailable Windows-Only Features & macOS-Specific Operational Quirks

Professionals running mixed OS fleets must design workflows around hard platform boundaries. Several driver paths and legacy hardware interfaces documented in the REW manual do not exist on macOS:

Feature / Subsystem Windows Status macOS Status & Engineering Impact
ASIO Drivers Supported (low-latency direct hardware & native multichannel) Unavailable (Windows-only). REW uses Java audio drivers exclusively over CoreAudio. Multichannel interfaces expose channels through CoreAudio + Audio MIDI Setup + REW Output Channel Mapping.
WASAPI Exclusive (EXCL:) Supported via Java EXCL: driver entries (including multichannel) Unavailable (Windows-only). There are no EXCL: devices on macOS; sample rate and bit depth must be locked externally in Audio MIDI Setup prior to launch.
RS232 Serial Control Supported on Windows 32-bit REW builds only Unavailable. Includes TAG McLaren AV32R DP / AV192R serial processor control and any equaliser paths dependent on Windows-only drivers.
MIDI Hardware Comms
(e.g., BFD Pro DSP1124P, FBQ2496)
Available Disabled by default on macOS because Java MIDI enumeration can crash the JRE on certain Macs. Can be toggled on in Preferences → Comms (may work on JRE V6+), but treat as an unstable path; prefer network/file biquad exports where supported.
In-App Volume Sliders Supported on compatible drivers Disabled on macOS. CoreAudio constrains volume changes to discrete, device-dependent dB increments. Control all gain in Audio MIDI Setup or the interface’s hardware/DSP mixer.
Preferences Storage Windows Registry Stored as a standard macOS plist at:
~/Library/Preferences/com.apple.java.util.prefs.plist
UI / Dialog / Keyboard Behavior Ctrl/Alt shortcuts; manual DPI font scaling
  • Shortcuts map to ⌘ (Command) and ⌥ (Option).
  • Display DPI font scaling option is absent because macOS handles Retina/HiDPI scaling automatically.
  • Use macOS file dialogs option exists, but the native chooser lacks REW’s measurement preview panel and has occasionally failed to render on screen—if a dialog hangs or vanishes, revert to the Java file chooser.
  • An AZERTY keyboard checkbox is available to work around a Java shortcut-mapping bug on French layouts.
DSP & Clock Engineering · V5.30+ Sample Rate Limits (up to 1536 kHz), Nyquist Math & Asynchronous Clock-Domain Traps

Sample rate ceilings on macOS: Historically, macOS Java audio drivers topped out at 192 kHz. Starting with REW V5.30+, sample rates up to 1536 kHz may appear selectable on macOS, subject entirely to interface firmware and host hardware capability (for example, an M1 Mac mini has reportedly operated cleanly at 768 kHz). In all cases, the target sample rate must be selected in Audio MIDI Setup prior to starting REW.

Why 44.1 kHz or 48 kHz remains optimal for acoustics: The highest representable frequency is Nyquist (fs / 2—24 kHz at 48 kHz). Frequencies above Nyquist cause aliasing unless suppressed by the interface’s anti-aliasing filter. Running 192 kHz or 384 kHz for room acoustics offers zero low-frequency resolution benefit—in fact, because FFT bin spacing is Δf = fs / FFT_length, raising fs by 4× actually makes frequency resolution 4× coarser unless you also quadruple FFT length, which multiplies RAM usage and convolution latency without improving room-mode accuracy. Reserve high sample rates strictly for ultrasonic driver measurements, high-order harmonic distortion sweeps beyond 20 kHz, or DAC filter characterization.

The Asynchronous Split-Device Trap (Java I/O): Because macOS Java drivers allow Output and Input to use completely separate hardware (e.g., an HDMI AVR output + a USB measurement mic), the DAC and ADC operate on independent crystal oscillators. Small clock-rate ppm drift between two unsynchronized crystals causes phase rotation over time. Consequently:

  • Do NOT use multiple-sweep averaging with a USB microphone or split input/output devices—clock drift between repetitions smears high-frequency coherent averaging and degrades the impulse response. Use a single longer sweep instead (doubling sweep length gains ~3 dB S/N cleanly).
  • Periodic noise RTA lock: In RTA mode, periodic pink/white noise eliminates windowing leakage only when generator and capture clocks are locked to the same hardware oscillator; split clocks reintroduce bin leakage unless windowed.
  • Always use an acoustic timing reference (or wired loopback when on a single synchronized interface) if you intend to compare phase or delay across multiple sweeps.
File Format Specification · Metrology Calibration File Syntax, Sensitivity Headers & Electronic Vrms Referencing

REW microphone, SPL meter, and soundcard calibration files are plain-text ASCII files containing space- or tab-delimited columns of Frequency (Hz), Gain (dB), and an optional third column for Phase (degrees). Comment lines are ignored except for recognized sensitivity metadata headers:

