STI Report Template and Analysis

By Sarah Okonkwo · May 5, 2026

STI Report Template and Analysis

1) Introduction: Context and Why STI Analysis Matters

The Speech Transmission Index (STI) is one of the most actionable single metrics for predicting speech intelligibility in rooms and sound systems. For audio professionals, STI sits at the intersection of electroacoustics and outcomes: it translates measurable properties—noise, reverberation, and system nonlinearity—into an index that correlates with how well listeners understand speech. That makes STI valuable in scenarios where “sounds clear” is an expensive and unreliable acceptance criterion.

STI-based reporting is common in commercial AV handover packages, transportation hubs, stadium voice alarm and public address systems, classrooms, courtrooms, worship spaces, and hybrid meeting rooms. In many jurisdictions and verticals, intelligibility targets (often stated as STI, STIPA, or CIS equivalents) are tied to safety, accessibility, or performance requirements. Even where not mandated, STI offers a defensible baseline for system tuning and retrofit decisions because it can be measured repeatably and compared to targets over time.

This report template and analysis framework focuses on what audio practitioners need: the variables that move STI, how to measure them consistently, how to interpret results across listening areas, and how to turn findings into specific design or commissioning actions.

2) Key Factors (Variables) Analyzed

Signal-to-noise ratio (SNR) at listener positions: including HVAC noise, crowd noise, and operational noise sources.
Reverberation time and temporal decay behavior: typically RT60 and early-to-late energy balance; strongly linked to room volume and absorption distribution.
Direct-to-reverberant ratio (D/R) and loudspeaker coverage geometry: distance, aiming, directivity index (DI/Q), and array behavior.
System frequency response and speech band balance: octave-band energy distribution, tonal shaping, and excessive low-frequency buildup.
Nonlinearities and time variance: distortion, limiting/compression artifacts, and AGC behavior that can reduce modulation depth.
Alignment and arrival-time structure: delays between distributed loudspeakers, strong early reflections, and echoes that smear modulation.
Measurement method and uncertainty: STIPA vs full STI, test signal level, mic placement, calibration, and operational state during tests.

3) Detailed Breakdown of Each Factor (With Supporting Reasoning)

3.1 SNR: The Most Direct Lever

STI is fundamentally sensitive to how well amplitude modulations in speech survive at the listener’s ear. Background noise reduces modulation depth by filling in the valleys of the speech envelope, lowering the modulation transfer function (MTF) in multiple modulation frequency bands. Practically, this means the same room and loudspeaker layout can produce materially different STI results depending on HVAC state, occupancy, and nearby equipment noise.

What to document: A-weighted and octave-band ambient noise at each measurement position; operational conditions (HVAC mode, doors open/closed, typical equipment on); and the speech test signal level at the microphone position. Report SNR by band when possible, because ventilation noise often dominates low and mid bands, while projectors and electronics can elevate mid-high bands that matter for consonant cues.

Engineering implication: If STI is marginal and the room is already acoustically “treated,” raising speech level is not always the best fix—especially in voice alarm contexts where maximum SPL may be constrained by distortion, feedback margin, or listener comfort. Reducing ambient noise (silencers, diffusers, equipment relocation) frequently yields a more stable improvement than adding gain.

3.2 Reverberation and Temporal Smearing

Reverberation reduces intelligibility by smearing temporal modulations, particularly in faster modulation rates that align with syllabic and phonetic transitions. STI penalizes this because the modulated test signal loses contrast as reflections arrive beyond the temporal integration window. Rooms with long RT, insufficient absorption at mid frequencies, or poor diffusion can produce decent tonal balance but still measure weak STI.

What to document: RT60 (or T20/T30) in octave bands, at minimum 500 Hz to 2 kHz, where speech intelligibility is most sensitive. Note whether decay is smooth or exhibits flutter/late discrete reflections. Where possible, include EDT (early decay time) because early energy conditions can be more predictive of subjective clarity than long-tail RT alone.

Engineering implication: Adding absorption changes both RT and D/R. In large volumes, targeted treatment at reflection points and rear-wall control can yield more STI gain than uniform low-value absorption. For retrofit feasibility, compare improvements from acoustic treatment versus adding more directional loudspeakers that increase direct sound at listeners.

