Multi-directional comb filtering in a digital video decoder

1. A method of filtering, using a multi-directional comb filter, pixel data from a sequence of fields of pixel data, wherein the fields of pixel data comprise a previous field, a current field and a next field, and wherein each of said fields includes a previous line, a current line, a next line, the method comprising:

determining a temporal chrominance velocity (K — 3dcomb) relative to a pixel position of the current field based on

a first video signal value (cvbs_p) for the pixel position on the current field;

a second video signal value (cvbs_c) for the pixel position on the previous field; and

a third video signal value (cvbs_n) for the pixel position on the next field,

wherein

K — 3dcomb is determined as one or more of

C — 3dcomb — n /(C — 3dcomb — p +C — 3dcomb — n ), and  (1)

C — 3dcomb_ p /(C — 3dcomb_ p +C — 3dcomb_ n ), wherein  (2)

C — 3 d comb p =(cvbs — c −cvbs — p ), and

C — 3 d comb n =(cvbs — c −cvbs — n ); and

determining a temporal luminance based on

the first video signal value;

the second video signal value;

the third video signal value; and

the temporal chrominance velocity.

2. The method of claim 1 , further comprising scaling the temporal luminance based on the temporal chrominance velocity.

3. The method of claim 1 , wherein the temporal luminance (Y3D_comb) is: K — 3dcomb*Y 3dcomb_p+(1−K — 3dcomb)*Y — 3Ddcomb_n;

wherein:

Y — 3 d comb — p =(cvbs — c +cvbs — p )/2; and

Y — 3 d comb — n =(cvbs — c +cvbs — n )/2.

4. The method of filtering of claim 1 , wherein the determining of the temporal luminance comprises scaling a summation of the first video signal value and second video signal value, and a summation of the second video signal value and the third video signal value.

5. The method of filtering of claim 1 , wherein the determining of the temporal luminance comprises determining an overall temporal luminance by selecting from among a spatial luminance, a temporal luminance, and a mix of the spatial luminance and the temporal luminance.

6. A multi-directional comb filter for filtering pixel data from a sequence of fields of pixel data, wherein the fields of pixel data comprise a previous field, a current field and a next field, and wherein each of said fields includes a previous line, a current line, a next line, the multi-directional comb filter comprising:

a notch filter to generate a notched luminance value;

a comb filter to generate a spatial luminance value and a temporal luminance value;

a spatial chrominance filter to generate a first spatial chrominance value and a second spatial chrominance value;

a temporal chrominance filter to generate a first temporal chrominance value and a second temporal chrominance value; and

a final mixer to combine the notched luminance value, the spatial luminance value and the temporal luminance value to generate a mixed luminance value and to combine a first raw chrominance value, the first spatial chrominance value and the first temporal chrominance value to generate a first mixed chrominance value, and to combine a second raw chrominance value, the second spatial chrominance value and the second temporal chrominance value to generate a second mixed chrominance value.

7. The multi-directional comb filter of claim 6 , wherein the final mixer comprises three mixers.

8. The multi-directional comb filter of claim 7 , wherein the final mixer further comprises a stillness factor.

9. A multi-directional comb filter for filtering pixel data from a sequence of fields of pixel data, wherein the fields of pixel data comprise a previous field, a current field and a next field, and wherein each of said fields includes a previous line, a current line, a next line, the multi-directional comb filter comprising:

means for determining a temporal chrominance velocity (K — 3dcomb) relative to a pixel position of the current field as one or more of

C — 3dcomb_n /(C — 3dcomb_p+C — 3dcomb_n), and  (1)

C — 3dcomb_p /(C — 3dcomb_p+C — 3dcomb_n), wherein  (2)

C — 3 d comb — p =(cvbs — c −cvbs — p ),

C — 3 d comb — n =(cvbs — c −cvbs — n );

cvbs_p is a first video signal value for the pixel position on the current field;

cvbs_c is a second video signal value for the pixel position on the previous field; and

cvbs_n is a third video signal value for the pixel position on the next field; and

means for determining a temporal luminance based on the first video signal value, the second video signal value, the third video signal value, and the temporal chrominance velocity.

10. The multi-directional comb filter of claim 9 , wherein the means for determining the temporal luminance comprises a means for scaling a summation of the first video signal value and second video signal value, and a summation of the second video signal value and the third video signal value.

11. The multi-directional comb filter of claim 9 , wherein the temporal luminance (Y3d_comb) is:

K — 3dcomb*Y — 3dcomb_p+(1−K — 3dcomb)*Y3d_comb_n;

wherein:

Y — 3 d comb — p =(cvbs — c +cvbs — p )/2; and

Y — 3 d comb — n =(cvbs — c +cvbs — n )/2.

12. The multi-directional comb filter of claim 9 , wherein the means for determining the temporal luminance comprises a means for scaling a summation of the first video signal value and second video signal value, and a summation of the second video signal value and the third video signal value.

13. The multi-directional comb filter of claim 9 , wherein the means for determining the temporal luminance comprises a means for selecting from among a spatial luminance, a temporal luminance, and a mix of the spatial luminance and the temporal luminance.