The Most Stubborn Tape so Far

[dellsam34]

One of the members contacted me with a need to capture some tapes that he was unable to get a stable capture off of with his full workflow he bought, This full size VHS NTSC tape appear to be a second gen dub from a small cassette, probably from another format too, or at least shot with a fluorescent tube camera judged by the red/green tones at objects' edges. The biggest problem though is this nasty flagging on top of the frame in addition to occasional miss-tracking and loss of signal, As shown in the customer's capture with frequent blue screens across the entire footage, This is his sample:

Customer's Capture

On my end I tried a normal Sony VHS VCR, a JVC D-VHS VCR but none of the two was able to produce less flagging, Albeit no blue screen or miss-tracking, Then I played it with one of my JVC HR-S7600AM but the flagging is the same, I turned the line TBC off which dramatically reduced flagging but has some on the top still, in addition to more chroma noise since the DNR is tied to the TBC, See samples with and without LTBC:

JVC Line TBC ON Less chroma noise.
JVC Line TBC OFF More chroma noise.
VCR (S-Video) -> S&W TBS-800 (SDI) -> BM UltraStudio (USB3)

This tape has narrow tracks with an off the norm angle, I was barely able to scan the tracks properly with the JVC DD system with the tape moving, (I wish the DD system feature could be modded to work in playback mode, it will solve a lot of tape problems), Sample attached:

JVC DD System @ 2X

With tape in pause the frame is garbage, this indicates how short the tracks are, Here is a step by step capture:

JVC DD System in step by step

So at this point I needed something as passthrough to see if I can reduce the remaining flagging even further, For this I used the Sony HVR-M15AU HDV deck, (S-Video in - S-Video out with digital processing, No DV out), I was able to mitigate the problem to an acceptable visual level, as you can see here:

Sony HVR-M15AU Pass-through Chroma noise set to low in the M15AU menu.
VCR LTBC OFF (S-Video) -> HVR-M15AU (S-Video) -> S&W TBS-800 (SDI) -> BM UltraStudio (USB3)

So what do you guys think? Can this occasional horizontal jitter on top of the frame be improved further? like post restoration in software? My job was just to do a capture though, so any software suggestions will be forwarded to him.

[Alwyn]

VCR (S-Video) -> S&W TBS-800 (SDI) -> BM UltraStudio (USB3)

This is the one I wanted to see... No link though.

Also, some info on the customer's capture would be helpful when comparing: what VCR, what digitiser, DVD recorder used for stabilising...

[dellsam34]

The links are right on top of it, The customer is using a JVC VCR with TBC ON and probably an external frame TBC, I didn't ask for details since I have to fix this problem on my end, not on his end.

[Alwyn]

Fair enough, that wasn't clear.

QTGMC won't open that M15AU file because the YV12 conversion fails:


[Attachment 78209 - Click to enlarge]

486...

My script:

SetFilterMTMode ("QTGMC", 2)
    AVISource("H:\Videohelp\Dellsam\Sony HVR-M15AU Passthrough.avi")
    ConvertToYV12(interlaced=true)
    AssumeBFF()
    QTGMC(preset="Fast", EdiThreads=8)
    Prefetch(24)

It's also BFF which is unusual for an analogue capture.

[Alwyn]

The customer is using a JVC VCR with TBC ON and probably an external frame TBC, I didn't ask for details since I have to fix this problem on my end, not on his end.

We're here to learn, Dellsam. If they used cr@p gear, then that's a good lesson for us. If they're using good gear, that's also a good lesson.

[Alwyn]

I'd love to see an ES-10/15 on that tape.

[dellsam34]

I believe NTSC is BFF, I could be wrong. 486 is broadcast standard capture, If you can't process it, crop it down to 480.

[lordsmurf]

If you want to pass me the tape, with permission of course, I have a hardware arsenal to try.

[Alwyn]

...

[dellsam34]

The horizontal jitter is still there, that was my concern, De-interlacing and color correction are personal preferences.

[dellsam34]

If you want to pass me the tape, with permission of course, I have a hardware arsenal to try.

That's not my say, It's part of a lot of 4 tapes, I'm only allowed to send them back to him with a pre paid postage.

[Alwyn]

Dellsam, I want to understand the two JVC "TBC On" and "TBC Off" captures. Were both were done with the Snell TBC in the chain?

