Processing captured VHS from a bad recorder/tape

[GrofLuigi]

Hi all,

It's been a while since I've captured VHS and today I discovered two of my three recorders had died. I was left with a Panasonic NV-G14, one of the earliest models, with a transparent window above the tape mechanism.

This is what it produced:

Sample (mjpeg (direct stream copy), 130.985.000 bytes).

And after the capture it died too.

Obviously, the best solution would be to buy/rent/borrow a good recorder and recapture, but it might take a while to find sensible option, and I don't want to spend too much on a new unit that I might never use again. Used is out of the question as the ones I find are usually just as bad.

The tape is also in bad shape, eaten in many places.

The goal is to make a DVD. What processing options would you recommend? I think no fancy <1 fps scripts are necessary, because I think the worst problems are the head noise, color correction and/or brightening the night-shot video, and the dropouts that are aplenty (like in frame 26). I usually mask these with freezeframe (), but this time there are many tape head dropouts (white lines) that sometimes don't cause dropped frame, but are taking half of the screen. I doubt there is an automated solution, but maybe automatic detection?

Thank you,

GL

[Ghitulescu]

Head noise is not a problem because of the overscan (the last lines are reversed). You probably mean here the CCD noise.
The dropout line should have been corrected by a TBC (by masking) if available, and it's due to the mechanical tape damage, one can see it it he looks at the tape, probably caused by dirt on the guides. It cannot be repaired unless one repairs the tape (flattens it with an iron - children don't do this at home) or by using various methods to increase the contact between the head and the tape (provided the magnetic layer was not damaged - for both cases).
I would lower the brightness (it was shot at night) and raise a bit the contrast (I always do this for VHS). Maybe I'll increase a bit the gamma.

[yup]

GrofLuigi!
Try.
http://forum.doom9.org/showthread.ph...02#post1465602
I am not promise magic, but source will be watchable.
I see source, my script remove only random lines, for remove constant line need use mask approach and spatial filter. You can use
http://forum.doom9.org/showthread.ph...08#post1465608
yup.

[GrofLuigi]

I might need to go with another recorder after all. The sound is also damaged, very dull, and only the lower frequencies exist.

I cleaned the heads and removed a ton of grime, but I'm afraid to put another tape in this one (doesn't have strength to rewind or FF, and on eject doesn't pull the tape back into the casette). I guess the rubber parts of the mechanism are solidified/baked.

@yup, I migt try that script just out of curiosity, but I need to gather the correct versions of plugins. And, does it make a difference that my noise is white?

GL

[2Bdecided]

Originally Posted by Ghitulescu

The dropout line should have been corrected by a TBC (by masking) if available

IMO that's not the job of a TBC at all. There's a dropout compensation circuit in "modern" (i.e. not the oldest) VHS machines (typically "VHS HQ" machines (remember them?), and newer) that copies the previous line over when the signal drops out. This may or may not be happening here - it's difficult to tell because it doesn't always work very well - especially when there's more than one line of damage.

A TBC fixes the timing - it doesn't do a thing for drop outs.

Cheers,
David.

[TheSkiller]

Originally Posted by GrofLuigi

[...] and today I discovered two of my three recorders had died.

If they just don't switch on at all or switch off by themselves when you try to rewind for example, then it's probably just some dead capacitors in the power supply. Not too expensive to get that repaired, the question is just if these VCRs are worth it. If these are high quality S-VHS decks or at least solid VHS decks then it's definitely worth it (assuming there are no additional defects).

The first thing I would do before using a VCR to capture important footage is to clean the head drum and the other parts that come in contact with tape (capstan etc). The most effective way to do that is is to get some 99%(!) Isopropyl and clean print paper (never use q-tips on the head drum!). Cut the paper in stripes of about two fingers height. Open up the VCR and put Isopropyl on one stripe of paper and gently press the wet stripe against the head drum. Use your other hand to gently rotate the head drum from above (don't touch the sides). Repeat this as long as there is dirt on the paper. Here is a guide (it's German though, but the pictures should give you an idea). Don't be afraid, if you're careful there's not much you can do wrong.

Your capture looks a lot like the VCR's head drum was dirty.


