Human Factors: Speech Transmission Index (STI) (1970-2005+)

In 1970, Houtgast and Steeneken working at the TNO Instituut voor Zintuigfysiologie (‘human factors’) in Soesterberg developed the Speech Transmission Index (STI). The absolute measurement of speech intelligibility is complex. STI is a method to determine the quality of speech transmission via communication channels such as through electro-acoustic equipment, a radio or satellite link, etc., using test signals.
The STI depends on:

• the sound level of the speech
• background noise at the sending side
• quality of the sound reproduction equipment
• the frequency response of the transmission channel end-to-end
• non-linear distortions in the transmission path
• background noise at the receiving side
• echoes (reflections with delay > 100ms)
• the reverberation time
• psychoacoustic effects (masking effects)

 
In 1971, the Speech Transmission Index Device using Artificial Signals (STIDAS) was developed by TNO. STIDAS was an apparatus for measuring the STI of communication channels using an artificial signal with a measuring time below three seconds. With this device, an objective indication of the performance of a communication channel to transport speech could be determined. The method is based upon (1) a simple artificial test signal by which a ‘standard’ spectral difference is introduced at the talker side of the channel, and (2) an analysing procedure to be applied to the test signal received at the listener side to quantify the degree of preservation of the spectral difference introduced.

In 1979, a fast-to-measure Room Acoustics STI (RASTI) method and RASTI-measuring devices were developed.

In 1980, STI was accepted by the Acoustical Society of America as a standard. Special measuring hardware and software were developed by the industry to measure STI.
In 1988, the RASTI method became an international standard: IEC 60268-16 Sound system equipment – Part 16: Objective rating of speech intelligibility by speech transmission index. This standard has been revised several times but is still based on TNO’s initial and subsequent work in this area.

In 1999, the publication ‘Evaluatiemethode voor operationele spraakcommunicatiesystemen‘ followed, internationally known as ‘Speech Transmission Index for Public Address Systems‘ or STIPA. This method has been included in several national standards for public address systems for emergencies. A 2004 leaflet about TNO’s services in this area can be found here.

Work at TNO on the Speech Transmission Index continued over a long period. A wide range of STI measurements has been performed both for Defence and for public and private entities, for instance on public address systems for road tunnels, gas masks, cockpit systems, and theatres.

In 2004, a publication “Using the Speech Transmission Index for predicting non-native speech intelligibility” followed. Work was performed on how to interpret STI values when non-native talkers or listeners are involved. Based on subjectively measured psychometric functions for sentence intelligibility in noise, for populations of native and non-native communicators, a correction function for the interpretation of the STI can be derived. This function can be applied to determine the appropriate STI ranges with qualification labels ‘bad’ to ‘excellent’ for specific populations of non-natives. The correction function is derived by relating the non-native psychometric function to the native psychometric function by a single parameter (v). For listeners, the parameter is found to be highly correlated with linguistic entropy. The proposed correction function is also valid for conditions featuring bandwidth limiting and reverberation.

References/further reading

Houtgast, T. and Steeneken, H.J.M. (1971), “Evaluation of Speech Transmission Channels by Using Artificial Signals”, Acustica 25, 355–367. download location
The measurement of Speech Intelligibility, Herman J.M. Steeneken, TNO Human Factors, Soesterberg, the Netherlands

More on the STI can be found on Wikipedia.