Jitter reduction device and method

1. A method, instantiated in a non-transitory computer readable medium storing programming instructions, executable by a processor, for reducing the effect of jitter in a sampled signal, the method comprising:

obtaining a frequency-domain data set representing the sampled signal;

obtaining an average sideband amplitude distribution generated by jitter around one or more principal frequencies of the signal;

estimating jitter phases for sidebands generated by jitter in the frequency-domain data set;

subtracting a jitter contribution in the sampled signal characterized by the average sideband amplitude distribution and the estimated jitter phases from a data record.

2. The method of claim 1 , further comprising a step of selecting a set of correlated sideband components that are correlated in phase and amplitude in the frequency-domain data set.

3. The method of claim 1 , wherein the jitter introduces a phase modulation in the sampled signal.

4. The method of claim 2 , wherein the step of estimating the jitter phases includes a step of finding, for each correlated set of sidebands, a phase which minimizes the sideband energy after the step of subtracting the jitter contribution.

5. The method of claim 2 , wherein the correlated sidebands components are sideband components of the same frequency offset below and above the principal frequencies.

6. The method of claim 2 , wherein the step of estimating the jitter phases comprises a step of creating a correlation mask from the frequency-domain data set.

7. The method of claim 6 , wherein the correlation mask is obtained by: multiplying each frequency bin of the frequency-domain data by a its scalar frequency value, obtaining a weighted spectrum; shifting left and right the weighted spectrum of a number of bins equal to the order of the sideband component whose phase is being estimated and multiplying the weighted spectrum by A m e jθ and by A m e j(π-θ) , A m being a value of the average sideband amplitude distribution, j the imaginary unit, and θ a variable phase; add together the resulting left-shifted spectrum and right-shifted spectrum.

8. The method of claim 7 , further comprising a step of finding the phase of the mask that minimizes the energy in the data record when subtracted from an original data record.

9. The method of claim 1 , wherein the jitter phases are estimated as equal to the phase of the sampled signal.

10. The method of claim 1 , further using prior knowledge of characteristics of a determined signal component to improve the estimation of the jitter contribution to subtract.

11. The method of claim 1 , further including a step of introducing a known calibration signal into the sampled signal.

12. A method as claimed in claim 1 , wherein the processor is either a signal processor or signal digitizer.

13. A computer program product embodied on a non-transitory computer readable medium, loadable in a memory of a digital processor, containing executable code arranged to carry out a method the when executed by the digital processor, the method comprising,

obtaining a frequency-domain data set representing a sampled signal;

obtaining an average sideband amplitude distribution generated by jitter around one or more principal frequencies of the signal;

estimating jitter phases for sidebands generated by jitter in the frequency-domain data set;

subtracting a jitter contribution in the sampled signal characterized by the average sideband amplitude distribution and the estimated jitter phases from a data record.