Example Calibration File Structures (USB vs Analog Sensitivity Headers)
* USB Measurement Mic Header (auto-configures SPL without external meter)
"Sens Factor =-1.42dB, SERNO: 7001234"
10.054  -3.42   0.00
20.000  -0.85   0.00
1000.0   0.00   0.00
20000.0 +1.18   0.00

* Analog Measurement Mic Header (valid only at fixed calibrated preamp gain)
Sensitivity -12.34 dBFS
20.000  -0.40
1000.0   0.00
20000.0 +0.65
  • Sens Factor = ... (USB Mics): Read directly by REW when loaded as a mic calibration file so SPL readings are referenced to 20 µPa automatically without needing an external SPL meter.
  • Sensitivity -12.34 dBFS (Analog Mics): Encodes the exact digital input level corresponding to a reference SPL (e.g., 94 dB SPL) for a fixed, detented preamp gain. If your interface lacks stepped/recallable preamp gain, rely on live SPL meter or acoustic calibrator verification instead.
  • Limit cal data boost to 20 dB: Prevents inverse calibration curves from applying >20 dB gain at extreme frequencies where the microphone or ADC has already rolled off—avoiding a deceptive +30 dB shelf of amplified quantization/thermal noise masquerading as acoustic output.
  • C-Weighted SPL Meter Option: When calibrating against a legacy handheld SPL meter that lacks flat (Z-weighted) response, enable REW’s C-weighted SPL meter compensation so REW applies the matching C-weighting curve when matching the band-limited speaker cal signal.
  • Electronic Measurements (FS sine Vrms): When testing preamps, crossovers, or DACs electrically, bypass acoustic SPL calibration and enter the ADC’s exact full-scale sine RMS voltage in FS sine Vrms (measured with a true-RMS multimeter at a clean 100 Hz–1 kHz sine tone). This converts all dBFS readouts to absolute Vrms, dBV, and dBu across RTA, Distortion, and Scope windows.
macOS Internals · CLI & Automation Logs, 7-Day Temp Crash Recovery, Plist Migration & Localhost API Launch

For repeatable studio workstations or automated QA test benches on macOS, REW exposes clean filesystem artifacts and CLI launch arguments:

macOS Terminal — Launch REW with API Server & Inspect Paths
# Start REW with local REST API on default port 4735 (localhost / 127.0.0.1 only)
open -a REW.app --args -api

# Start REW API on a custom port (e.g., multi-instance or port conflict avoidance)
open -a REW.app --args -api -port 4567

# Open REW Logs & 7-day temporary measurement recovery folder in Finder
open ~/Library/Logs/REW

# Inspect or back up REW macOS preferences plist
ls -la ~/Library/Preferences/com.apple.java.util.prefs.plist
  • Automatic 7-day crash recovery (/temp): REW continuously writes temporary copies of measurements into a /temp subfolder inside the REW logs folder in ~/Library/Logs, retaining them for 7 days. If a macOS power event or JRE crash terminates a session before saving an .mdat, recover your sweeps directly from this directory.
  • API Security & Licensing Boundary: The API binds strictly to localhost / 127.0.0.1 (default port 4735). Read-only GET endpoints are available out of the box; automated sweep measurement control via PUT/POST endpoints requires the REW Pro upgrade.
  • Clean Preference Reset: If a corrupted macOS preference (such as an invisible native file dialog or broken MIDI port entry) locks up REW, quit REW and back up/inspect ~/Library/Preferences/com.apple.java.util.prefs.plist before resetting.

5. Worked Example — Commissioning a Dual-Path macOS Measurement Rig After a macOS Update

Goal: Commission a MacBook Pro (Apple Silicon, macOS 14+) after a macOS point update for a studio calibration job using a 2-channel balanced USB audio interface (48 kHz / 24-bit) with an analog phantom-powered measurement microphone—and establish a verified backup profile for a calibrated USB measurement microphone—ensuring zero clipping, verified loopback flatness, and traceable 75 dB SPL calibration.

