Akadēmiskais centrs
Abstract:
Time-frequency analysis can be described as Fourier analysis simultaneously both in time and in frequency. Its origins are in quantum mechanics, in signal processing, and in pseudo-differential operators. A time-frequency transform is a translation-modulation-invariant sesquilinear mapping from test functions in time to functions in time-frequency plane. The class of time-frequency distributions is further restricted by imposing extra conditions, trying to capture the idea what an idealized energy density could satisfy. Sharp time-frequency resolution would allow sharp operations on signals. Spectrograms are widely used examples of time-frequency distributions, but they have always bad time-frequency resolution; however, spectrograms can be replaced by sharp methods. We describe the time-frequency analysis for both analog and digital signals, in both non-periodic and periodic cases. The involved Fourier integral formulas are naturally discretized, and computed fast with Matlab. Real-life examples will be demonstrated.
Aicināti visi interesenti.
Semināram būs iespēja pieslēgties arī attālināti MS Teams platformā.
(Interesenti piedalīties attālināti aicināti pieteikties semināram līdz 11. oktobrim, rakstot ziņu uz e-pastu sergejs.smirnovs@lu.lv.)
 ar referātu Time-frequency transforms, with applications.){kind=link}