3.3 D/R Ratio and Coverage Geometry

STI benefits when the direct field is strong relative to late energy and noise. Loudspeaker directivity and placement determine D/R more than amplifier power does. As distance doubles, direct sound drops by roughly 6 dB in free field, while reverberant field may remain relatively constant once beyond critical distance. The result is predictable: farther seats trend lower STI unless compensated by distributed loudspeakers or tighter directivity.

What to document: Loudspeaker model and pattern control, mounting height, aiming angles, and the measured level uniformity across the audience plane. Include a coverage map with STI points overlaid so the relationship between geometry and results is visible.

Engineering implication: A single “loud enough” cluster can still underperform in STI at perimeter zones due to low D/R and increased reflections. Distributed systems reduce distance, improving D/R and level consistency, but create alignment complexity (see Section 3.7). The STI report should show whether poor results cluster with distance and off-axis listening, indicating geometry as the primary limitation.

3.4 Frequency Response and Speech Band Balance

STI is not a simple tonal metric, but frequency response matters because speech intelligibility depends heavily on mid-band energy and the audibility of consonant information (often 1–4 kHz). Excessive low-frequency energy can mask speech cues, and aggressive high-frequency roll-off reduces articulation. Even with good RT and SNR, an improperly equalized system can under-deliver STI.

What to document: Measured transfer function (or at minimum, octave-band SPL) at representative positions, with EQ targets and any system voicing applied. Note whether equalization is global or zone-specific, and whether any speech enhancement processing is engaged.

Engineering implication: In practice, STI improvements from EQ are usually incremental compared with improvements from SNR or D/R, but EQ becomes decisive when the system has band-limited components (small ceiling speakers, narrowband paging horns) or when LF masking is elevated by room modes and subwoofer coupling.

3.5 Nonlinearities, Limiting, and Dynamics Processing

STI assumes that modulation depth is preserved through the electroacoustic chain. Heavy compression, aggressive limiting, clipping, and certain noise reduction algorithms can flatten or reshape modulations. In emergency paging, the system may be run close to maximum output, making distortion and limiting common failure modes even when coverage plots look correct.

What to document: Processor settings (compression ratios, limiter thresholds, AGC behavior), amplifier headroom, loudspeaker power handling, and measured distortion indicators if available. Include the test signal level relative to operational paging level; STI measured at a quiet commissioning level can overstate real-world performance.

Engineering implication: If STI drops at higher playback levels, the system is likely hitting nonlinear limits. Remedies are typically architectural (more loudspeakers, more efficient/appropriate models, different zoning) rather than “more EQ.” The report should explicitly compare STI at normal and peak operational levels where safety messaging must remain intelligible.

3.6 Time Variance and Environmental Variability

Occupied conditions change both noise and absorption. Crowds add broadband noise (reducing SNR) while also increasing mid/high absorption (potentially reducing RT). Which effect dominates depends on venue type. Transport stations and arenas often see STI degrade during peak noise, while theaters can see modest RT improvements in occupied conditions but still face SNR constraints for unamplified or low-level announcements.

What to document: Measurement conditions and expected operational scenarios. If measurements are performed unoccupied, the report should state likely deltas and whether additional tests under representative noise conditions were performed.

Engineering implication: Commissioning targets should be stated with context (e.g., “STI ≥ 0.50 under nominal ambient noise of X dBA”). Without this, pass/fail decisions can be contested because STI is not an intrinsic property of the room alone.

3.7 Alignment, Echoes, and Multi-Source Interactions

Distributed loudspeakers and fills can improve D/R but introduce inter-source timing issues. If arrivals are misaligned, listeners receive multiple comparable-level signals separated in time, reducing modulation clarity and producing echo perception. STI will typically reflect this as reduced modulation transfer, especially in mid bands.

What to document: Delay settings, reference alignment strategy (e.g., precedence-based, main-to-fill timing), and any locations where two sources are within a few dB of each other. Include impulse response snapshots or note audible echoes where observed.

Engineering implication: The STI report should not only report “low STI here,” but also identify whether the cause is noise/RT versus multi-source interference. The corrective actions differ: alignment and level shading address timing conflicts; acoustic treatment and noise control address smear and masking.

