We introduce and analyze a method for stabilizing test parameters and estimating the nonlinearity of mildly non-linear quantizers by Fourier analysis of the quantizer’s output to sinusoidal input signals. Stabilization for the combined effects of the rounding operation of the quantizer and the variations in offset and amplitude of the applied sinusoid is achieved by a deterministic form of subt...

An improved version of the original parametric formulation of the generalized spectral subtraction method is presented in this study. The original formulation uses parameters that minimize the mean-square error (MSE) between the estimated and true speech spectral amplitudes. However, the MSE does not take into account any perceptual measure. We propose two new short-time spectral amplitude esti...

Using femtosecond time-resolved x-ray diffraction, we directly monitor the coherent lattice dynamics through an ultrafast charge-density-wave-to-metal transition in the prototypical Peierls system K(0.3)MoO(3) over a wide range of relevant excitation fluences. While in the low fluence regime we directly follow the structural dynamics associated with the collective amplitude mode; for fluences a...

Being motivated by the widely accepted principle that the more training data we have, the better performance the recognition system has, we conducted experiments asking human subjects to do test on a mixture of real English handwritten textlines and textlines altered from existing handwriting with various distortion degrees. The idea of generating synthetic handwriting is based on a perturbatio...

The advancement of electronic integrated circuit technology and photonics in the last decade have allowed the design and demonstration of ultra-high bit-rate transmission over advanced optical fiber transmission line employing in-line optical amplification as well as electronic compensation and equalisation of linear and nonlinear distortion effects. The research and development of novel modula...

For one thing, some measurements which are very hard in the time domain are very easy in the frequency domain. Consider the measurement of harmonic distortion. It's hard to quantify the distortion of a sine wave by looking at the signal on an oscilloscope. When the same signal is displayed on a spectrum analyzer, the harmonic frequencies and amplitudes are displayed with amazing clarity. Anothe...

Audio rendering systems are always slightly nonlinear. Their non-linearities must be modeled and measured for quality evaluation and control purposes. Cascade of Hammerstein models describes a large class of non-linearities. To identify the elements of such a model, a method based on a phase property of exponential sine sweeps is proposed. A complete model of non-linearities is identified from ...

The well known and established methods used for distortion measurements cannot be used for detecting the errors that appear in digital audio systems based on perceptual coding and decoding. A test stimulus has to be used that has the time-varying properties similar to real signals such as speech and music. Furthermore, the procedure of quality evaluation has to include an estimate of the tempor...

A directional sound beam can be generated by a parametric loudspeaker system through the nonlinear interaction between finite-amplitude ultrasonic waves in air. However, this nonlinear interaction also produces harmonic components in addition to the desired audible sound. In order to investigate this nonlinear phenomenon, a baseband distortion model is developed from nonlinear system identifica...

Designing a neural network (NN) for processing complex signals is a challenging task due to the lack of bounded and differentiable nonlinear activation functions in the entire complex domain C. To avoid this difficulty, 'splitting', i.e., using uncoupled real sigmoidal functions for the real and imaginary components has been the traditional approach, and a number of fully complex activation fun...