Oh, and as Ghitulescu suggested, the brightness of your capture is way to high. Histogram of frame 12:

That small spike on the left of the luma histogram (top one) is most likely that small piece of blanking on the very left of your capture. This is supposed to be black (it should touch the red bar, just like the spike on the very right hand side touches the red bar). However, I would not touch the contrast, it seems to be correct.


Edit: I once had that problem with very dull sound on a VCR too: the problem was a thick film of dirt on the normal-audio head (which is not located on the rotating head drum by the way).

Edit2: Just read your newest post, yes that sounds like dead capacitors in the power supply.

[GrofLuigi]

@TheSkiller: Yes, I clean it like that, it took one whole A4 sheet of paper until there was no more dirt. And I used smaller pieces too.

But, I don't want to put another cassette in this one because I don't want to do any more damage.

#1 is (was) S-VHS JVC HR-S9200 that shuts down half a minute after the power button is pressed or a tape is inserted. A repairman told me a logic chip is needed, but nowadays it cannot be found. Before, it did it interminnently, now it can't stay on ever.

#2 has worn out heads (every tape dropout causes similar white streaks for minutes, often until the end, and this tape has many ). It is used now only for rewinding.

And, I used both "Extend luma white/black point" in VirtualDub because I forgot which one is needed (this is a new installation; previously I had tinkered a lot with it and had it correctly). And I didn't even have a chance to play with BT Tweaker this time.

Edit2: Also, I turned off "Encode/Assume normalized YUV" in Picvideo MJPEG codec on a whim since I THINK I read somehere recently it's not needed for this... Jeez, I had forgotten all about this color thing in the last few years and it's just one big mess in my head right now.

@2Bdecided: It says HQ on the front.

GL

[GrofLuigi]

Off topic:

TheSkiller, now you got me thinking about this capacitor thing. Are you sure it can/will help to replace them in the power supply? This repairman was the only one I could find nowadays, and although he wouldn't dupe me, he's probably not very skilled.

Off off topic:

Once upon a time I "repaired" an AIWA VCR by replacing the diodes in the power supply (on visual inspection they were charred) and put much stronger ones. But it extended its life for only a month. In the end, the capstan got deformed (not the shaft, its bearing/chamber/how do you call it - the hole it goes into). So to repair it, the whole mechanism would need to be replaced.

GL

[TheSkiller]

Originally Posted by 2Bdecided

There's a dropout compensation circuit in "modern" (i.e. not the oldest) VHS machines (typically "VHS HQ" machines (remember them?), and newer) that copies the previous line over when the signal drops out.

Really? I never noticed something like that and all the VCRs I've ever used were HQ ones (non-HQ ones are pre-1985). To "copy" a line the VCR would have to digitize every line and store it in memory. Sorry, but I doubt that any VCR does that, especially ones without a TBC.

Originally Posted by GrofLuigi

TheSkiller, now you got me thinking about this capacitor thing. Are you sure it can/will help to replace them in the power supply?

With a VCR that old, they are not at their full power anymore for sure. Look at the VCRs display (you know, where it displays the tape counter). Do the numbers and symbols look rather dim and change brightness when you try to play or rewind a tape?

Edit: Actually, every electrician should be able to replace those capacitors with new ones, you don't have to be a VCR expert to do that.

[Ghitulescu]

The dropout corection is what most people like to call it Line-TBC. Sorry for the confusion I created. Of course a TBC should be able to keep at least one whole field in the memory.

The capacitors thing is a common source of failure for most if not all consumer gear*. If one is not so skilled in replacing them, the only alternative solution, which I know for sure to work with Panasonic, is to keep them uninterrupted under tension. Never take their plug out of the socket. Panasonics are also prone to capacitors failure on the head pre-amp (located near the drum) - they loose the colours and generate dropout-like symptoms.

Repairing a VCR to save a tape is the best solution next to have another one, a good one, a pro model if you can have one (they are populating eBay, as they are discharged from TV stations, studios etc.).

*I had once a NV-J35EG (PAL Germany). It was used, then deposited. After some 3 years I had to sell it, for space reasons. I had a short test and the symptoms were "snow" and playback problems. After letting it on for 1 day, things went back to normal. The pro decks I have from Panasonic never had this issue, although I haven't used them in 5 years.

