How can we help?

Installation

Download Spektral from the App Store. The app requires macOS 14.0 (Sonoma) or later on Mac, or iOS 17.0 or later on iPhone and iPad.

On first launch, your device will ask you to grant microphone access. Spektral needs this to capture audio for analysis. You can manage this permission at any time in System Settings > Privacy & Security > Microphone.

First Launch

When you open Spektral, audio analysis starts automatically. You will see two views stacked vertically:

• RTA (top) — Real-time spectrum showing frequency on the horizontal axis and level (dB) on the vertical axis.
• Spectrogram (bottom) — Scrolling time-frequency heatmap. Newest data appears at the bottom, scrolling upward. Bright horizontal lines indicate sustained frequencies — often feedback.

Selecting an Audio Input

macOS: Open Settings (Cmd+,) to choose your audio input device. Spektral lists all available Core Audio input devices, including USB interfaces, Thunderbolt devices, and the built-in microphone. The status bar at the bottom of the window shows the current device, sample rate, FFT size, and bin resolution.

Selecting an Audio Output (macOS): Directly below the input picker, the Output Device setting controls where Spektral plays its test signals — the pink noise generator and the STIPA signal. The default, Follow input device, routes the signal back through your input interface, which is correct for a single duplex interface (your mic and outputs on the same unit). If your output lives elsewhere — a separate interface, or your Mac's own output — choose it from the list. This is also how you loop Spektral's pink noise back as a reference for transfer-function delay and phase measurements.

iOS: Spektral uses the device's built-in microphone by default, or any connected audio interface.

RTA (Real-Time Analyser)

The RTA displays the current audio spectrum on a logarithmic frequency scale from 20Hz to 20kHz. The vertical axis shows level in decibels.

• The green trace is the live spectrum.
• The red trace (when enabled) shows peak hold — the maximum level reached at each frequency.
• The dashed white trace (when enabled) shows a frozen snapshot for comparison.

On macOS, hover your cursor over the RTA to see the frequency and dB level at that point. On iOS, tap to place the cursor. By default the cursor snaps to nearby spectral peaks for precise readings — turn Snap cursor to peaks off in Settings to let it free-run for pinpointing an exact frequency.

Understanding dB readings in the RTA

Uncalibrated, the RTA displays levels in dBFS (decibels relative to full scale), where 0 dB is the maximum level the digital audio system can represent. All readings are negative — for example, a moderately loud signal might show around −20 dBFS.

This is different from an SPL meter, which shows dB SPL (decibels relative to the threshold of hearing). The same sound might read −20 dBFS in the RTA but 70 dB SPL on a sound level meter — both are correct, they simply use different reference points.

For frequency analysis and ringing out a PA, dBFS is exactly what you need. What matters is the relative level between frequencies — “this peak is 15 dB above its neighbours” — and that relationship is the same regardless of the reference scale.

Once you calibrate against a reference meter or calibrator, the RTA's dB axis reads in true dB SPL, and the dedicated SPL Monitor page provides full metering on both macOS and iOS.

SPL overlay

Enable Show SPL overlay on RTA in Settings to float a live SPL card over the RTA: the current level in the centre with rolling Leq readouts either side. Tap either Leq readout to choose its window (1 to 60 minutes) without leaving the plot. When Limit Guard is enabled, the overlay also shows your headroom or an exact countdown to the limit.

Spectrogram

The spectrogram shows frequency over time using a thermal colour map. Cool colours (blue/cyan) indicate low levels; warm colours (yellow/red) indicate high levels.

Persistent horizontal bright lines typically indicate feedback frequencies or strong resonances — exactly what you need to find when ringing out a PA system.

Range controls: Adjust the floor and ceiling independently to focus the colour ramp on the level range that matters. Drag the spectrogram vertically to scroll back through history, or double-tap to return to live. You can also hide the spectrograph entirely in Settings for a full-height RTA.

Instant Replay

Heard something odd? Right-click anywhere on the spectrograph (long-press on iPhone) to replay the audio behind that moment — Spektral keeps a rolling recording of roughly the last minute, matched row-for-row to the display. Playback goes to your system's default output, so it works over headphones while the measurement continues. Scroll back through the history first to replay something that has already scrolled past.

Feedback Sentinel

The Sentinel watches the live spectrum for feedback as it forms. Click SENTINEL in the toolbar to arm it.

Feedback has a signature that music does not: a near-pure tone that grows over hundreds of milliseconds and then sustains. The Sentinel tracks every tonal component across frames and alerts when one is growing fast (feedback taking off) or ringing (grew earlier, now sustained). Because a track must have grown since it was first detected to alert, steady tones — mains hum, HVAC, held notes — cannot false-alarm.

