ABSTRACT
The stationary-gradient method requires the lowest investment in technology for one-dimensional experiments. Numerous methods of solving deconvolution problems have been published. The goal is the deconvolution of the measured distribution to of the measured spectrum S, with concentration invariance of the spectrum presupposed. The general problem with deconvolution is the division by zero or values near zero. Upon deconvolution of noisy functions, the results can be distorted by oscillations of the sine function which suggests that the window function is too small. The measured spectrum was used for the deconvolution by the measured derivative spectrum. In the following discussion, deconvolution is described for applications without prior knowledge of the spectrum parameters. Noise limit is chosen greater than the squared noise amplitude in the power spectrum and is derived from the signal-to-noise ratio of the measured spectrum S and to. The spectrum is not composed of only different separated lines.
