Is this ghosting from a 1984 VHS tape ?

[hydra3333]

Hello.

Attached is a 17Mb example source (from VHS recorded in 1984, no other capture available).

It has what looks like may be

(a) a small colour shift right (VHS tape copy?) eg on the white shirts at the 1 second mark, although that looks "wrong" on the flagpoles at the 17 second mark
(b) and something which may be a heavy "Ghost" but could be something else I don't know the name of, given what appears to be an associated heavy colour shift to the right

Attached is 13Mb beginning of to address it, deblock/cropbottom/cnr2/mdegrain3/chromashiftSP(x=2)/qtgmc.

Your expert advice would be greatly appreciated about what the large ghosting-type artifact is, and perhaps suggestions on how to address it.

Thank you.

[jagabo]

Better in some areas, worse in others.

src = last
Overlay(last, last.Invert(), x=11,  y=0, Opacity=0.20).ColorYUV(cont_y=165, cont_u=165, cont_v=165)
Overlay(last, src, mask=ColorYUV(cont_y=150).Invert(), opacity=0.7)

Chroma may need slightly different treatment.

[Sharc]

.... what the large ghosting-type artifact is ....

Could be signal reflections due to long cables with wrong impedance termination between the VCR and the capture device.

Edit: Unlikely. A second thought indicates that for practical cable lengths the echo ghost is less than 1...2 pixel delayed.

[jagabo]

OTA multipath reception.

[hydra3333]

Thank you. It is ghosting, then. Wasn't sure what "Halo" meant (if it was different) and if it was that or something.

Better in some areas, worse in others.

src = last
Overlay(last, last.Invert(), x=11,  y=0, Opacity=0.20).ColorYUV(cont_y=165, cont_u=165, cont_v=165)
Overlay(last, src, mask=ColorYUV(cont_y=150).Invert(), opacity=0.7)

Chroma may need slightly different treatment.

Good oh, I'll look these up in the Wiki and try them as a starting point.