• Live alerts — a growing ring is flagged directly on the RTA at its frequency, with a suggested notch: the cut depth from how far it stands above the spectral floor, and the Q from the ring's own measured bandwidth.
• Ring-out list — every confirmed ring is collected into a panel on the RTA with its frequency, musical note, suggested cut and Q, and a hit counter. A frequency that rings on three separate gain pushes is the one to notch first.
• Copy to console — the copy button puts the whole list on the clipboard, formatted for show notes or dialling into the console EQ. The bin button clears the session.
• iPhone — the Sentinel also fires a haptic tap when a new ring is confirmed, so you feel it while walking the room.

Transfer Function (macOS, and iOS with a dual-channel interface)

Press T or use the toolbar mode toggle to switch to Transfer Function view. This provides dual-channel H1 measurement showing:

• Magnitude — System frequency response in dB
• Phase — Phase relationship between input and output
• Coherence — Measurement confidence (1.0 = perfect correlation)

Auto-EQ Suggestions: Click the Auto-EQ button in Transfer Function mode to analyze your system response and receive up to 6 intelligent parametric EQ suggestions. Spektral uses 1/3 octave smoothing and targets musical frequency bands (80Hz, 125Hz, 250Hz, 500Hz, 1kHz, 2kHz, 4kHz, 8kHz) with musically appropriate Q values. Gain suggestions are capped at ±12dB with additional high-frequency protection to prevent harsh over-brightening. Perfect for getting a rough system tune before fine adjustments.

Comparing with Other Analysers

If you run Spektral side by side with another dual-channel analyser and the phase or coherence traces look rougher in Spektral, compare the analysis settings before suspecting the measurement — three controls dominate how "tidy" a transfer function looks, and different analysers ship very different defaults.

• FFT size — Longer FFTs give finer frequency resolution. Some analysers express FFT length as a power of two (2¹⁴ = 16,384 points, 5.9 Hz bins at 96 kHz); Spektral's default 4,096 trades resolution for display speed (23.4 Hz bins at 96 kHz). In a reflective room the response has dense comb structure, and a wide bin averages straight across several comb notches — which collapses that bin's coherence and turns high-frequency phase into hash. For measurement work, set FFT size to 16,384 in Settings, and check the other analyser's window length matches.
• Averaging — Coherence is a statistical estimate: its stability is a direct function of how much averaging feeds it. Many analysers integrate over half a second or more by default; Spektral's Avg control at OFF keeps only a light baseline so the display stays responsive. Set Avg to 16 (or infinite) for a comparably smooth, stable trace. A tidy coherence curve in any analyser is heavy averaging at work, not better data.
• Coherence blanking — The Coh slider in the Transfer toolbar blanks the magnitude and phase traces wherever coherence falls below the threshold, so untrustworthy regions simply don't draw. Around 0.40 is a good starting point. At 0 everything draws, including the noise.

A like-for-like starting recipe: FFT 16,384 · Avg 16 · Smoothing 1/6 · Coh 0.40. With matched settings the delay readouts and traces should agree closely between analysers.

Microphone position matters more than any setting. Coherence reflects the ratio of direct sound to room reflections at the microphone. Moving closer to the source raises that ratio and genuinely improves the measurement in every analyser — two analysers on the same rig are seeing the same room; they just present it through different amounts of smoothing.

Impulse Response & Delay

Spektral measures the propagation delay between your reference and measurement signals using FFT-based cross-correlation, and locks onto it automatically for coherent transfer function measurement. The Delay button in the action strip opens the delay controls, and shows the current locked delay in milliseconds.

Smart delay finding: Rather than simply picking the loudest peak in the cross-correlation — which can be a strong reflection or boundary bounce rather than the direct sound — Spektral scans for the first significant peak in time. The direct sound always arrives first, so this locks onto it correctly even in reflective rooms where a later reflection is louder.

Manual delay: When the room defeats the auto-finder, open the Delay popover and switch to Manual. Type the delay in milliseconds, or nudge it by ±1 ms or ±1 sample. You can also click anywhere on the impulse response strip to set the delay to that lag.

Impulse response view: Press I (or toggle it from the Delay popover) to show the impulse response strip beneath the magnitude and phase displays. Two display scales are available:

• Linear — rectified amplitude, with the direct sound as a single tall spike.
• ETC (dB) — the Energy Time Curve, 20·log₁₀|IR| over a 60 dB range. The direct sound sits at 0 dB and reflections cascade clearly down the scale, visible against the noise floor. This is the view for diagnosing reflections.

