PA Measurement and Calibration with Smaart: Transfer Function, Coherence Spectrum, Phase and Delay.

Rational Acoustics Smaart acoustic measurement software (v8/v9 Suite) represents the established global standard for electroacoustic metrology, calibration, optimisation and time alignment of professional sound-reinforcement (PA) systems. For system engineers, FOH technicians and live-audio optimisers, Smaart turns subjective listening impressions into rigorous, quantitative laboratory data for amplitude (magnitude), time (delay and impulse) and phase (angular relationship).

Correct PA system setup requires understanding everything from capture hardware and SPL calibration to interpreting dual-channel transfer functions, precisely aligning frequency bands (subwoofers, tops, front fills, side fills and delay towers), and applying corrective EQ based on spatial averaging.

---

1. Measurement Hardware Architecture: Microphones, Interface and Loopback

Reliable acoustic measurements require a calibrated hardware chain without bottlenecks or phase distortion introduced by the capture electronics.

                  ┌────────────────────────────────────────────────────────┐
                  │                 PINK NOISE / SWEEP GENERATOR           │
                  └───────────────────────────┬────────────────────────────┘
                                              │
                                              ▼
┌─────────────────────────┐        ┌──────────────────────┐        ┌─────────────────────────┐
│ REFERENCE INPUT (CH1)   │◄───────┤ MIXER AUX/DSP OUTPUT ├───────►│ DSP PROCESSOR / AMPS    │
└────────────┬────────────┘        └──────────────────────┘        └────────────┬────────────┘
             │                                                                  │
             │ [CLEAN ELECTRICAL REFERENCE LOOPBACK]                            ▼
             │                                                          ┌──────────────┐
             │                                                          │ PA SPEAKERS  │
             │                                                          └──────┬───────┘
             │                                                                 │ [ACOUSTIC WAVE]
             │                                                                 ▼
┌────────────┴────────────┐                                        ┌───────────┴─────────────┐
│ SMAART SOFTWARE (v9)    │                                        │ MEASUREMENT MICROPHONE  │
│ Dual-Channel Transfer F.│◄───────────────────────────────────────┤ OMNIDIRECTIONAL (CH2)   │
└─────────────────────────┘      [ACOUSTIC MEASUREMENT INPUT]      └─────────────────────────┘

1.1 Measurement Microphone

The measurement microphone should be an electret condenser with an entirely omnidirectional polar pattern and small capsule ($1/4”$ or $1/2”$).

• Flat frequency response: A neutral curve from $20\text{ Hz}$ to $20\text{ kHz}$ ($\pm 1\text{ dB}$). $1/4”$ capsules cause less high-frequency disturbance of the diffuse field.
• Dynamic range and maximum SPL: Handles dynamic levels above $130 - 140\text{ dB SPL}$ without mechanical diaphragm overload or internal FET preamplifier distortion.
• Calibration files (.miccal): Load the serial-number-specific correction file to cancel small deviations at the spectral extremes ($>10\text{ kHz}$ or $<30\text{ Hz}$).

1.2 Professional Audio Interface

Requires highly linear AD/DA converters with ultra-low background noise ($THD+N < 0.001%$), preferably stepped gain knobs or digital control (to avoid accidental changes), and stable $+48\text{ V}$ Phantom power.

1.3 Reference Loopback Channel

Loopback sends an identical, unaltered copy of the signal generated by Smaart (or DJ/FOH console output) to audio-interface Input 1 (Reference).

• Function: Mathematically cancels sound-card nonlinearities, oscillator level variability and digital-clock fluctuations. Smaart subtracts the reference channel from the measurement channel captured by the microphone in air.

---

2. SPL Calibration (Pressure Calibration)

Sound-pressure-level (SPL) calibration converts relative dBFS readings (digital Full Scale) into absolute physical dBSPL values with metrological precision.

1. Using an Acoustic Calibrator (Pistonphone): Insert the measurement microphone capsule into an approved acoustic calibrator (IEC 60942 Class 1 or Class 2) emitting a pure $1\text{ kHz}$ sine tone at a calibrated $94.0\text{ dB SPL}$ or $114.0\text{ dB SPL}$.
2. Smaart Procedure:
   • In Input Config, select Calibrate SPL.
   • Start reading with the pistonphone on and manually enter the exact calibration figure ($94.0\text{ dB}$).
   • Smaart sets the dBFS-to-dBSPL Offset factor.
3. Enabled Metrics: During the event, monitor LAeq, LEQ, Peak SPL, LAF, LAS and exposure in real time according to hearing-protection and noise-pollution-control regulations.

---

3. Smaart Measurement Modes: RTA, Spectrogram and Transfer Function (TF)

Smaart divides its operational architecture into two principal groups: Single-Channel analyzers and Dual-Channel Transfer Function analyzers.

