Post by William Martin on Jul 7, 2003 16:22:54 GMT
This was started on the old forum I said something like the old domestic recordings should be fed slowly into a computer then analysed basically the video signal consists of 5 parts
1- clocking signal usually on a track at the bottom of the tape but sometimes in the video signal, this is used to synchronise the signal 2-end of line signal to mark the end of one line scan 3-end of frame signal for each 1/2 frame 4-the luminance signal the actual picture signal simply strong for light and weak for dark 5- sound track usually at the top
these will have to be decoded but there were only 4 or 5 types of encoding schemes used on old domestic video. this is only for B/W colour is more complicated but do-able
once enough control siganls have been recorded it will be possible to repair incomplete or missing controls by calculating where they should be so this should stop rolloing and other errors and even insert blank signal for missing picture info.
to answer one posting from the old forum, yes i agree it is like the baird video recovery and if that can be done then so can this. the important thing is to get the raw signal stored while the tape still exists in good enough condition.
so how would you read the tape? and with what? Do those little bits of rust present themsleves in an order that that identifies themselves as frame? as the sync pulses did in the old days of physical video edting
Post by William Martin on Jul 9, 2003 15:06:06 GMT
to read the tape, a special machine (or machines) would have to be constructed this would be geared to run slow it would have adjustable read heads it would be capable of reading the whole signal on the tape
speparating control signals from image signals would be left to a computer program It would be nescessary to aquire encoding data from all the most common video standard of the day from the signal pattern the system used could be narrowed down missing data could be bridged as far as control is concerned but only for small gaps magnatic powder could always be used to see the pattern and angle of scan on the tape which would also aid with identification and machine settings once a first frame has been identified, this could be used to set interval and control signals for the others based on signal frequency and identifiibly different signals that occur regularly.
Post by Andy Henderson on Jul 9, 2003 17:00:05 GMT
Looking at that problem with a bit of lateral thinking, it is (of course) possible to create a computer system which would process old recordings. The real question is why would we want to do this at present? There are still plenty machines around, tended by experts such as Martin Loach and Terry Martini. Playback will become a problem in the future, but at present there is little that can't be played back. There are exceptions of course, but relatively few.
I have had similar thoughts to William many times in the past.
We all know when we recover old tapes how some have such severe mistracking that it's nigh-on impossible to get a steady readback of the signal.
Computer restoration has, until now, simply worked on the principle of digitizing the already decoded signal as it leaves the VCR, rather than intercepting the mechanism of the VCR itself.
As William says, it should be possible to patch-over eroded control signal with a voltage-synthesis feedback loop which cuts in where the recorded control signal has eroded. This would have a similar effect to the Dynamic Track Following process used in the Video 2000 system, except it would be worked out via microprocessor logic rather than picking up signals adjacent to the video tracks themselves.
Alternatively, the recorded control signal could be ignored altogether and a graphic tracking sweep editor used, like the Tone Curve feature in Corel Photo Editor/Adobe Photoshop etc., whereby mesh points can be slid around until the skewiness of the video tracks is matched - producing a stable image which would THEN be digitized and cleaned up. Such software could, presumably, have an Auto function which would parametrically adjust the curve at pre-defined points looking for a best fit, requiring of course the video head output to be directly analysable by the computer.
As William says the demodulation, de-emphasis, etc. could also feasably be done by computer, so that the old electronics of the original format VCR wouldn't have any chance to kill off any horizontal resolution, colour-under carrier, etc. etc.
Mind you, this would of course only work on first-generation tapes of course - I'm not suggesting for one minute it would work on copies, where such problems will be burned-in.
Post by William Martin on Jul 14, 2003 15:34:55 GMT
sorry to keep on.. but would one of those old magnetic computer tape readers be any good for this?
not at all bang on as much as you want
reading the tape is not simple, a helical scan head would be best although a multiple read head with 50-100 fixed heads may work, an important thing is to treat the tape as carefully as possible hence the machine would have to be geared to run dead slow.
Post by William Martin on Jul 14, 2003 15:45:24 GMT
exactly what i was thinking, this can be done it would just be time consuming but would lead to the recovery of lost material and an improvement on existing material such as the steptoe "wesgrove" and "cv2000" recordings. old video should not be played back on old equipment unless both are of good quality. first thing to do is get it transferred as best as possible signal noise can be cleaned up later. then there would be a good long breathing space to decode the recording by which time the origional tape may have degraded beyond recovery. with care even "sticky" tape can be read by reading the signal from the back of the tape whist it is on the spool then removing one layer cleaning it and doing the same again. redaing from the back of the tape is one trick i havn't seen done yet, the back should have less damage than the front and be smoother but would be more fragile.
Post by William Martin on Jul 16, 2003 15:06:18 GMT
I don't know, it may be possible (perhaps using magnatic powder) to read a visible pattern this would save wear and tear and any distortions due to poor tape condition and bad contact with the scanning head. I've thought about this but i don't know what a scan pattern would look like opticaly. The FBI did do something like this to check if the watergate tapes had been tampered with