Reflection labelling: Spektral automatically detects the strongest reflections after the direct sound and labels them R1, R2, R3 on the impulse response. Each is annotated with its arrival time after the direct sound (+ms), the equivalent extra distance travelled (m), and its level relative to the direct sound (−dB) — helping you identify which surface is causing it.

Comb-filter prediction: Every reflection causes comb filtering — a series of regularly-spaced notches in the frequency response. Press M (or click the Combs button) and Spektral overlays the predicted notch frequencies from the strongest reflection directly onto the magnitude trace. When the predicted notches line up with the dips you're measuring, you've confirmed the cause: that notch is the reflection in your impulse response, not the loudspeaker.

Polarity: If the direct-sound peak in the cross-correlation is inverted, a POLARITY indicator appears in the impulse response header — meaning your measurement and reference are 180° out of phase at the direct arrival.

Multi-Position Capture Sessions

Capture transfer function measurements at up to 12 listening positions to build a comprehensive picture of your system's response across the coverage area. Captures live in a sidebar (the Captures button in the action strip), each with its own colour, and are quality-scored automatically based on coherence, anomaly detection, and drift from the running average.

• Capture — Click Take Capture in the sidebar (or press Cmd+G) to freeze the current measurement. Name each capture by position (e.g. "FOH Centre", "Balcony Left").
• Quality scoring — Each capture shows a quality indicator. Low coherence, anomalous nulls, or significant drift from the average are flagged automatically.
• Include/exclude — Toggle individual captures in or out of the coherence-weighted average to remove outliers. Hover a row to highlight its trace on the plot.
• Display traces — Show or hide the Average, the Live trace (hide it to compare captures without distraction), and the Position spread band, which shades how much the positions disagree.
• Autosave — The session saves automatically and restores when you reopen Spektral, so an end-of-gig quit never loses your captures.
• Relative delay reference — Lock to a reference delay for alignment work across positions.
• PDF export — Export the full session as a report with an overview page, averaged magnitude plot, and per-capture detail pages showing magnitude and coherence traces.

Delay Alignment Wizard

The Delay Alignment wizard guides you through measuring relative delays between your main PA and delay speakers. It calculates sample-accurate timing values you can enter directly into your processor, with temperature-compensated distance calculations.

Setup:

1. Open Transfer Function mode and click the Alignment button
2. Enter the venue name (optional) and ambient temperature — this is used to calculate the speed of sound accurately
3. Name your zones. The first zone is your reference (typically the main PA). Add delay zones with the + button and name them (e.g. "Delay 1", "Front Fill")
4. Click Begin Alignment

Capture:

1. For each zone, Spektral waits for the delay finder to lock, then captures transfer function frames
2. Play pink noise through only the zone being measured
3. Position the measurement microphone at the listening position
4. The wizard shows live coherence during capture — green is good (≥0.85), orange is acceptable, red indicates poor measurement conditions
5. Review each capture before moving to the next zone. You can accept, retry, or skip

Results: When all zones are measured, the summary table shows:

• Relative delay — milliseconds relative to the reference zone
• Sample values — delay in samples at both 48kHz and 96kHz, ready to enter into your processor
• Distance — calculated distance in metres and feet, using temperature-compensated speed of sound
• Polarity — whether the zone is normal or inverted relative to the reference
• Coherence — measurement confidence for each zone

Click Copy as Text to copy the results table to the clipboard for pasting into show notes or a system design document.

Note: Transfer Function mode requires a dual-channel audio interface with both reference (loopback) and measurement (mic) inputs. On macOS, the measurement side can alternatively come from an iPhone running Spektral — see iPhone Wireless Probe below.

Selecting Input Channels: By default, Spektral uses input 1 as the measurement channel and input 2 as the reference. On interfaces with more inputs you can route either to any channel in Settings > Audio Input — choose the Measurement channel (your mic: feeds the RTA, SPL metering, and the measurement side of the transfer function) and the Reference channel (the loopback). Picking the channel the other is using swaps the two. The status bar shows the active routing, e.g. CH 5→6. The RTA and SPL displays follow the measurement channel only, so other inputs on the device never pollute the spectrum.

iPhone Wireless Probe

The Wireless Probe streams your iPhone's microphone to Spektral on your Mac over the local network — turning the phone into a walking measurement mic. It works peer to peer (Wi-Fi or direct), with no servers and nothing leaving your network.

Setup

1. On the Mac: open Settings > Wireless Probe and enable Listen for iPhone probe.
2. On the iPhone: open Spektral's settings, find your Mac under Mac Link, and tap it. When it shows connected, audio flows automatically — a streaming pill on the phone's main view confirms it (tap the pill to stop).
3. On the Mac, the probe's spectrum overlays the RTA as a violet trace at the probe's own sample rate — walk the room and watch coverage change live.