[jmac698]

Where I am, VCR's are getting dumped on the street every other week. At some used stores you can still buy them for USD $10. I would just get any cheap VCR and digitize it. People seem to worry so much about getting a good VCR, I don't agree that there's a significant quality difference. To test this, I took a capture of the same tape through 3 different VCRs. I have posted the results. It's pretty hard to see the difference, but there are some slight changes in color.
I think that if I calibrated each one, they would be nearly indistinguishable.
Yes, there can be problems in a particular random VCR - so I have 5 stacked on my workbench, I've come to discover which one works best. The total cost was far less than a top-end SVHS (and I have two of those as well, and they aren't my favorites VCR's either).

[yup]

jmac698!
I agree with You. I make capture one bad VHS tape on 5 different VCR, 2 was from company which business digitize old VHS, 2 from my friends and last DVD recorder with VCR Mohican's at dead market VCR. Result practically the same. If this copy not original and VCR have dirty heads during recording it is impossibly find better VCR for capturing. Also if heads camcorder was dirty the same result.
Only digital processing using Avisynth or other software can help. I am also try solve this problem but now not at end point.
yup.

[GrofLuigi]

Well, since this topic has evolved it might be moved to hardware or capture...

Anyway, that was my idea too - since the two better units were unusuble, I tried to capture with the one that worked somewhat - but forgot to plan ahead (head cleaning/checking histogram), I guess hoping to fix any irregularities afterwards. Oh well. There is a good chance I will find a VCR to borrow.

But the idea of replacing bad caps is still itching me. I have access to few more presumed dead units. TheSkiller, I'm familiar with the phenomenon of dimmed display, but I'm still unsure new caps would help the motors to gain strenght. I think those VCRS are more in need of replacing rubber parts and/or logic chips, which are both hard to find nowadays.

GL

[Ghitulescu]

The pro units don't automatically have a better image quality - as sometimes people tend to say, especially when they sell one on eBay , they are more reliable and generally more versatile. Some don't even work if auxiliary equipment is not present (various cards and/or extensions).

However, the digital image is different from what the eye sees. I tend to do all the processing in hardware and I am positive that images that otherwise appear identical to the eye are different to the encoding engine (I'm using a pro DVD recorder), especially when the bitrate is limited. Of course if one heavily relies on postprocessing on a PC, the results are generally the same.

[jmac698]

both yup and I have worked a lot on the white lines problem. I found a 100% fix, but it involved recording several copies of the same tape. I don't agree the magnetic particles have flaked off, I think it's just a matter of signal strength too low for various reasons, tape wrinkles, dirty heads, maybe alignment problem. The video is FM modulated so if there's a signal loss, the circuit seems to default to white. The comet-like appearance would then be the based on the electronics frequency detector loop or something like that, it should be fairly predictable. Therefore the white lines should have a definite profile.

[2Bdecided]

Originally Posted by TheSkiller

Really? I never noticed something like that and all the VCRs I've ever used were HQ ones (non-HQ ones are pre-1985). To "copy" a line the VCR would have to digitize every line and store it in memory.

No it wouldn't - there have been glass delay lines in TVs for nearly 50 years. All PAL VCRs have 2-line delay lines for chroma processing, and 1-line delay lines for luma filtering. Typical delay lines in VCRs use CCD "buckets" - 13.3MHz analogue sampling.

I was wrong about VHS HQ specifically having drop-out compensation - AFAICT all machines have this facility (Videocassette recorders - A serving guide, 3rd Ed - Steve Beeching, Heinemann, 1988 if you're interested pp44-48 and p33). VHS HQ extends the filtering over more lines (amongst other things).

Apparently it's possible to bypass the drop-out compensator in older VCRs (discrete circuity!) and see just how awful VHS looks.

Cheers,
David.

[2Bdecided]

Originally Posted by jmac698

I would just get any cheap VCR and digitize it.

You really need a decent fast-acting TBC somewhere - either in the VCR, stand-alone (though many dedicated stand-alone TBCs aren't fast acting) or in the capture device itself (e.g. in some DV camcorder's analogue input stage).

This isn't for drop outs - it's to keep the signals in-time, which visually means vertical lines stay vertical (rather than bent or very wobbly).

Cheers,
David.

[Ghitulescu]

Yes, drop-outs should be corrected in the player. After the signal left the player, there's no room left for correction, other than masking (similar to drop-line compensator).