[dellsam34]

The two links are followed by the workflow bellow them, that means it applies to both, The same with passthrough, the workflow is posted below the link. I used line spacing, It should be obvious.

[Sharc]

Try Dejitter (Avisynth32 plugin). It seems to give some improvement.
http://www.avisynth.nl/users/vcmohan/DeJitter/DeJitter.htm

AVISource("Sony HVR-M15AU Passthrough.avi")
crop(0,0,-0,-6)
DeJitter(jmax=80,th=0.1, wsyn=8, extend=true)
crop(0,4,-32,0).addborders(8,2,8,2) #centering, 704x480

[Alwyn]

@Sharc, I don't think that improves things overall; while the top is steadier, the detail of the vertical surfaces is compromised. Check the legs and edge of the dresses around frame 1200.


[Attachment 78222 - Click to enlarge]

[Sharc]

@Alwyn: Yes, I noticed this as well. I didn't try to optimize anything. It was just a quick shot using this filter. It may work better when applied to the deinterlaced video, I don't know.
Initially I thought to just crop/mask the top ~32 pixels as there is not much significant picture content there.
Maybe there are better proposals around.

[sophisticles]

I am going to ask a question that will probably piss some off:

Has the owner of these tapes considered sending them to experts that have specialized equipment for digitizing these types of tapes?

I am not implying that dellsam34 doesn't know what he is doing but the equipment he has mentioned using is commodity hardware.

Wouldn't a service like Kodak's digitizing have more specialized equipment:

https://kodakdigitizing.com/

Maybe I am being naive, but if Kodak can't digitize this properly, no one can.

[lollo]

I am not implying that dellsam34 doesn't know what he is doing but the equipment he has mentioned using is commodity hardware.
...
Wouldn't a service like Kodak's digitizing have more specialized equipment

And what hardware do you think this "Kodak" is using? Probably worse than many of the members here.

Maybe I am being naive, but if Kodak can't digitize this properly, no one can.

The best candidates here for finding a solutions are Bogilein and maybe LordSmurf, having a large set of different VCRs.

The online service are generally crap.

If you refer to having "professional" VCRs for digitizing impossible tapes, it does not work like that.

[Bogilein]

As this is NTSC material I wouldn't be much help as I don't have the equipment for NTSC material.

If it was PAL material I would try a Panasonic DMR first, then Canopus NX and various VCRs (Panasonic, JVC pre-1998) and play around with manual tracking to see if it gets better.

[lollo]

As this is NTSC material I wouldn't be much help as I don't have the equipment for NTSC material.

Sure, mine was just a general comment about the difference between some of the members here and the proposed online services

If it was PAL material I would try a Panasonic DMR first, then Canopus NX and various VCRs (Panasonic, JVC pre-1998) and play around with manual tracking to see if it gets better.

Thanks for the hint, I agree

[dellsam34]

Maybe I am being naive, but if Kodak can't digitize this properly, no one can.

They may have a fancy website and multi-billion camera business but I can guarantee you this tape will be rejected by Kodak's service and the likes, such as Legacy Box, I honestly don't know what Kodak uses to transfer tapes but I can assure you not the S-VHS microwave size decks for a very simple reason, No technicians available to service those decks regularly, and keeping an old technician on the payroll will not be a wise corporate decision since it would cost too much, My guess is they use later models Funai VCRs and combo DVD/VCR's, So if one breaks down they are easy to replace by the personnel working there, yellow to yellow composite, red to red, white to white for audio and plug in power and they're good to go.

So yeah I'm one of the very few options available to him, And also those microwave size S-VHS machines are not recommended for consumer tapes anyway, at least the faulty ones like this tape, they are built for studio quality recordings that work with no problem. So it is not that I was unable to capture this problematic tape, I did so successfully given how bad it is, So my job is done, but being a perfectionist, there is just a sliver of occasional shaking on top of the frame that I want to see if the customer can eliminate further. I let him know about this thread, so this is for him, not for me.

[Bwaak]

Maybe I am being naive, but if Kodak can't digitize this properly, no one can.

They may have a fancy website and multi-billion camera business

They don't.

[dellsam34]

So they are just like Legacy Box then, Legacy Box also likes to brag about the TV shows and other organizations rating, It's like those TV shows hosts and their audience are experts in capturing old media.