Probe as the transfer function measurement mic

In Transfer Function mode, click Probe Mic in the action strip to use the connected iPhone as the measurement input. The reference keeps its loopback input on the interface; the phone becomes the mic. Walk the room, capture positions, and run the Delay Alignment wizard with no microphone cable run.

• Delays read as relative. The wireless link adds latency that is common to every measurement, so delays between positions (the alignment wizard's job) are valid, while the absolute delay number includes the link and is marked accordingly in the readout.
• Sample rates must match. The iPhone runs at 48 kHz, so the interface must be at 48 kHz too; the button explains if not.
• Brief network hiccups pause the trace and recover automatically — the measurement stream is never corrupted.

SPL Monitor (macOS & iOS)

The SPL Monitor mode provides dedicated environmental noise compliance monitoring for live sound events. Select SPL in the mode picker to switch to it. This mode is optimized for long-term measurement sessions and regulatory reporting.

Features:

• Live SPL Display — Large real-time sound pressure level readout with color-coded VU meter showing current level and user-configurable caution/warning thresholds
• Frequency Weighting — A-weighting (default), C-weighting, or Z-weighting (flat) based on IEC 61672 methodology for environmental noise monitoring
• Time Weighting — Slow (1s), Fast (125ms), and Impulse (35ms) time constants based on IEC 61672 methodology. Choose Slow for steady-state monitoring, Fast for general measurement, or Impulse for capturing transient peaks
• Microphone Calibration — Calibrate via acoustic calibrator (94dB/114dB) or reference SPL meter for absolute measurements. Enter your reference level, tap Calibrate, and Spektral averages 3 seconds of stable signal to compute the offset automatically
• Rolling Leq — True rolling equivalent continuous levels: a five-tile grid shows Leq over the last 1 minute, 15 minutes, 30 minutes, 1 hour, and 8 hours. Rolling means exactly that — always the most recent window, the number a licensing officer would measure, never an average anchored to when you pressed start
• Limit Guard — Choose an enforcement window (LAeq 1–60 min) and your venue limit in dB SPL. Spektral shows your live headroom, colour-codes the relevant Leq readouts green/amber/red, and when the current level would breach the limit, forecasts the exact time to breach by simulating the real measurement window — a countdown you can mix to. A dashed limit line appears on the history graph. Requires calibration, since the limit is compared in dB SPL
• Historical Tracking — Scrolling time-history graph showing SPL over the duration of your session (up to 8 hours), color-coded by threshold zones
• Session Statistics — Automatic tracking of minimum, maximum, and average SPL values with START/STOP/RESET controls
• PDF Export (macOS) — Generate professional compliance reports with metadata (event name, location, notes), microphone and calibration details, time-history graphs, and session statistics
• Configurable Thresholds — Set caution and warning levels (default 85/95dB) with quick presets for Environmental, Occupational Safety, Live Event, and Studio use cases
• SPL Aura — Optional visual glow in both RTA and SPL Monitor modes showing at-a-glance when levels exceed thresholds (green/orange/red)

Calibration Workflow: Couple an acoustic calibrator to your mic (or position a reference SPL meter) → enter the reference level in Settings (default 94 dB) → tap Calibrate → Spektral captures 3 seconds of signal, checks stability, and applies the offset to all SPL readouts. If the signal is too noisy the calibration is rejected with a clear message.

Uncalibrated? Until a calibration is set, SPL readouts are labelled dBFS(A) with a clear warning — the values are relative to digital full scale (and typically negative), not sound pressure. Calibrate for absolute levels.

On iPhone and iPad: Spektral Mobile includes the full SPL Monitor mode — hero readout, the rolling Leq tiles, min/max/average, and the session history graph — using your calibrated phone. Select SPL in the mode tabs. PDF export remains on the Mac for now.

RT60 Reverberation Time (macOS only)

RT60 measures how long sound takes to decay by 60 dB — the single most useful number for characterising a room. Select RT60 in the mode picker. Spektral measures the decay across six octave bands (125 Hz, 250 Hz, 500 Hz, 1 kHz, 2 kHz, 4 kHz) using Schroeder backward integration.

To measure:

1. Press MEASURE. Spektral plays an interrupted pink-noise burst through your output, then captures the room's decay. (Use MANUAL if you'd rather excite the room yourself — a balloon pop, starter pistol, or clap — and have Spektral capture from the impulse.)
2. The large readout shows the reverberation time for the selected band; tap a band (125–4k) to inspect it, and the decay curve plots that band's energy decay over time.
3. The RT60 BY BAND chart shows the result across all bands at a glance, so you can see whether the room is bright, balanced, or boomy.