Mode / Tool	Channels	Measured Parameters	Main Application
RTA (Real-Time Analyzer)	1 (Measurement)	Amplitude vs Frequency (dB SPL)	Visual spectral analysis of live music, feedback detection and relative tonal balance.
Spectrograph (Spectrogram)	1 (Measurement)	Frequency vs Time vs Amplitude (Colour)	Identifying room resonances, persistent modes and decay over time.
Transfer Function: Magnitude	2 (Ref + Meas)	Spectral gain/attenuation ($\text{dB}$)	Actual system frequency response excluding original music/source.
Transfer Function: Phase	2 (Ref + Meas)	Angular offset ($\text{degrees } ^\circ$)	Time alignment, crossovers, phase coherence and polarity inversion.
Transfer Function: Coherence	2 (Ref + Meas)	Data quality ($0.0 \text{ to } 1.0$)	Distinguishing background noise, room reflections and destructive interference.
Impulse Response (IR)	2 (Ref + Meas)	Temporal Response / Time ($\text{ms}$)	Determining physical delay, early/late reflections and RT60/EDT parameters.

3.1 RTA (Real-Time Analyzer)

RTA displays spectral energy distribution in fractional-octave bands ($1/1, 1/3, 1/6, 1/12, 1/24$ or FPPO - Fixed Frequency Resolution Per Octave). It requires no reference channel, but cannot distinguish whether a frequency peak comes from the playing track or a loudspeaker defect.

3.2 Spectrograph (Spectrogram)

Displays a three-dimensional waterfall in 2D: Y is continuous time, X frequency, and colour intensity (dB). It is exceptionally useful for locating acoustic resonances persisting in the room after a sub-bass hit.

---

4. Dual-Channel Transfer Function

The dual-channel Transfer Function is Smaart’s advanced core. It calculates the complex quotient between the microphone signal ($H_{\text{med}}(f)$) and original reference signal ($H_{\text{ref}}(f)$):

[ H(f) = \frac{H_{\text{med}}(f)}{H_{\text{ref}}(f)} = |H(f)| \cdot e^{j \phi(f)} ]

4.1 Magnitude Response

Shows the gain or attenuation in decibels imposed by speaker and room on each frequency. If emitted sound has a valley at $100\text{ Hz}$ but the recorded music had a peak there, TF magnitude shows only loudspeaker/room behaviour at $100\text{ Hz}$, ignoring musical content.

4.2 Phase Response (Phase Trace)

Represents angular phase difference ($\phi(f)$ between $-180^\circ$ and $+180^\circ$) against frequency.

• A flat phase trace indicates all frequencies arrive perfectly aligned in time (linear phase / constant delay).
• A downward slope indicates uncompensated delay (acoustic sound arrives later than the internal reference).

4.3 Coherence Spectrum

Coherence ($\gamma^2(f)$) quantifies the degree of linear causality between reference input and measured output:

[ 0.0 \le \gamma^2(f) \le 1.0 ]

• High Coherence ($\gamma^2 > 0.8$): Excellent signal-to-noise ratio ($S/N$). Sound at the microphone comes directly and cleanly from the loudspeaker under test. Magnitude and phase data are valid for EQ and alignment decisions.
• Low Coherence ($\gamma^2 < 0.5$): Data is entirely unstable and unusable because of:
  1. Severe room reflections (reverberant field dominating the direct field).
  2. Comb-filter cancellation between two speakers covering the same zone with different arrival times.
  3. High background noise (audience, traffic, generators).
  4. Time mismatch (Delay Finder not correctly synchronized).

[!CAUTION] SYSTEM METROLOGY GOLDEN RULE: Never apply EQ at frequencies where coherence falls to minimum levels. Boosting $+6\text{ dB}$ in a phase or room-reflection cancellation overloads the power amplifier and destroys the compression driver without increasing the acoustic level at that point by a single dB.

---

5. Impulse Response (IR) and Delay Finder

For mathematically coherent phase and magnitude responses, Smaart needs to synchronize reference-signal timing with the acoustic wave’s flight time in air.

          [Sent Tone / Noise   ] ───> Audio Output
                                         │
                                         ▼ [Acoustic Flight: d = 10.2 metres]
                                         ▼ [Propagation time: t = d / c = 29.73 ms]
                                         │
[Impulse Response IR    ] ───────────────┼───────────────► Main peak at +29.73 ms

5.1 Acoustic Flight Time

The speed of sound ($c$) at $20^\circ\text{C}$ is approximately $343\text{ m/s}$. Time taken to travel distance $d$ is:

[ t_{\text{vuelo}} = \frac{d}{c} = \frac{d}{343} \quad (\text{seconds}) ]

For $10\text{ metres}$, $t_{\text{vuelo}} \approx 29.15\text{ ms}$. Besides physical flight, converters and DSP filters add internal latency ($1.5 - 4\text{ ms}$).