[Alwyn]

Pity we couldn't see an ES-10 capture of that. Would have been good comparing it with the higher-end setup.

I can guarantee you this tape will be rejected by Kodak's service and the likes, such as Legacy Box

The problem with that being?

[sophisticles]

And what hardware do you think this "Kodak" is using? Probably worse than many of the members here.

If you refer to having "professional" VCRs for digitizing impossible tapes, it does not work like that.

I honestly don't know what Kodak uses to transfer tapes but I can assure you not the S-VHS microwave size decks for a very simple reason

I was thinking that they may be able to use different approaches than what has been put forth in this thread.

The two things i envisioned were taking the tape out of the VHS and scanning it the way film is scanned, you guys are thinking with a traditional VCR mindset, I was thinking about treating like a reel of film, which it basically is, and using film scanning equipment.

For a more extreme approach, you can treat the VHS tape as a sort of specialized LTO tape backup and use data recovery techniques to extract the video data.

These are just theories of mine, since if you think about it a VHS tape is really a specialized LTO tape that only stores video and audio data, but i see no reason why you can't treat it like any other tape archive and use similar equipment to extract the raw data, which in this case is the audio/video information.

Long story short, I think the approach being advocated by the experienced members of this forum, may be a bit too linear.

Time to think a bit out of the box.

[Cornucopia]

No, it's not like LTO, and high-speed transfer wouldn't work here.

1. LTO is digital, so the sensitivity of 1 vs 0 voltages is a clear demarcation of difference compared to analog, and LTO has a form of error detection and correction that takes advantage of the tape transport modality as well as the speed.
2. Although LTO is very fast, it is still "realtime" as far as its intended transfer bandwidth is concerned.
3. VHS high speed duping is done NOT by playing on a deck at high speed, but rather it uses magnetism's nearby field to "contact print" a blank tape that is placed side by side, running at the same speed, simple spool to spool. But this only works for dubbing, and it does still produce copies that are inferior to standard methods.
4. Capturing/digitizing would be different, requiring ultra-high bandwidth electronics, and very specialized consistent high speed motors and playback head gear. In fact, it would require everything that existing vhs decks require, but at several orders of magnitude tighter tolerance, higher caliber. This makes things like timebase correction & stability that much more critical. All of these add up to very expensive ($15-20k+) gear with almost zero niche market.
Physics & good Engineering still have to be adheared to.


Scott

[dellsam34]

The problem with that being?

For the obvious reasons I stated.

[dellsam34]

The two things i envisioned were taking the tape out of the VHS and scanning it the way film is scanned, you guys are thinking with a traditional VCR mindset, I was thinking about treating like a reel of film, which it basically is, and using film scanning equipment.

When you design and make a machine like this let us know. Kodak does not have this magic machine, I can guarantee it, it's not worth high dollar investment in R&D.

[dellsam34]

Capturing/digitizing would be different, requiring ultra-high bandwidth electronics, and very specialized consistent high speed motors and playback head gear. In fact, it would require everything that existing vhs decks require, but at several orders of magnitude tighter tolerance, higher caliber. This makes things like timebase correction & stability that much more critical. All of these add up to very expensive ($15-20k+) gear with almost zero niche market.
Physics & good Engineering still have to be adheared to.


Scott

Dynamic tracking should have been the norm for all VCRs, Unfortunately only one or two models ever made with this technology probably because it was too expensive back then, It's like the DD system for trickle play but it tilts only the heads instead of the entire drum, and unlike the DD system it works for normal playback, This would give a degree or two for offset video tracks like the tape being discussed here.

[lollo]

Long story short, I think the approach being advocated by the experienced members of this forum, may be a bit too linear.

Time to think a bit out of the box.

We are too linear? Time to think a bit out of the box? Useless comments.

You do not have know what you are talking about: "treating like a reel of film, which it basically is, and using film scanning equipment"? "you can treat the VHS tape as a sort of specialized LTO tape backup"?

I do not know if you are more incompetent on VHS capture or on encoding, as master poisondeathray often points out (https://forum.videohelp.com/threads/413919-Is-Intel-s-QSV-encoder-better-than-X264-and...65#post2729340 just as example)

So avoid that kind of criticism to us.

BTW, have a look to VHS Decode project for an alternative to our standard methods. Not really a breakthrough, but a different approach.