Horizontal lines - TBC, tape, or hardware problem?

[Name]

Hello,

today during capturing a tape I came across some horizontal line artifacts.


[Attachment 85020 - Click to enlarge]

I'm using a JVC S9500, and TBC was enabled. The VHS is "Babe" from 1996.

The artifacts just last for around 1 frame and are occurring on different positions.

Furthermore, it looks these artifacts are becoming more frequent, the further I'm into the tape. At the beginning they are seldom and at the end quite frequent.

As this didn't happen with my previous test tapes, I'm wondering now what the reason is?

Cheers

[aramkolt]

If you have the TBC off, does it still do it? Also, if you recapture that same section, does that artifact show up on the same frame it did with the prior capture?

[themaster1]

Can you upload a video sample ?

[dellsam34]

It's called tape dropouts, There is nothing you can about it.

[Sharc]

Hello,

today during capturing a tape I came across some horizontal line artifacts.


[Attachment 85020 - Click to enlarge]

I'm using a JVC S9500, and TBC was enabled. The VHS is "Babe" from 1996.

The artifacts just last for around 1 frame and are occurring on different positions.

Furthermore, it looks these artifacts are becoming more frequent, the further I'm into the tape. At the beginning they are seldom and at the end quite frequent.

As this didn't happen with my previous test tapes, I'm wondering now what the reason is?

Cheers

See here. It is probably the tape.
https://www.avartifactatlas.com/artifacts/video_dropout.html

[jagabo]

The artifacts just last for around 1 frame and are occurring on different positions.

That can often be fixed after capturing. Using AviSynth.

[Sharc]

The artifacts just last for around 1 frame and are occurring on different positions.

That can often be fixed after capturing. Using AviSynth.

Yes, using spotless() for example, with careful settings not to destroy details or small moving objects.

[Name]

That's a 15 sec sample at 1h 15m, where it is already quite frequent:

https://files.videohelp.com/u/310756/sample.avi


So this means, most likely, it is the age of the tape!?

[Sharc]

Not perfect and maybe too distructive:

AVISource("sample.avi")
assumeTFF()
converttoYV24(interlaced=true) #convert to planar
separatefields()
e=selecteven()
o=selectodd()
e=e.spotless(RadT=2,ThSAD=2000) 
o=o.spotless(RadT=2,ThSAD=2000) 
interleave(e,o)
weave()

[lordsmurf]

It's called tape dropouts, There is nothing you can about it.

This. DOC compensation.

[Name]

AVISource("sample.avi")
assumeTFF()
converttoYV24(interlaced=true) #convert to planar
separatefields()
e=selecteven()
o=selectodd()
e=e.spotless(RadT=2,ThSAD=2000) 
o=o.spotless(RadT=2,ThSAD=2000) 
interleave(e,o)
weave()

Can you explain it a bit what this is doing?
Not just that the drop-outs are better, also the image itself looks smoother.
My sample was a bit wavy / shimmery (not sure how to call it).

I just hope for the tapes where it matters, I don't face this issue.

Some crazy question, as it is just happening for 1 frame at a time, and the issue is always a straight line, a shift.
Would there be a tool where I could specify the frame, line, and how much it would need to be shifted?
I know this would be a hell of a task, but just out of curiosity.

[jagabo]

The dropouts are the least of the problems. First, there's a lot of horizontal time base wiggle. A line TBC like those on some Panasonic DVD recorders may help with that -- disable the VCR's TBC. Then there are vertical time base problems. The fields bounce up and down randomly. A better capture device might help with that. The Panasonic DVD recorder may help with that. Or a full frame TBC. And the video is very noisy. All this together make me think it was recorded at a very low tape speed. SLP?

Note that the underlying video is progressive and could normally be field matched with TFM() to give 25p. But the vertical bounce is interfering with that.

[Sharc]

AVISource("sample.avi")
assumeTFF()
converttoYV24(interlaced=true) #convert to planar
separatefields()
e=selecteven()
o=selectodd()
e=e.spotless(RadT=2,ThSAD=2000) 
o=o.spotless(RadT=2,ThSAD=2000) 
interleave(e,o)
weave()

Can you explain it a bit what this is doing?

It applies the avisynth script 'Spotless' to the even and odd fields of the video, then re-weaves the filtered fields. The separation into fields was a workaround to TFM() which didn't work as intended because of the wobbly and bouncing video. My script was a quick sledgehammer fix of your capture which suffers from the defects jagabo mentioned.

