| 摘要 |
The present invention is a process for reducing the noise of a video signal (S1), in particular when reproducing a video signal stored on a video recorder, in which a window is defined around every signal value P (0,0,0) as an output signal value. A mean deviation ( sigma ) of the signal values in the window is established from the output signal values P (0,0,0), and an averaging of the output signal value is carried out with those neighboring signal values in the window which lie within limit values (So, Su) for the deviation from the output signal value P (0,0,0) dependent on the established mean deviation ( sigma ). The method is improved in its effectiveness by at least one threshold value (a1, a2) being fixed for the mean deviation ( sigma =a/2), by the limit values (So, Su) for a mean deviation lying below the threshold value (a1, a2) being set with values (b) dependent on the established mean deviation ( sigma ) and by the limit values (So, Su) for a mean deviation ( sigma ) lying above the at least one threshold value (a1, a2 ) being determined with a substantially constant, preselected maximum value (b2).
|
