Capturing on a hard drive with long stored files a bad idea?

[lollo]

For some reason I can still see filed blending in your sample, someone can correct me if I'm wrong.

I do not see field contamination in Recapture.avi. But each field is very low quality and their timing architecture is weird, as Sharc alreay pointed out.

Is not a simple (constant) phase shift because a DoubleWeave().SelectOdd() does not fix the video. It must be some dynamic phase shift, embedded in the tape or introduced by a faulty device.

It can be fixed with a field matching operation, here a comparison comp.avi. But OP should solve his capture issue first.

video_org=AviSource("Recapture.avi")

#video_org_rep=video_org.DoubleWeave().SelectOdd()

video_org_rep=video_org.TFM(order=1, mode=0, PP=0, slow=2, field=1)

stackhorizontal(\
subtitle(video_org,"video_org",size=20,align=2),\
subtitle(video_org_rep,"video_org_rep",size=20,align=2)\
)

[jagabo]

Yes, 25p content was broadcast BFF but captured TFF. It can also be made progressive again with something like:

AssumeTFF()
SeparateFields()
Trim(1,0)
Weave()

Though this has the same problem as DoubleWeave().SelectOdd(). The filter chain gets confused because there are an odd number of fields left after removing the first field. If you work linearly from start to end there's no problem. But if you go to the end and step backward the filtering gets messed up.

On the other hand, TFM() can sometimes mismatch or deinterlace inappropriately.

[lollo]

On the other hand, TFM() can sometimes mismatch or deinterlace inappropriately.

With PP=0 TFM does not look for combed frames, so no deintarlacing is applied. And I add mode=0 for this type of material.

The only case where I noticed a such configured TFM in confusion is in rare footage where the picture is static and there is some vetical/horizontal shifting (i.e. end credits on uniform background); but the outcomes is just a frame pattern change, so nothing serious.

But again, if the defect is not baked on tape, better to fix the capture.

[Gary34]

The one i saw was a European model apparently. Slightly different layout on the front as it doesn't have the Tint buttons. One of the later models anyway

You saw one that had 2004 on it that was PAL? I’m pretty sure you just saw a board image of mine. I haven’t heard of anyone else having one that was made that late.
[Attachment 89878 - Click to enlarge]

[jagabo]

On the other hand, TFM() can sometimes mismatch or deinterlace inappropriately.

With PP=0 TFM does not look for combed frames, so no deintarlacing is applied. And I add mode=0 for this type of material.

Sorry, I didn't notice you used pp=0.

[Gary34]

Well I guess we know that a recap doesn’t fix that now. https://www.digitalfaq.com/forum/vcr-repair/14060-tbc-3000-causing.html LS has the same issue with his 3rd and 4th gen TBC 3000s and from what he said there is not a known way to fix it.