Jumpy / jittery VHS capture

[VideoYak]

Hi all,

I have been having some success doing my Video8 transfers (Hi8 (S-video) - TBC - Capture Card). A family member gave me a panasonic VCR with S-video out and a VHS tape that they asked if I could digitize.

I gave it a shot, I understand the VHS might not be ideal - there is this jitter/jumpiness to footage. I have attached a samples of the footage (interlaced) and then the footage post QTGMC.

I tried StabMod but it didnt do too much. Is there anything I can do to improve it? Or is it basically un-fixable unless I use a better VHS player for capture?

# script by g-force, avsi by McCauley, modified by lordsmurf
# Note: stab() was proving too erratic for most VHS sources, hence this mod

function StabMod (clip clp, int "range", int "dxmax", int "dymax")
{
    range = default(range, 4) #2
    dxmax = default(dxmax, 4) #2
    dymax = default(dymax, 8) #10
    temp  = clp.TemporalSoften(7,255,255,25,2)
    inter = Interleave(temp.Repair(clp.TemporalSoften(1,255,255,25,2)),clp)
    mdata = DePanEstimate(inter,range=range,trust=0,dxmax=dxmax,dymax=dymax)
    DePan(inter,data=mdata,offset=-1)
    SelectEvery(2,0)
}

[dellsam34]

The frame jump is film source related, It has nothing to do with VHS, Normally the source material is not interlaced so it should be de-interlaced in a specific way, However with that frame jump now it became interlaced, Hope one of the members like Jagabo will chime in and give you the right scripts.

[davexnet]

Just out of interest, if you hook up the VCR to the TV and watch the tape there, does it still bounce up and down?
If so then narrow it down with a different tape and/or different VCR

[jagabo]

I looks to me like the main problem is that a rolling shutter device was used to digitize the original film, and the film was sometimes still moving vertically while frames were digitized. This has caused the content of some frames to be taller than others. If you look at individual fields you'll see that some are stretched vertically relative to the previous and next. That cannot be fixed by a de-bounce filter.

[VideoYak]

I looks to me like the main problem is that a rolling shutter device was used to digitize the original film, and the film was sometimes still moving vertically while frames were digitized. This has caused the content of some frames to be taller than others. If you look at individual fields you'll see that some are stretched vertically relative to the previous and next. That cannot be fixed by a de-bounce filter.

Thanks jagabo, which component in my workflow would be the rolling shutter device? The TBC, VHS player or the capture card?

I used the s-video out on the VHS/DVD player. Would using an alternate component out help instead?
[Attachment 74728 - Click to enlarge]

[dellsam34]

The original film, Before the VHS cassette was made.

[poisondeathray]

You can check on a longer sequence, but every 3rd field looks almost ok in terms of vertical movement and warping - like it belonged to the original film sequence. You might have to adjust the pattern depending on where you cut the file

Double rate deinterlacing then selecting every 3rd frame gives you 16.667fps . You can leave it or use RIFE or some other interpolator like FRC, mvtools2 etc... to interpolate frames

(You can look at deflickering, descratching, dirt removal too, as well as cleaning but the borders)

1.mp4

AVISource("Interlaced Sample.avi")
AssumeTFF().QTGMC(border=true)
SelectEvery(3,2)

2.mp4 adds spotless and rife

[Alwyn]

PDR, marvellous.