News and Projects

You are here: Home : News : Our contributions At The Royal Society, London

Our contributions At The Royal Society, London

26th Oct, 2006 to 27th Oct, 2006

Its not every day you are invited to express your own professional views at the most prestigious and well known establishment in the World, most famously noted for announcing cutting edge scientific achievements.

I was very privileged to be able to attend a lecture at The Royal Society on the Mall, overlooking St James Park, London. This feature also includes an insight into our acoustic work with speech intelligibility, and the way it is measured. 

Founded in 1660, the Royal Society's purpose has been (and still is) to promote and advance the understanding of science, technology and engineering worldwide. Many eminent scientists have been elected as Fellows of the Society including Isaac Newton, Charles Darwin, Ernest Rutherford, Albert Einstein, Dorothy Hodgkin and Stephen Hawking. Unlikely that I would be immortalised on the walls of this great place, but I was just very happy to be there, and making some contribution.

The fundamental purpose of the lecture was to discuss and demonstrate different ways of measuring speech intelligibility to the controlling standard IEC 60268-16 (2). This is a vital part of our sound and acoustic work, as there is little point in having a sound system installed unless you can actually hear it when it's on. 

The Speech Transmission index..

For those who are really not interested in the jargon of how to measure speech, now is the time to go to the next article! As scientists, engineers, consultants we need to check the speech transmission index (STI) in a vast number of situations of everyday life. For example, imagine not being able to hear flight announcements clearly at an airport, or not being able to hear which train was arriving next, and on which platform.

Sometimes, it is life threatening; military radio equipment needs to be tested for speech intelligibility as vital messages need to be carried, pilots headsets undergo rigorous testing, and the law will no longer tolerate muffled weak sounding PA systems for evacuating large gatherings in case of emergency. These are vital issues. 

How do we test STI?

This is a very complex subject, but we have adopted a number of ways of testing STI. One such method is the Modified Rhyme Test (MRT). This test uses sets of closely rhyming monosyllable words such as pin, sin, tin, fin, din, and so on, that are specifically selected with half to differ in the initial consonant and the other half in the final consonant. The listener's task is to respond to the stimulus by indicating which of the six rhyming words presented before them is being said. 

Word lists

Phonetically balanced word lists (sometimes referred to as PB-50) have also been used for STI testing. This time the test data consists of 50 word lists that are phonetically balanced (such as piece, place, planes, play and so on). Again, the listener is asked to evaluate, and write down the stimulus, whatever he/she can hear.

As you can imagine, these tests were invented a long time ago, and although very useful as a full STI test, does not in fact guarantee total accuracy. For instance, some words can be unfamiliar, some too familiar, smaller numbers of people may not represent a much larger crowd with background noises, listener distractions etc. It is interesting to note, in an identical experiment, with a neutral talker, Australians scored lower points than UK counterparts. This anomaly gives rise to another interesting discussion that phonetics are often perceived differently between races.     

RASTI

In 1979, a quicker instrument-based test known as the room acoustic speech transmission index (RASTI) was developed using modulated signals. This simplified the testing procedure, but the original RASTI test had limitations due to limited bandwidths being tested, and the results often showed a good measurement when background noise became higher. This is obviously not consistent with good measurement as we all know if someone is talking loudly at the theatre next to you, it is impossible to hear the best of actor's voices on the stage properly.

A new standard?

STIPa (speech transmission index public address) is fast being adopted as a more accurate standard for STI testing. It still uses a modulated signal instead of rhyming words, but more frequencies (and thus more combinations) are used in the test. This means, there is less error in subjective measurements, and results are far more consistent.
The test is carried out with a small loudspeaker and a measurement device at the listener position. A CD with the modulated carrier is played through the speaker, and the measurement taken. The entire test can be done in 15 seconds, a far cry from the long rhyming tests, which can take hours, especially if a large number of readings are to be used. There is no chance of listener error, and the test also takes into consideration background noise. Here endeth the brief lesson!

Our contribution..

One contribution I put forward along with Leon Pieters (MD of Ampetronic induction loop equipment) was to see much tighter regulation on the installation of audio and induction loop systems, especially for the hard of hearing. Very few that I have been asked to check as a consultant, meet RNID specification, and even less attention is paid to pick microphone balance, which is vital if they are to hear it properly. Our extensive work in hearing assisted systems have been developed over considerable time by a mixture of practical, and theoretical input.    

It already looks like our comments regarding STI measurement will be heard (some even implemented) at the next IEC conference, which is where the legislation is made, and agreed for Europe.   

The old adage of being in the right place, at the right time, is true. Like many others, I always considered the Royal Society to be the most respected platform on which to air your views and discoveries. What I didn’t realise was that if your comments were voiced at The Royal Society, you can make changes for the better. Now, I think that is an achievement!