Back

STI / STIPA | SonaVyx

IEC 60268-16

Speech Intelligibility Measurement

Single 15-second STIPA measurement. Position your device at the listener location.

RT60 Reverberation Time

Measure room reverberation time. Excessive RT60 is a primary cause of low STI.

Impulse Response Analysis

Capture room impulse response for full STI computation via wasm_sti_from_ir.

SPL Meter

Measure sound pressure level to assess signal-to-noise ratio for STI prediction.

Understanding Speech Transmission Index (STI) and STIPA Measurement

The Speech Transmission Index (STI) is the internationally accepted metric for quantifying speech intelligibility in acoustic environments. Defined by IEC 60268-16, STI evaluates how faithfully a transmission channel preserves the temporal modulation patterns of speech, producing a single number between 0 (completely unintelligible) and 1 (perfect clarity). STI is indispensable for the design, commissioning, and verification of public address systems, voice alarm installations, classrooms, courtrooms, houses of worship, airports, railway stations, and any venue where spoken communication is safety-critical or functionally important.

The measurement works by analysing the Modulation Transfer Function (MTF) across seven octave frequency bands centred at 125 Hz, 250 Hz, 500 Hz, 1 kHz, 2 kHz, 4 kHz, and 8 kHz. Within each band, the modulation depth of a test signal is compared before and after transmission through the acoustic channel. Reverberation, noise, echoes, and nonlinear distortion all reduce modulation depth, which lowers the measured MTF and therefore the STI. The band-specific MTF values are weighted and combined according to the IEC 60268-16 algorithm to produce the final STI score.

STIPA: The Practical Field Measurement Method

STIPA (Speech Transmission Index for PA systems) is a standardised simplification of the full STI method that reduces measurement time from minutes to just 15 seconds. STIPA uses a specially designed test signal containing two modulation frequencies per octave band (14 modulation frequencies total) rather than the 98 required for full STI. Despite this simplification, STIPA produces results that correlate with full STI to within +/- 0.03, making it the preferred method for on-site PA and voice alarm system verification to EN 54-16, IEC 60849, and BS 5839-8.

SonaVyx implements STIPA measurement entirely in the browser using a Rust-compiled WebAssembly DSP engine. The test signal is generated via wasm_generate_stipa_signal, played through the device speaker, captured by the microphone, and processed through wasm_sti_from_ir to extract the modulation transfer function and compute the final STI value. For environments where a measured impulse response is already available, the wasm_sti_from_rt60_snr function can estimate STI from RT60 and signal-to-noise ratio data without requiring a live measurement.

STI Rating Scale and Compliance Requirements

The qualitative interpretation of STI values follows a standardised five-level scale. Scores below 0.30 are rated Bad, indicating that less than half of spoken words will be understood. Scores between 0.30 and 0.45 are Poor, typical of reverberant spaces with inadequate sound reinforcement. The Fair range (0.45 to 0.60) represents the minimum threshold for many building codes and PA system standards. Good intelligibility (0.60 to 0.75) is the target for classrooms under BB93, ANSI S12.60, and most voice alarm systems. Scores above 0.75 are rated Excellent and indicate that speech is clearly understood even by listeners who are not native speakers of the language being spoken.

Factors Affecting Speech Intelligibility

Several acoustic parameters directly influence STI. Reverberation time (RT60) is often the dominant factor: as RT60 increases beyond approximately 1.0 second, late reflections smear temporal modulation and reduce intelligibility. The relationship between RT60 and STI is nonlinear and frequency-dependent, which is why octave-band analysis is essential. Background noise reduces the signal-to-noise ratio (SNR) at the listener position, masking modulation patterns particularly in the 500 Hz to 4 kHz range that carries the majority of speech consonant information. Echoes and discrete reflections arriving more than 50 milliseconds after the direct sound further degrade modulation by creating temporal overlap between successive syllables. System nonlinearity and distortion in amplifiers, loudspeakers, and signal processing can introduce additional modulation artifacts that reduce the effective MTF.

By combining STIPA measurement with RT60 reverberation analysis, impulse response capture, and SPL metering, engineers can identify the root cause of poor intelligibility and make targeted acoustic or electroacoustic corrections. The SonaVyx platform provides all of these measurements in a single browser-based toolkit, eliminating the need for expensive dedicated hardware while delivering results traceable to international standards.

Frequently Asked Questions

What is Speech Transmission Index (STI)?
STI is a standardized metric defined by IEC 60268-16 that quantifies speech intelligibility on a scale from 0 (completely unintelligible) to 1 (perfect intelligibility). It accounts for how a transmission channel (room, PA system, communication link) degrades the modulation of speech signals across seven octave bands from 125 Hz to 8 kHz.
What is the difference between STI and STIPA?
STIPA (Speech Transmission Index for Public Address systems) is a simplified variant of the full STI measurement that uses only two modulation frequencies per octave band instead of fourteen. STIPA produces results highly correlated with full STI (within +/-0.03) but requires only a 15-second measurement instead of several minutes, making it practical for field use in PA and voice alarm system commissioning.
What STI value is considered acceptable for a PA system?
For public address and voice alarm systems, IEC 60849 and EN 54-16 require a minimum STI of 0.50 (Fair). For critical applications such as emergency voice communications, an STI of 0.60 or higher (Good) is recommended. Classrooms typically target STI >= 0.60, while concert halls and theatres aim for STI >= 0.75 (Excellent) for unamplified speech.
How does ambient noise affect STI measurements?
Background noise reduces the signal-to-noise ratio (SNR) at the listener position, which directly degrades modulation transfer and lowers the STI value. Measuring ambient noise before the STIPA test allows the system to account for the noise floor. Typical masking effects mean that even moderate background noise (NC-35 to NC-40) can reduce STI by 0.10 to 0.20 points, particularly in the 500 Hz to 2 kHz range critical for speech consonant recognition.
Can I use a phone or laptop microphone for STI measurement?
While a calibrated measurement microphone gives the most accurate results, modern laptop and smartphone microphones can provide useful indicative STI readings, especially for comparative measurements within the same room. The browser-based STIPA measurement in SonaVyx compensates for typical consumer microphone frequency response curves. For formal compliance testing to IEC 60268-16, a Class 1 or Class 2 measurement microphone is recommended.