Before and after. If you've ever upgraded your antenna from the proverbial piece of wet string to a six element beam, the improvement will have been nothing short of staggering. The change would have been easy to quantify using 'before and after' signa strength measurements. In other areas, judging whether a change has been effective is so much harder. If you add a speech processor, for example, your received peak signal strength will remain the same. Similarly, adding a noise filter won't change the strength of incoming signals at all. Yet in both cases, and in many others where you won't see a change in signal strength, you would hope that the effort had not been in vain. Of course, in time, an improvement in your DXCC score might just hint at a beneficial change to your station but surely there's a better way of assessing changes to your line-up. What is needed is an accurate and repeatable method of assessing the intelligibility of speech.rnIntuitively, we might expect that it will be blatantly obvious if a change has achieved the desired beneficial effect but experience suggests otherwise. I'm involved in the development of cave radios (see RadCom May 2008 and creg.org.uk for more details) that are used by the UK's volunteer rescue groups to transmit through solid rock to establish contact between the surface and an underground rescue team. These radios operate in the LF spectrum where they are plagued by interference from the LORAN-C navigation system that operates on 100kHz but has very wide sidebands. I recently conducted tests on a number of noise filters aimed at the amateur radio market in the hope that they would reduce this interference. All the filters gave the impression that the interference was reduced and that the speech was more readable. It was a big surprise, therefore, when a formalised speech intelligibility test proved that, in this particular application, the intelligibility had actually been reduced! This provided a clear indication that some sort of formal metric is required in assessing the readability of a signal.
展开▼
