Abstract :
A study is made of the theoretical signal/noise gain of an ideal strip transducer; this gain is the actual signal/noise ratio compared with that for a `point¿ receiver (i.e. one whose dimensions are small compared with the wavelength at the lowest significant frequency). The results are not restricted to narrow-band systems and relate generally to noise fields which may be non-isotropic and have non-uniform frequency spectra. The theory is used to investigate the correlation between the noise outputs of the various sections of a subdivided transducer, and it is shown that, at least in the narrow-band case, the usual assumption of negligible correlation is generally justified in practice. Multiplicative systems are also considered, and it is shown that, in certain circumstances, the overall signal/noise ratio of such a system may be better than that of either of its component groups. The results obtained are also generally applicable to ideal radio receiving arrays.
