The Common Intelligibility Scale (CIS) was proposed by Barnett and Knight of AMS Acoustics in 1995. 

It is a method of relating all potential measures of speech intelligibility to a single reference.  This reference was subsequently embodied into IEC 60849 “Sound Systems for Emergency Purposes”, the International Standard. 

At the time the CIS was conceived RASTI was the primary objective measure of Speech Intelligibility.  The relationship between RASTI and Phonetically Balanced (PB) wordscores was given by Houtgaast and Steenehen the inventors of STI/RASTI. 

In order to determine the relationship between the metrics in more detail, the AMS Acoustics team looked to PB wordscores. 

ANSI S3.2-1989 contains 20 lists of phonetically balanced words; the phonetics of US English appear in each list equally.  Firstly we analysed the lists and then used voice trained actors to record the lists anechoically.  This was a significant challenge as the actors were required to maintain a steady speech rhythm, rate, tone and timbre. 

These lists can be played through a sound system into any space and recorded binaurally at a number of positions. 

The recordings are presented to a jury using calibrated, monitor quality headphones. 

A proper wordscore measurement is robust and repeatable, not a subjective judgement made by an individual under specific circumstances, often referred to as a Golden Ears Test. 

A “Golden Ears Test” is about as useful for determining speech intelligibility as a “wet finger” in the air to determine wind direction, speed and temperature or a “toe in the water” to determine temperature, content and tidal activity.  It is too often overlooked that we are in fact trying to adjudicate a contractual pass or fail scenario.