As the speech signal has some redundancy, some loss of information during transmitting is considered acceptable to provide the sufficient intelligibility. However, the allowed value of the transmitted signal loss depends significantly on ambient conditions. Because of different noise, distortions and reverberation interference, the total loss of information may be so considerable that it may lead to the impossibility of speech understanding, in other words, to decreasing of speech intelligibility level. Thus, the communication systems which provide the sufficient speech intelligibility in normal conditions may be inapplicable in case of the high-level ambient noise.
Overtones
It is commonly believed that the band of the main speech signal energy is quite narrow: about 500 Hz - 2 kHz. That is why the standard telephony systems are intended for the full operating frequency band within 300 Hz – 3,5 kHz. This band is considered sufficient for voice transmitting under normal conditions.In fact, the human voice contains a wider frequency spectrum.
The frequency components beyond the main range of sounds are referred to as overtones. Under normal communication conditions they are redundant information and can be discarded without any sufficient loss for the message understanding. However, in the conditions of strong noise, the “main” sounds merge into one and become illegible. In this case, overtones become crucial for the message understanding. Overtones are so-called “punctuation marks” which help divide the sound stream into separate words and phrases which increase the speech intelligibility.
Standard and Extended Bandwidth
The diagrams below show important differences between the standard telephony and industrial communication.The left diagram shows the standard 3.5 kHz bandwidth, while as at the right diagram the transmitted spectrum is extended to 7 kHz which allows transmitting not only the main sounds of the voice (the yellow field at the diagram) but also the more high-frequency overtones. If you combine the typical wavelength characteristic of industrial noise (which overlaps significantly the main sound information field) with the diagrams, you will see that the transmission efficiency of the signal with the extended bandwidth is higher. This will lead to the intelligibility increasing.
Practical Application
In telephony the digital sound signal is transmitted via channels with the traffic capacity of 64 kbit/s which provides the 3.5 kHz bandwidth. Combining these two channels for transmitting one sound signal, it is possible to extend the bandwidth up to 7 kHz; it will guarantee the higher speech intelligibility under conditions of strong ambient noise. On the other hand, this solution increases the specific cost of the communication system as it requires more resources for each subscriber.In this case, if you combine not only two channels but three (or more) channels, extending the bandwidth up to 10.5 kHz (or higher), the results won’t compensate the expenses. It is bound to the fact that it is hardly possible to distinguish between 7 and 10,5 kHz by ear. Second, the majority of high-performance horn speakers used at industrial enterprises have the bandwidth of about 7-8 kHz which won’t allow reproducing sound signals of the higher quality.
Thus, choosing of the effective solution for communication system requires taking into account the field operating conditions and the balance of quality of the voice transmitting and the resource intensity of the system. While developing communication systems intended for strong ambient noise conditions, one of the main requirements is the speech intelligibility. In this case, the specialized systems in industrial communication are the most efficient solution.