Reading the results:

• T20 / T30 — reverberation time extrapolated from the first 20 dB or 30 dB of decay. T30 is the usual reference; T20 is useful when the noise floor is high. Switch between them (and EDT) with the toggle.
• EDT (Early Decay Time) — derived from the first 10 dB of decay; correlates with the perceived liveness of a room.
• R² — goodness-of-fit of the decay line (0–1). Values near 1.0 mean a clean, trustworthy measurement; low values flag a noisy or too-short capture — raise the level or measure again.

Use the export button to copy the full per-band results. RT60 is ideal for room tuning, verifying acoustic treatment, and commissioning documentation.

STIPA Speech Intelligibility (macOS only)

STIPA measures the Speech Transmission Index (STI) — how intelligible speech will be through a sound system or in a space — following the IEC 60268-16 modulation-transfer method. Select STIPA in the mode picker. It's the measurement required for PA and voice-alarm intelligibility sign-off, including EN 54 / BS 5839 life-safety commissioning.

To measure:

1. Play the STIPA test signal through the system under test. Press GENERATE SIGNAL to play Spektral's built-in IEC-compliant test signal through your output, or feed in an external STIPA signal.
2. Position the measurement microphone at the listening position and press MEASURE. Choose a measurement duration of 15, 20, or 30 seconds — longer captures are more robust in noisy or reverberant spaces.
3. Spektral reports the overall STI score with a clear rating, plus a full per-band breakdown.

Reading the results:

• STI score — a single 0–1 value rated Bad (<0.30), Poor (0.30–0.45), Fair (0.45–0.60), Good (0.60–0.75), or Excellent (>0.75). Most voice-alarm standards require at least 0.50 (Common Intelligibility Scale ≈ 0.70).
• Per-band MTF — modulation transfer for each of the seven octave bands (125 Hz–8 kHz) at the two STIPA modulation frequencies, so you can see which bands are dragging intelligibility down.
• SNR — the signal-to-noise ratio Spektral derived per band; low or negative values indicate the band is being masked by background noise.

Use the export button to copy the score and per-band data into your commissioning report.

Remote Monitoring

Remote Monitoring turns any iPhone running Spektral into a wireless analysis station — capturing both the full RTA spectrum and SPL at every site. Deploy devices at multiple measurement points around your event and watch each site's live spectrum, noise levels, and history from the dashboard on your Mac.

What you need

• Spektral for macOS (£29.99) — this is where you purchase the Remote Monitoring add-on and view your dashboard
• Remote Monitoring add-on (£19.99) — one-time in-app purchase within the macOS app
• Spektral for iOS (£9.99) — one universal app for iPhone and iPad, buy once. Each device you deploy as a monitoring station needs this app installed
• All devices signed into the same Apple ID

Complete multi-point monitoring system for under £60. No subscriptions, no servers, no accounts to create.

Step 1: Purchase on your Mac

1. Open Spektral on your Mac and switch to SPL Monitor mode
2. At the bottom of the SPL Monitor you will see a Remote Monitoring bar — click Learn More
3. Purchase the Remote Monitoring add-on (£19.99, one-time)
4. The dashboard unlocks immediately on your Mac

Step 2: Unlock appears on your iPhones

Once purchased on Mac, the unlock syncs automatically to every iPhone and iPad signed into your Apple ID via iCloud. Within seconds, a new Start Remote Listener button appears in Spektral's settings on each iOS device. No code to enter, no account to link — it just works.

If you don't see it immediately, ensure the iOS device is connected to the internet and signed into the same Apple ID as your Mac.

Step 3: Deploy your iPhones

1. Open Spektral on an iPhone and go to Settings > Start Remote Listener
2. Name your measurement location (e.g. "Stage Door", "Nearest Residence")
3. Tap Start Monitoring — the device begins capturing RTA spectrum and SPL, syncing readings via iCloud
4. Position the iPhone at the monitoring point — it will continue measuring in the background
5. Repeat for each location (up to 5 sites)

Step 4: Monitor from your Mac

On your Mac, the dashboard updates as new data arrives from each remote iPhone. Each iPhone captures RTA spectrum and SPL continuously, and syncs back to your Mac every 10 seconds via iCloud — so the dashboard shows near-live levels with a 10-second cadence rather than instantaneous streaming.

