Longitudinal Video Recording

~ the obit ~

Every videotape machine you have ever seen spins its heads. The tape crawls past and a drum of heads rotates fast against it at an angle, writing diagonal stripes. That arrangement is what makes helical scan expensive: a precision drum on bearings plus a servo to keep it phased. Longitudinal Video Recording proposed the opposite arrangement. Hold the head still, run the tape past it very fast in straight lines, then step the head sideways to the next track and run the tape again.

BASF developed the system and first presented it in 1974. Getting it in front of the public took another five years. English Wikipedia says LVR "wasn't demonstrated until Autumn 1978 at the Berlin Radio Show", and that cannot be right: the Internationale Funkausstellung ran strictly biennially in odd years from 1971 through 2005, so there was no 1978 edition to demonstrate anything at. The real showing was the 1979 Funkausstellung, which is the year the Magnetbandmuseum dates its detailed spec sheet from as well. Toshiba brought a rival LVR prototype to the Consumer Electronics Show in Chicago in June 1979.

By the end of that same year the whole thing was finished. Eumig in Vienna, owner of the Swiss camera maker Bolex since 1969, had taken the BASF design and was building the portable version when the project was canceled, in the phrasing repeated across the German and English sources, because the chances of succeeding for this system in the market were rated as too low. VHS and Betamax had already divided the ground between them. New Scientist, writing on 3 January 1980, recorded that the LVR never went to market.

~ the rap sheet ~

~ the idea ~

There is no single LVR spec sheet, because at least three different machines were described under the name and their numbers do not reconcile. The version New Scientist described in January 1980 used half inch tape with 300 tracks and a tape speed of 240 inches per second. The Eumig and BASF portable prototype shown in 1979 had 48 video tracks plus 2 audio and gave about two and a half minutes of play per track. Two hours in total, with a stated plan to double the track count and reach four.

The production specification recorded by the Magnetbandmuseum is the most detailed and the least like the others. It used eight millimeter tape 8.5 micrometers thick, carrying 72 longitudinal tracks each 68 micrometers wide with roughly 30 micrometers of guard band between them. The tape ran at 4 meters per second for 150 seconds per pass, which came to about three hours in total. Audio rode on two FM subcarriers at 162 and 204 kHz, and color was handled by sequential amplitude modulation of the color-difference signals. Grundig sold a related quarter inch variant under the LVC and CVC names.

~ the flaw ~

At the end of every pass the head had to find the next track, and that took about a tenth of a second during which there was no picture coming off the tape. The prototypes covered it by buffering a fragment in memory, in 1979, when memory was not something you spent freely. So the format's headline advantage, a mechanism with no rotating assembly in it, was paid for with a defect that recurred every two and a half minutes for as long as the tape ran.

The deeper problem was geometric. A head parked over a 68 micrometer track has to stay over that track for 150 seconds while the tape moves at 4 meters per second, and tape stretches. This is the same wall the BBC hit with VERA in 1958 and hit again with an experimental digital longitudinal recorder in 1974. I work on hardware where the sensor and the compute are argued over in millimeters and milliwatts, so the appeal of deleting the single most expensive moving part is obvious to me, and so is the reason it keeps not working.

~ nineteen seventy nine ~

Compress the timeline and LVR gets an unusually short life for a format with this much engineering in it. Announced in 1974, then not shown to the public until the autumn of 1979 and canceled before that December was out. Blaupunkt had started work on a competing longitudinal system by then, which on its own tells you the idea was taken seriously by people who built consumer electronics for a living.

Koichi Sadashige published "An Overview of Longitudinal Video Recording Technology" in the SMPTE Journal in July 1980. It is a survey of the state of the art in a field that had stopped existing six months earlier.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »