Hilbert Transforms Hilbert transforms are used to change the phase of a signal by 90°. For example H cos 0 t sin 0 t . What is the process by which this operation is performed? Let us consider this operation in frequency domain. We start with the Fourier transforms of cosine and sine: d(+0) F {cos 0t} d(-0) j d(+0) F {sin 0t} -j d(-0) What frequency operation is performed to go from sine to cosine? A useful function in formulating this operation is the signum or sign function: 1 ( 0 ), sgn( ) 1 ( 0 ). This function is like a unit step function with a nonzero negative part. sgn () We readily see that, in frequency domain, the Hilbert transform is equivalent to multiplication by j sgn( ) as can be seen on the following slide. F {cos 0t} -j sgn () F {sin 0t} j -j So, a Hilbert transform can be reduced to a filtering operation, where H ( ) j sgn( ). We can design a digital filter with this frequency response: H d ( ) j sgn( ). The coefficients cn for the digital filter can be found using the Fourier series method: cn 1 2 1 2 H d ( ) e jn j sgn( ) e d jn j 0 jn (1) e d 2 d 0 ( 1) e jn d j 1 1 jn jn 1 e e 1 2 jn jn 1 2 n 1 2 n 1 n 2 e ) jn 2 2 cos 1 cos n e n ) jn ) ) The plot of the magnitude and phase of this filter is shown on the following slide. Frequency Response 2 j |H(e )| 1.5 r = 20 r = 50 1 r=5 0.5 0 -0.5 -1.5 -1 -0.5 -1 -0.5 0 0.5 1 1.5 0 0.5 1 1.5 , x j H(e ), x 1 0.5 0 -0.5 -1 -1.5 , x