Main Page Related Pages A/V Systems Classroom Intelligibility Cluster Design for Intelligibility Intelligibility Calculator Loudspeaker Directivity Sound System Design Video Conference Facilities |
 |
Design for Speech IntelligibilityContrary to popular belief, it is possible to define, design for, and predict speech intelligibility in sound system design. Many people (including design professionals in architecture and engineering disciplines) seem surprised that it is possible to predict the speech intelligibility of sound systems before they are installed. The performance parameters that govern speech intelligibility for most germanic and western languages have been well documented and researched. More research is coming to light on preferences for other language groups as well. Language develops to suit the human organism, so most languages share the same important criteria, since we all have the same general shape of ear. Because of this universal application of human speech characteristics, we can design sound systems to deliver speech intelligibility anywhere in the world. We can also identify sound system deficiencies that cause poor speech intelligibility. Many sound systems that sound just fine for music playback do not deliver good speech intelligibility. This is the reason that any sound system intended for speech reinforcement should be demonstrated with recorded speech or a live microphone, not a music recording. The ear / brain processor can fill in a substantial amount of missing information in music, but requires more detailed information for understanding speech. The speech power is delivered in the vowels (a, e, i, o, u and sometimes y) which are predominantly in the frequency range of 250Hz to 500Hz. The speech intelligibility is delivered in the consonants (b, c, d, f, g, h, j, k, l, m, n, p, q, r, s, t, v, w) which requires information in the 2,000Hz to 4,000Hz frequency range. People who suffer from noise induced hearing loss typically have a 4,000Hz notch, which causes severe degradation of speech intelligibility. Most telephones have a limited usable bandwidth of less than 4,000Hz, and you may have noticed that it can make it difficult to hear the difference between the letters b, d, t, and v, and also between f and s, on the phone. Vowels have more sound energy available as well, the consonants that give us the detail and information in speech are quite low in level. Just think of how loud you could shout the letter "O" versus the letter "T". It is ironic that the most critical aspect of speech is also the lowest in energy, but then our ear has a hearing mechanism that is finely tuned to 2kHz to make it very sensitive to sounds in that range. There are many everyday examples that point out the difference in energy between vowels and consonants in speech, oddly enough, English language operas are a prime example, but a more familiar one may be Broadway stage musicals. There's a good reason why the Broadway musical that became famous was set in "Oklahoma" and not "Tennessee", because no one in the audience would have been able to hear them sing Tennessee on a Broadway stage, there are so many consonants and so few vowels to allow the singers to belt out the word at high volume. And they couldn't have held that first note the way they did in... "OOOOOOOOOOOOOOOOOOOOOOOklahoma, where the wind comes sweeping down the plain." (Note: To all people from Tennessee, it's nothing personal, it was just an example of a state that started with a consonant and had a challenging combination of consonants in the name, we could have picked Texas, or Saskatchewan on our side of the border... "SSSSSSSSSSSSSSSSSSSSSSSSSSSSSaskatchewan, where the arctic winds come sweeping cross the plain") It is quite easy to build a sound system that does not provide uniform coverage in the critical speech intelligibilty bands. In fact it seems that many sound systems are assembled without adequate attention to this detail. Many sound systems are built using packaged "Rock and Roll" loudspeakers, and the resulting sound systems can be loud enough and yet still be deficient in speech intelligibility. We are often asked to evaluate existing sound systems for this deficiency. Sound systems in drama theatres and churches are often sold on the basis of their ability to deliver terrific sounding music, but fall short on the delivery of speech coverage. It is very important to identify the primary purpose of a sound system when designing it. It is much easier to extend the frequency range of well designed speech system so that it can provide pleasant sounding music, than it is to add speech intelligibility to a music system without uniform mid frequency coverage. If you have a JavaScript capable browser you can try this calculator to see what factors influence the calculation of the expected speech intelligibility: JavaScript Intelligibility Calculator Direct sound source lacks intelligibility |
