odd  line 4*n+1: [ Yo1, Co1 ] (Y: luminance, brightness)
even line 4*n+2: [ Ye1, Co1 ] (C: chrominace, color)
odd  line 4*n+3: [ Yo2, Ce1 ]
even line 4*n+4: [ Ye2, Ce1 ]

Odd and even lines interlace with two different fields
but share the same chrominace value... obviously wrong.

Note that I never said the interlaced display is wrong.
I mean the color decoding of interlaced frames is imperfect.

I favor Japanese animation, so-called Anime.
Different from RC1 movies, they are almost coded with interlaced frames.

It's impossible to convert such kind of 30/29.790fps interlaced frames
back to 24/23.976fps progressive frames (Inverse Telecine) as usual.
Moreover, the quality of adaptive deinterlace downgrades also.

That's why I develop DVD2AVI.

odd  line 4*n+1: [ Yo1, Co1 ]
even line 4*n+2: [ Ye1, Ce1 ]
odd  line 4*n+3: [ Yo2, Co1 ]
even line 4*n+4: [ Ye2, Ce1 ]

This is the correct decoding process of interlaced frames.
In order to get better quality, DVD2AVI treats interlaced frames as

odd  line 8*n+1: [ Yo1, (  Co0 + 7*Co1)/8 ]
even line 8*n+2: [ Ye1, (3*Ce0 + 5*Ce1)/8 ]
odd  line 8*n+3: [ Yo2, (5*Co1 + 3*Co2)/8 ]
even line 8*n+4: [ Ye2, (7*Ce1 +   Ce2)/8 ]
odd  line 8*n+5: [ Yo3, (  Co1 + 7*Co2)/8 ]
even line 8*n+6: [ Ye3, (3*Ce1 + 5*Ce2)/8 ]
odd  line 8*n+7: [ Yo4, (5*Co2 + 3*Co3)/8 ]
even line 8*n+8: [ Ye4, (7*Ce2 +   Ce3)/8 ]

The result approximates to the 6-tap FIR filter of MSSG codes.