Abstract

The nonlinear characteristics of subcarrier network electronic, optoelectronic, and optical components set an upper limit to the available carrier to noise ratio (CNR). Conventionally, the harmonic and intermodulation distortion of system components such as semiconductor lasers has been assessed with multitone tests (e.g., Ref. 1). While these measurements give an accurate account of laser nonlinearity, they do not reflect the spectral characteristics of active communication channels. The statistical properties of the subcarrier channel signal ensemble must also be considered since peak signal clipping can introduce further CNR degradation.² These issues are common to frequency-division-multiplex radio systems where it has been found that Gaussian statistics only apply with more than sixty speech signals (e.g., Ref. 3). Consequently, a new test procedure for the emerging subcarrier optical network systems is required.

© 1990 Optical Society of America