Host: MacBook Pro (macOS 14+, REW V5.30+)
Primary Path: 2-in/2-out Interface + Analog Omni Mic (+48V)
Backup Path: Calibrated USB Mic (Sens Factor header)
Target Format: 48 kHz / 24-bit · 75 dB SPL @ -12 dBFS out
  1. Hardware attach & post-update TCC clearance: Connect the USB audio interface before opening REW. Because the Mac just installed a macOS security update, preemptively run tccutil reset Microphone in Terminal so REW triggers a clean CoreAudio mic entitlement prompt rather than silently recording digital zeros.
  2. Audio MIDI Setup lock: Open Audio MIDI Setup, select the 2-in/2-out interface, and set both Input and Output streams to 48,000 Hz, 24-bit Integer. Open the interface’s native mixer utility and disable direct hardware input monitoring on Ch 1 & Ch 2. Confirm +48V phantom power is OFF.
  3. Soundcard Preferences binding: Launch REW, approve the macOS Microphone permission dialog, and open Preferences → Soundcard. Change Output and Input from Default Device to the explicit interface name. Confirm Sample Rate: 48 kHz, Output: L (Ch 1), Input: L (Ch 1). Verify the bottom-left REW status bar reads 48 kHz 24-bit.
  4. Loopback soundcard calibration (Analog Path): Patch balanced Line Out 1 directly into Line In 1. Click Calibrate Soundcard…, run the level check (adjusting Line In 1 preamp trim until headroom reads ~15 dB green with zero feedback), capture the loopback sweep, inspect the trace (flat ±0.15 dB from 20 Hz–20 kHz, -0.8 dB at 22 kHz anti-aliasing skirt), and save as StudioInterface_48kHz_Ch1.cal. Disconnect the loopback cable completely.
  5. Analog mic calibration & SPL referencing: Connect the analog XLR measurement mic to Input 1, enable +48V phantom power, and load the mic’s 90° calibration file in Preferences → Cal Files with Limit cal data boost to 20 dB checked. Open SPL Meter → Calibrate, play REW’s band-limited speaker cal pink noise through the left monitor, hold a C-weighted Slow SPL meter next to the capsule (enabling REW’s C-weighted SPL meter checkbox), and enter the measured 75.4 dB.
  6. Gain staging & Check Levels verification: Open Measure (⌘+M), set sweep output level to -12 dBFS, and click Check Levels. Adjust the monitor controller until REW reads ~75 dB SPL and verify Input 1 RMS sits at -21.5 dBFS with 18.5 dB green headroom. Run a verification sweep (20 Hz–20 kHz, 256k) and inspect the Captured graph and Measurement Info panel.
  7. Backup USB mic profile hygiene: Before switching to the USB mic for a quick spot-check, open Preferences → Cal Files and clear the Soundcard Calibration file, then select the USB mic as Input Device (keeping the interface as Output Device), load the USB mic’s 90° cal file containing Sens Factor = ..., verify SPL reads ~75 dB automatically without external meter calibration, and confirm Number of Sweeps = 1 (disabling multi-sweep averaging due to the split DAC/ADC clock domains).
Expected Output & Engineering Interpretation
  • Status Bar & Measurement Info: Both confirm 48 kHz 24-bit. Analog capture Info lists both StudioInterface_48kHz_Ch1.cal and the analog mic cal file; USB mic capture Info lists only the USB mic cal file and no soundcard cal.
  • SPL & Phase Cal Traces: Enabling the Soundcard Cal overlay shows a smooth ±0.15 dB passband with gentle filter roll-off near Nyquist; no comb-filtering ripples (proving zero monitoring bleed or acoustic leakage during loopback).
  • Captured Graph: Raw sweep capture reaches ~-10 dBFS at a +11 dB room mode peak (leaving 10 dB headroom below 0 dBFS—no flat-topping), zero buffer-dropout gaps, and clean noise floor >45 dB below sweep RMS.
  • Session Discipline: Any subsequent move of Input 1’s preamp knob or switch to 96 kHz immediately invalidates the analog profile; the engineer documents the preamp detent position in Measurement Notes so recalibration is deterministic.

6. Check Your Understanding

0 / 4 Answered

Diagnose four real-world macOS I/O and calibration failure scenarios. Select the single most technically accurate answer for each case.

Q1. Silent Capture After a macOS Upgrade

After updating macOS, your USB audio interface’s hardware preamp LEDs jump normally when you speak into the mic, and the interface is explicitly selected in REW’s Preferences → Soundcard at 48 kHz / 24-bit. Yet REW’s Level Meters sit dead flat at digital silence. What is the correct first remediation sequence?

Q2. USB Microphone + Loopback Calibration + Multi-Sweep Averaging

You switch from your analog XLR mic to a calibrated USB measurement microphone (whose cal file contains Sens Factor = -1.25 dB) while keeping your external DAC for sweep playback on macOS. You leave your 48 kHz interface Soundcard Cal file loaded and enable 4-sweep averaging to improve S/N. Which diagnosis is completely accurate?

Q3. Sample Rate Change, Headroom & Loopback Calibration Validity

Mid-session on an analog interface, you change your measurement setup from 48 kHz to 96 kHz to inspect a tweeter’s breakup resonance above 24 kHz. You change the rate inside REW’s Soundcard Preferences only, keep your 48 kHz soundcard cal file loaded, and notice Check Levels shows 3 dB headroom. What is wrong with proceeding?

Q4. macOS Platform Boundaries & Hardware Control

A client asks you to bring your MacBook Pro running REW to calibrate a legacy system with a TAG McLaren AV32R DP over RS232 serial, load filters over MIDI to a BFD Pro DSP1124P, and run an automated measurement script against REW’s API from a remote laptop on the same LAN. Which statement accurately reflects REW’s macOS constraints?