Dashboard features:

• Live readouts — Current SPL, Leq, peak, RTA spectrum, and connection status for each site
• Site detail view — Click any site to see its full SPL and spectrum time-history with min/max/average statistics
• Satellite map — View all sites on a map with SPL-coloured annotations and heat map propagation overlay
• PDF export — Generate compliance reports with a dedicated page per remote site, including time-history graph and statistics
• Offline handling — Sites that stop monitoring are shown as dimmed rather than disappearing, so you can review their last readings

Calibration

The calibration offset corrects for each iPhone's microphone sensitivity — a property of the hardware, not the location — so you can calibrate all your iPhones together at FOH before deploying them to their monitoring positions:

1. Set up a calibrated reference SPL meter (or your calibrated Mac) at FOH
2. Play pink noise through the PA at a steady moderate level
3. Place each iPhone near the reference mic in turn
4. On each iPhone, go to Settings, enter the reference level, and tap Calibrate
5. Deploy the iPhones to their monitoring positions — the offset travels with the device

If using external measurement microphones (e.g. MicW i436), you can calibrate with an acoustic calibrator coupled directly to the mic at each site. Bluetooth headsets are not suitable for calibration. All calibrated readings sync automatically to your Mac via iCloud.

Privacy & data

All data stays in your personal iCloud account using Apple's CloudKit private database. No third-party servers, no accounts to create, no data shared with other users. Readings sync privately between your own Apple devices only. If an iPhone temporarily loses connectivity, readings are stored locally and synced when the connection resumes — no data is lost.

Note on accuracy

Spektral calibrated against a reference meter is a practical and affordable monitoring solution suitable for many noise management scenarios. For higher accuracy, third-party calibrated iOS measurement microphones (such as the MicW i436 or Dayton Audio iMM-6) can significantly improve performance and approach Class 2 levels of accuracy. However, formal IEC 61672 Class 1 or Class 2 certification requires the entire measurement chain — microphone, preamp, ADC, and software — to be type-approved as a complete instrument. No iOS app can carry this certification independently, regardless of microphone used. Where statutory noise measurement is legally required, a certified sound level meter remains necessary.

Run / Pause

The transport is a play/pause pair — both buttons stay visible, with the active state highlighted. Click either, or press Space, to start or pause audio analysis. When paused, the display freezes at the last captured frame.

Peak Hold

Click PEAK in the toolbar or press P to toggle peak hold. A red trace appears showing the maximum level at each frequency since peak hold was enabled.

Double-click PEAK or press R to reset the peak hold trace.

Identify (Soundcheck)

Click IDENTIFY or press F to enable automatic peak frequency identification. Spektral will detect and label the strongest spectral peaks directly on the RTA view, showing their exact frequency (and optional musical note).

Sentinel

Click SENTINEL to arm the Feedback Sentinel — live feedback detection with suggested notches and a session ring-out list. The button lights red while armed.

Snapshot Compare

Click SNAP or press C to capture the current spectrum as a reference trace. A dashed white line will appear, allowing you to compare your live audio against the frozen reference. Click again to clear the snapshot.

This is useful for seeing the effect of EQ changes — apply a notch filter, then compare the new response against the snapshot.

Pink Noise Generator

Click PINK or press N to toggle the built-in pink noise generator. Pink noise has equal energy per octave, making it ideal for system testing.

Adjust the output level from the slider icon next to PINK in the toolbar (it also appears in the RT60 and STIPA panels) — this level is shared with the STIPA test signal. Choose where it plays out with the Output Device setting in Settings (Cmd+,); the generator runs on a separate audio output so it doesn't interfere with the analysis input.

Smoothing

Use the SMOOTH dropdown or press S to cycle through smoothing options, from full resolution to 1/1 octave smoothing. Fractional octave smoothing groups FFT bins into bands, reducing visual noise at the cost of frequency resolution.

Averaging

Use the AVG dropdown or press A to cycle through temporal averaging options. Averaging smooths the display over time using an exponential moving average.

Higher averaging values produce a more stable trace at the cost of slower response to transient sounds.

Gain and Floor (macOS)

Use the GAIN slider to boost or attenuate the displayed signal level (-10dB to +40dB). This is a display offset only — it does not affect your audio signal.

Use the FLOOR slider to adjust the noise floor of the display (-90dB to -20dB). Raising the floor hides low-level noise.

Double-click either label to reset to the default value.

iOS: Use touch gestures instead — see the Touch Gestures section below.

Frequency Readout

The frequency and dB readout appears on the right side of the toolbar. When you hover over the RTA, it shows the frequency and level at the cursor position.

When the cursor snaps to a peak, a target icon appears and the reading uses parabolic interpolation for sub-bin frequency accuracy.