4) Comparative Assessment Across Relevant Dimensions

An STI report becomes decision-useful when it compares results across zones, operating states, and system configurations. For audio professionals evaluating upgrade paths, the following comparisons typically expose root causes:

Zone-to-zone STI distribution: median, minimum, and percentage of positions above target per zone (floor, balcony, concourse, platform, classroom rear seats). Averages alone can hide failure pockets.
STI vs distance/off-axis angle: identifies geometry-limited performance and validates whether tighter directivity or more zones are needed.
STI vs ambient noise condition: HVAC off/on, doors open/closed, simulated crowd noise; isolates SNR sensitivity.
STI at nominal vs maximum paging level: reveals limiting/distortion impacts and loudspeaker headroom constraints.
Before/after changes: EQ changes, delay alignment updates, added absorption, added loudspeakers; links investments to measurable gains.

For reporting, include a table with each measurement point: ambient noise, playback level, STI, and notes (audible echo, strong reflections, mechanical noise nearby). Then summarize by zone with percentile statistics and a compliance map.

5) Practical Implications for Audio Practitioners

System design: STI analysis pushes designs toward controlled directivity, shorter throw distances, and predictable coverage. In practice, this often favors distributed ceiling systems in low ceilings, column arrays in reflective worship spaces, or steered arrays where architectural constraints prevent optimal placement. The report should connect any low-STI areas to the geometry that creates low D/R.

Commissioning workflow: STI measurement should be scheduled after alignment, EQ, and limiter settings are finalized, and under defined operational conditions. If the end-user expects intelligibility during busy periods, incorporate representative noise. A commissioning report without documented conditions can misrepresent performance and complicate acceptance.

Troubleshooting: When STI is low, practitioners should triage in this order: (1) SNR (noise control or level strategy), (2) D/R (coverage and directivity), (3) time alignment (multi-source interactions), (4) reverberation control, (5) processing artifacts and tonal shaping. This order reflects how strongly each factor typically moves STI in real installations.

Procurement and upgrade decisions: STI reporting supports cost-effective choices. If low STI is driven by ambient noise, upgrading loudspeakers may provide limited return compared with HVAC mitigation. If low STI correlates with distance and off-axis positions, additional zones or narrower directivity yields clearer gains than adding amplifier power.

6) Data-Driven Conclusions and Recommendations (Report Template)

6.1 STI Report Template (What to Include)

Project details: venue, date/time, measurement standard/method (STI or STIPA), instrument model, calibration reference, and operator.
System description: loudspeaker types/quantities, zoning, processing chain, paging/music routing, and operational modes tested.
Test conditions: occupancy state, HVAC state, doors, typical noise sources active, and any simulated noise approach used.
Measurement grid: map and coordinates, microphone height, number of positions per zone, and rationale for point selection.
Per-point dataset: ambient noise (dBA and/or octave bands), playback level (dBA and/or octave bands), STI value, and observations.
Summary statistics: min/median/mean STI per zone, percentage of points above target, and identification of outliers.
Diagnostics: RT metrics, evidence of echoes or misalignment, processing headroom checks, and frequency response snapshots where relevant.
Corrective actions: prioritized list with expected mechanism of improvement (SNR, D/R, alignment, RT) and verification plan.

6.2 Conclusions and Recommendations Framework

Conclusions should be stated as findings tied to measured variables, not as general assessments. Examples of report-grade conclusions that remain evidence-based:

If STI tracks ambient noise: “Positions with ambient noise ≥ N dBA show STI reduced by X–Y compared with quieter positions at similar distance; mitigation should prioritize noise control or revised operational level strategy.”
If STI tracks distance/off-axis: “Rear/perimeter zones show consistently lower STI and reduced direct level, indicating D/R limitation; add distributed loudspeakers or increase directivity control to reduce throw distance.”
If STI drops at high level: “At operational paging level, STI decreases while system limiting indicates reduced modulation depth; increase headroom via additional loudspeakers/zones and revise limiter thresholds.”
If discrete timing issues appear: “Low-STI points coincide with overlapping sources within a few dB and delayed arrivals; implement precedence-based alignment and level shading, then re-verify STI.”

Recommendations should specify verification criteria (e.g., retest under the same noise state, same grid, same playback level) and focus on closing the gap between measured zone performance and the target threshold. For decision-makers, the most useful final deliverable is a zone-by-zone compliance statement plus a prioritized corrective plan that identifies the physical mechanism behind each expected improvement.

When STI reporting is executed with controlled conditions, sufficient spatial sampling, and supporting diagnostics (noise, RT, alignment), it becomes more than a compliance checkbox. It functions as an engineering decision tool: it identifies whether intelligibility is limited by the room, the system, or the operating environment, and it quantifies which interventions are most likely to move results in measurable, repeatable ways.