5.2 Using Delay Finder

The first job when measuring a band is to find the actual arrival delay, rather than its theoretical value.

1. Enable TF measurement and press Delay Finder (or Track).
2. Smaart calculates the impulse response’s time-domain cross-correlation ($IR$).
3. It identifies the initial arrival peak and calculates delay in ms (e.g. $31.45\text{ ms}$).
4. Entering this value in Smaart’s reference visually flattens the phase-trace slope, allowing frequency-cutoff and component-alignment analysis.

---

6. Complete Measurement Procedure: Pink Noise vs Sweeps

Two standardised methods provide the PA system’s test signal:

6.1 Pink Noise

• Properties: Constant energy per octave fraction (falling $-3\text{ dB/octave}$).
• Smaart Use: Continuous output with time averaging (Pink Sweep / FIFO Averaging). Preferred at festivals and for quick checks because magnitude and phase changes can be observed in real time while adjusting DSP.

6.2 Logarithmic Sine Sweep

• Properties: A continuous, variable-frequency sine tone sweeping from $20\text{ Hz}$ to $20\text{ kHz}$.
• Smaart Use: Requires Impulse Response Mode. Provides extreme ambient-noise rejection and isolates the first acoustic arrival from room reverberation decay (RT60) through FFT processing with Time Windowing.

---

7. Advanced Measurement and Alignment: Individual Cabinet, System, Sub-Top, Front Fills and Delays

Complete PA alignment proceeds sequentially from lower to higher complexity.

              ┌─────────────────────────────────────────┐
              │  STEP 1: Individual Cabinet Measurement │
              └────────────────────┬────────────────────┘
                                   │
                                   ▼
              ┌─────────────────────────────────────────┐
              │  STEP 2: Diagnose Polarity vs Phase     │
              └────────────────────┬────────────────────┘
                                   │
                                   ▼
              ┌─────────────────────────────────────────┐
              │  STEP 3: Sub-Top Phase Alignment        │
              └────────────────────┬────────────────────┘
                                   │
                                   ▼
              ┌─────────────────────────────────────────┐
              │  STEP 4: Front Fill & Delay Integration │
              └────────────────────┬────────────────────┘
                                   │
                                   ▼
              ┌─────────────────────────────────────────┐
              │  STEP 5: Spatial Averaging & PA EQ      │
              └─────────────────────────────────────────┘

7.1 Measuring an Individual Cabinet (Component Level)

Before enabling the whole rig, measure each band separately (bass, mids, high-frequency driver):

• Check each transducer’s unfiltered frequency response to identify the enclosure’s natural resonances.
• Set manufacturer-recommended crossover points (e.g. 24 dB/oct Linkwitz-Riley at $1.2\text{ kHz}$ between mids and compression driver).

7.2 Rigorous Diagnosis: Polarity vs Phase

A common live-sound mistake confuses polarity inversion with a time-delay mismatch:

• Polarity Inversion ($180^\circ$ Electrical):
  • The whole phase trace instantly jumps vertically by $180^\circ$ equally across all frequencies. Polarity is frequency-independent.
  • Solution: Flip polarity ($\varnothing$ / $180^\circ$) on the DSP or console channel.
• Time Offset (Acoustic Delay):
  • The phase trace shows a wrapping slope making multiple complete $360^\circ$ turns (Phase Wrap) as frequency increases, because the angular relationship depends on wavelength ($\lambda = c / f$).
  • Solution: Add milliseconds ($ms$) of delay in DSP.

       INVERTED POLARITY                            TIME OFFSET (DELAY)
  Phase (°)                                     Phase (°)
  +180 ────────────────────────────             +180 ───┐   ┌───┐   ┌───┐   ┌───
     0                                             0 ────\─/─────\─/─────\─/────
  -180 ────────────────────────────             -180 ─────┘   └───┘   └───┘   └─
       (Flat line shifted by 180°)                   (Rotating downward slope)

7.3 Sub-Top Phase Alignment

For maximum acoustic summing of $+6\text{ dB}$ in the crossover coupling region ($80\text{ Hz} - 100\text{ Hz}$):

1. Microphone placement: Position the measurement microphone in the critical transition area (where subwoofer and main PA cabinet sound pressures are identical).
2. Measuring the Top (Main PA):
   • Turn off the subwoofer. Turn on the main PA.
   • Run Delay Finder to capture the Top’s flight time. Save magnitude and phase traces (Capture Trace).
3. Measuring the Subwoofer:
   • Turn off the main PA and turn on the subwoofer.
   • IMPORTANT! Do not run Delay Finder again. Keep the reference calculated for the Top fixed.
   • Observe the subwoofer’s phase trace in the crossover band ($60\text{ Hz} - 100\text{ Hz}$).
4. Matching Phase Slopes:
   • Adjust subwoofer (or top) Delay in DSP until the subwoofer phase-trace slope perfectly overlays the top’s phase trace throughout the crossover region.
5. Summation Verification:
   • Enable both systems simultaneously (Top + Sub). Combined magnitude should show a $+6\text{ dB}$ sum peak above each individual cabinet’s level at the cutoff frequency. A cancellation or dip indicates nonparallel phase slopes or inverted polarity.