Gesture Reference

On iPhone and iPad, Spektral supports intuitive touch gestures for navigating the frequency and dB axes:

Settings (smoothing, averaging, peak hold, and other features) are accessed via the gear icon in the toolbar.

Shortcut Reference

FFT Size

Choose the analysis FFT size in Settings, from 2048 to 16,384 points. Larger sizes give finer frequency resolution — mainly useful in the low end for room modes and sub feedback — at the cost of a slower-responding display. The status bar shows the resulting bin resolution. Spektral follows your interface's sample rate, including 96 kHz on interfaces that support it.

SPL Calibration

Open Settings (Cmd+, on Mac, or the gear icon on iOS) to calibrate your microphone for absolute SPL measurements:

1. Couple an acoustic calibrator (typically 94 dB or 114 dB) to your microphone, or position a reference SPL meter at the same listening point
2. Enter the known reference level in the Reference level (dB) field — default is 94
3. Tap Calibrate — Spektral captures 3 seconds of signal at 10 samples/second
4. If the signal is stable (±1.5 dB or less), the calibration offset is applied automatically. If too noisy, calibration is rejected with an error — check the calibrator coupling and try again

After calibration, the offset and timestamp are shown and can be reset at any time. Calibration is captured using flat (Z) weighting internally, so it works correctly regardless of your display weighting (A/C/Z) and with calibrators at any frequency.

Calibration affects all SPL readouts (toolbar, SPL Monitor, Leq values, Remote Listener) but does not affect the RTA spectrum display.

Tips

• Use a wired or USB measurement microphone for calibration — Bluetooth headsets use compressed audio codecs that make SPL readings unreliable. Spektral will warn you if a Bluetooth mic is detected on iOS.
• If you change your audio input device (plug in a different mic, connect headphones, etc.), Spektral shows a warning that calibration may no longer be accurate. Re-calibrate after switching devices.
• The calibration offset is stored globally — if you use multiple microphones, re-calibrate when switching between them.

Accuracy & External Microphones

The built-in microphones on Mac and iPhone are designed for voice, not measurement. For serious SPL work, calibrating against a reference meter corrects for the offset and gives you reliable relative readings — more than adequate for monitoring trends and staying within limits.

For improved absolute accuracy, consider a third-party calibrated measurement microphone designed for iOS or Mac:

• MicW i436 — Class 2-equivalent MEMS measurement mic with calibration certificate
• Dayton Audio iMM-6 — Affordable calibrated measurement mic with individual frequency response data
• miniDSP UMIK-2 — USB measurement microphone with calibration file (macOS)

These microphones offer flatter frequency response and tighter tolerances than built-in mics, and when calibrated can approach the performance requirements of IEC 61672 Class 2.

Important: Formal Class 1 or Class 2 certification under IEC 61672 requires the entire measurement chain — microphone, preamp, analogue-to-digital converter, and software — to be type-approved as a complete instrument. No combination of iOS app and external microphone can carry this certification independently. Where a statutory noise assessment legally requires a certified instrument, a dedicated Class 1 or Class 2 sound level meter remains necessary. That said, many environmental health officers will accept calibrated iOS-based measurements for practical noise management — particularly when calibrated against a known reference meter on site.

Frequency Weighting

Choose between A-weighting, C-weighting, or Z-weighting (flat) in Settings, on both Mac and iOS. These are frequency weighting curves based on IEC 61672 methodology (see also Why are there two weighting settings?):

• A-weighting (default) — Mimics human hearing sensitivity. Required for most environmental noise regulations and occupational safety standards. Attenuates low and high frequencies.
• C-weighting — Flatter response than A-weighting. Used for peak measurements and entertainment noise assessments.
• Z-weighting (flat) — No frequency weighting applied. Measures the true acoustic energy across all frequencies.

Time Weighting

Choose between Slow, Fast, or Impulse time weighting in Settings, on both Mac and iOS. These control how quickly the SPL reading responds to changes in sound level, based on IEC 61672 time constants:

• Slow (1s) — Smooths out fluctuations for steady-state noise monitoring. Best for environmental compliance and long-term measurements.
• Fast (125ms) — General-purpose setting that tracks most level changes while filtering brief transients. Good default for live sound monitoring.
• Impulse (35ms rise, 1.5s decay) — Fast attack captures sharp transients (door slams, drum hits) with slow decay for readability. Use when peak events matter.