[Name]

All this together make me think it was recorded at a very low tape speed. SLP?

It is an original official tape. I bought it. So that's not a home copy.

Well, I'm just testing and playing around with some random tapes I still have at home before I lay hands on my family recordings.

Furthermore, my JVC S9500 + I-O Data GV-USB2 is just the first step. In the long shot, I'm planning to modify my VCR and try the VHSdecode project.

[dellsam34]

If that's a commercial tape then there is something seriously wrong with your VCR or capture workflow besides the drop outs, Too much noisy pattern in that sample, possibly ground loop issues or faulty VCR.

[Sharc]

As this didn't happen with my previous test tapes, I'm wondering now what the reason is?

Re-capture a tape which was good in your previous tests. If it is still good it is an indication of a bad tape, if it is worse than before it points to a VCR or setup issue. Consider careful heads cleaning.

[Sharc]

The glitches seem to be in the odd fields only. So here a variant based on the even fields. (Doesn't fix the wobbling and bouncing though):

AVISource("sample.avi")
converttoYV16(interlaced=true)
assumetff()
QTGMC(preset="fast").selecteven()
#BWdif(field=1) #alternative to QTGMC
MCDegrainSharp()

[lollo]

I also think the problem is in the tape. Its condition determines the drop outs (the bad horizontal lines), and related to that is the vertical jitter (the even or odd or both fields shifed up or down. The TBC of the JVC is not able to reconstruct all the lines in the frame, so the shift. In my experience this cannot be fixed by an external frame TBC, while a specific DVD-Recorder in pass-through mode may sligthly attenuate it.

A specific DVD-Recorder can also help with the "horizontal jitter", being its line correction stronger than the TBC inside the VCR.

The excessive noise is also a synthom of a bad tape.

As side note, for few shifted fields (i.e. 10/50 in a capture, otherwise is a long and boring work) in AviSynth is possible to fix them without impact on the final "visual" quality. More info here: https://www.digitalfaq.com/forum/video-restore/12929-how-replace-frames.html#post86442

[Name]

Alright, thanks to all of you for the input and the corrected test sample.
I re-capture some of the former tapes and also see how other tapes behave.

This sounds very interesting (and time-consuming)
https://www.digitalfaq.com/forum/video-restore/12929-how-replace-frames.html#post86442

[lollo]

This sounds very interesting (and time-consuming)
https://www.digitalfaq.com/forum/video-restore/12929-how-replace-frames.html#post86442

Not feasable in your sample. Even because after the shift the fields stay at that offset and do not return back to the proper line position, so you have to correct also the "good" frames.

It applies only if you have some field shift, as it very often happens. An example here: https://forum.videohelp.com/threads/409177-VHS-Horizontal-Stabilisation-part-2-Electri...oo#post2686220

[dellsam34]

I would worry more about playing the tape correctly first before moving to the restoration phase, I know it is common in public forums that people jump right into band-aid patching solutions, but if you can figure out the source of noise and eliminate it you may not have to do all this that should not be necessary in the first place.
Lollo when tapes wear out they don't developpe a noise pattern like that one, they get drop outs, they miss-track, they get occasional horizontal RF noise and streaks, but not in that noise pattern, Also those line drop outs could be just a side effect of bad playback, I'm not ruling out the possibility of bad tape with bad playback.

[lollo]

Sure, I agree, restoration only after anything else is fixed.

My point was that probably the tape is not that good. The OP has provided other samples with the same hardware (but different tapes) not showing that noise. It may be marginal, I don't know without having the material in my hands.

[Name]

I don't know without having the material in my hands.

Honestly, if you want, I can even send you the tape and you can test it.
This tape has no value at all to me. So nothing can go wrong.
Not sure where you are located, but if you don't feel comfortable sending it to your address, I could even send it to a locker.
Up to you, or to anyone else here. I'm located in Central Europe.

[lollo]

OK, I will PM you my address, so I can test what's wrong with this tape

[jagabo]

I don't think damage to the tape is likely since only one field is effected. It was either recorded with a misaligned/dirty head or the playback device has a misaligned/dirty head.

[dellsam34]

That's what I thought until he mentioned it is a pre-recorded tape, I'm not entirely sure how a commercially recorded tape develops such a noise pattern. Even a misaligned VCR with dirty heads cannot produce that noise pattern, it can cause the artifact shown in the first post, but not the noise.
One way to find out if the noise is in the original recording, you guessed it. Hit pause of the tape playback, if the noise freezes, it is in the tape, if it doesn't it is in the playback chain.