7.4 Front Fill Alignment

Front Fills cover the audience’s first rows outside the vertical beam of Line Arrays or main cabinets.

1. Measure the main PA at audience row 3 and save its time/phase trace.
2. Enable the Front Fill and measure at the same row 3.
3. Apply digital Front Fill delay in DSP ($t_{\text{fill}} = t_{\text{PA}} - t_{\text{directo_fill}}$) until main-PA and front-fill initial transients arrive together in the overlap zone.

7.5 Delay Tower Alignment

At festivals or large venues, delay towers reinforce sound pressure $50 - 100\text{ metres}$ from the main stage.

1. Measure main-PA acoustic arrival at the delay-tower position (e.g. $180\text{ ms}$ flight).
2. Set tower DSP delay so its processed signal arrives exactly $0 - 5\text{ ms}$ after the main-PA wave (using the Haas / Precedence Effect to keep sound localisation on the main stage).

---

8. Measurement-Based EQ and Spatial Averaging

[!WARNING] NEVER EQ AT ONLY ONE ROOM POSITION: A microphone in one seat captures local acoustic nodes and antinodes from nearby wall reflections. Correcting that isolated point ruins spectral response elsewhere in the venue.

8.1 Spatial Averaging Method

1. Multiple positions: Take TF measurements at 4 to 8 representative audience positions (centre axis, left, right, middle and rear).
2. Generating the Spatial Average Trace: Smaart calculates vector or power averaging of captured traces.
3. Identifying actual system modes:
   • Loudspeaker / system anomalies: Excess energy at $2.5\text{ kHz}$ with high coherence at all positions is a loudspeaker or horn defect. Correct it with DSP EQ.
   • Local room anomalies: A deep $160\text{ Hz}$ valley appearing only at one specific position is local reflection cancellation. Do NOT EQ it.

8.2 Why to Avoid a “Completely Flat” Adjustment

A perfectly flat $20\text{ Hz}$ to $20\text{ kHz}$ frequency response measured by a microphone sounds extremely shrill, thin and tiring to human ears. The temptation to EQ until the screen looks flat is a beginner’s most common measurement mistake: a microphone does not hear like a person, and the room is not a free field.

• PA Target Curve / House Curve: Professional live PA systems (d&b audiotechnik, L-Acoustics, Meyer Sound) use a target curve with progressive bass boost ($+6\text{ to } +10\text{ dB}$ below $80\text{ Hz}$) and gentle treble roll-off ($-2\text{ to } -4\text{ dB}$ above $10\text{ kHz}$) to respect Fletcher-Munson equal-loudness curves at high SPL.

---

9. Diagnosing Common Live-Sound Measurement Errors

Observed Smaart Symptom	Likely Root Cause	Corrective Rig Action
Phase trace wraps $360^\circ$ extremely quickly (Phase Noise).	Delay Finder is unset or the speaker moved.	Press Delay Finder again to update the reference flight time.
Drastic high-frequency Coherence drop ($>8\text{ kHz}$).	Wind, thermal air turbulence outdoors or ground friction.	Use a microphone windscreen and longer time averaging.
Periodic deep magnitude valleys with sawtooth coherence.	Comb Filtering from overlapping noncoplanar cabinets.	Readjust physical cabinet coverage angles or isolate channels for individual measurement.
Magnitude trace clipped to a flat $+6\text{ dB}$ plateau.	Digital or analogue overload in the interface input preamplifier.	Reduce audio-interface input-preamp Gain.
Sub and Top sum to $-6\text{ dB}$ at cutoff instead of $+6\text{ dB}$.	Polarity inversion or exactly $180^\circ$ offset at crossover.	Invert subwoofer-band polarity in DSP and verify phase alignment.

---

10. Good Measurement Practice for System Technicians

1. Check physical loopback before going into the venue, ensuring the interface introduces no internal offsets.
2. Calibrate the microphone with a pistonphone to verify sound-pressure levels meet the event’s legal regulations.
3. Require high coherence ($>0.8$) before touching any DSP parametric-EQ setting.
4. Align phase in the crossover coupling region, adjusting mechanical or digital delays until both bands’ phase slopes are parallel.
5. Use spatial averaging across multiple audience positions to correct actual cabinet tonal deviations rather than local room acoustic nulls.