SPL Monitoring Thresholds (macOS only)

Configure caution and warning thresholds in Settings to match your compliance requirements. The VU meter, history graph, and SPL aura all respond to these user-defined levels:

• Caution Threshold (default 85dB) — Orange visual indicators appear when SPL exceeds this level
• Warning Threshold (default 95dB) — Red visual indicators appear when SPL exceeds this level

Quick Presets:

• Environmental (85/95 dB) — Typical residential noise ordinance limits
• Occupational Safety (85/90 dB) — OSHA/HSE workplace exposure limits
• Live Event (90/100 dB) — Higher thresholds for concert venues
• Studio (80/90 dB) — Lower thresholds for controlled environments

SPL Aura / Glow

Enable "Show SPL Aura" in Settings to display a visual glow around the main window when sound levels exceed your configured thresholds. This works in both RTA mode (when toolbar shows SPL readout) and SPL Monitor mode.

The glow smoothly fades between green (safe), orange (caution), and red (warning) based on your threshold settings. Useful for at-a-glance monitoring without needing to read exact numbers.

What is the Reference channel?

The Reference channel is only used by Transfer Function mode. It carries the signal your measurement mic is compared against — the console output or a pink noise loopback — and the difference between the two is the system's magnitude, phase, and coherence.

For RTA-only use you can ignore it: the RTA reads the Measurement channel alone. But don't patch your measurement mic into both inputs — in Transfer mode you would be comparing the mic against itself and get a flat, meaningless trace. Patch a desk output into the Reference input and Transfer will show the system's actual response.

The Ref button on the RTA toolbar overlays the reference channel's spectrum, a quick way to check what that input is seeing. See Selecting an Audio Input for how the pink noise loopback is routed.

Why does the RTA slope down at high frequencies?

Because Spektral's RTA is a narrowband display, the convention used by professional dual-channel analysers — and on a narrowband display, a healthy system playing pink noise or music is supposed to slope downward. There is no low-pass filter anywhere.

Pink noise carries equal energy per octave. Up high, each octave's energy is shared across far more FFT bins than at the bottom, so each individual bin reads lower: the trace falls by about 3 dB per octave even through a perfectly flat system. Traditional 1/3-octave RTAs sum the bins in each band before displaying, which is why pink noise reads flat on those. Same signal, two display conventions.

• A gentle downward tilt on music or venue ambience is what "flat" looks like here. Judge the shape — bumps, notches, and peaks against the local trend — not the overall tilt.
• The Smoothing control averages neighbouring bins for readability; it does not change the slope.
• To judge system response independent of the display convention, use Transfer Function mode — it divides out the source signal, so flat means flat.
• If the top end looks more rolled off than a gentle tilt, check your Spectrum weighting is Z (flat).

Why are there two weighting settings?

Because the SPL number and the spectrum display answer different questions, Spektral weights them independently:

• SPL meter weighting shapes the SPL and Leq readouts — the toolbar figure, the SPL overlay, SPL Monitor mode, and Limit Guard. A-weighting is the regulatory standard; C-weighting is common for music levels. The current choice is shown next to the reading, e.g. dB(A) or dB(C), and in the status bar.
• Spectrum weighting shapes the RTA trace and spectrograph themselves. This should normally stay on Z (flat) so you see the system's full bandwidth — A or C weighting visibly pulls down the lows and the top octave on the display. The plot label tells you at a glance: it reads "RTA" when flat, or "RTA (A)" / "RTA (C)" when weighted.

A common combination is dB(C) on the meter for show levels with a flat Z spectrum for analysis. SPL calibration is captured flat internally, so it stays correct whichever weightings you choose.

No audio input detected

• Check that you have granted microphone permission in System Settings > Privacy & Security > Microphone.
• Verify your audio input device is connected and selected in Spektral Settings (Cmd+,).
• Try selecting a different input device.
• Check that no other application has exclusive access to the audio device.

Spectrum appears flat or very low

• macOS: Increase the GAIN slider to boost the displayed level, or lower the FLOOR slider to reveal low-level signals.
• iOS: Drag vertically to shift the dB range, or use a two-finger vertical pinch to widen the visible range.
• Check your audio interface input gain.
• Ensure the correct input channel is selected on your interface.

Transfer Function says "Dual-Channel Input Required"

• Transfer Function needs at least two inputs on the selected device: one for the measurement mic and one for the reference loopback. Built-in Mac microphones are mono, so an audio interface is required.
• Check the correct interface is selected in Settings (Cmd+,), and that the Measurement and Reference channels point at the inputs you have patched.
• On macOS you can use an iPhone running Spektral as the measurement mic instead — see iPhone Wireless Probe. The reference still needs its loopback input on the interface.

High CPU usage

Spektral uses efficient rendering via Apple's Canvas and Accelerate frameworks. If you notice high CPU usage:

• Check that you are running macOS 14.0 or later.
• Close other apps that may be competing for GPU resources.
• Use a higher smoothing setting to reduce per-frame computation.