gemini://zaibatsu.circumlunar.space/~solderpunk/gemlog/video-tapes-and-helical-scan.gmi

For whatever reason, I really enjoy reading up on all the nitty gritty technical details of old media formats. It usually proves extremely interesting and education. Sometimes that's because I'm reading about something that I've never seen before because it wasn't commercially successful or launched at all in Australia where I grew up. LaserDisc, for example, was a video format from the late 70s which stored analogue video (in the familiar NTSC or PAL format) on something which looks an awful lot like a CD or DVD, but which was closer in size to an LP record, being 30cm in diameter! The same "pits and lands" structure used to store digital data on a CD or DVD is used on LaserDisc, but the discretely varying reflectivity is interpreted as a pulse-width modulated signal whose continuously varying duty cycle is the actual information content. Other times, I end up learning a bunch of stuff about a format that I used extensively but just never understood in any technical detail because, even as a curious and technically-inclined kid, in the pre-Wikipedia days it just wasn't so easy to learn stuff you couldn't easily figure out with your own hands and eyes. This entry is about one of those cases.

I grew up with casette tapes as the dominant audio format of my youth, and they were pretty easy to understand. When the door of a casette deck opens to accept the tape, you can peek inside it and see what's there , and even when it's closed with a tape inside, the doors were always transparent so you could keep looking and be sure there were no surprises. There's nothing too involved, mechanically. When you press play, the tape head moves into position to contact the tape, and the tape is drawn across it, off one spool, onto the next. You can visualise the audio signal as a straight line along the tape, read off in a simple, linear fashion. Easy! I had always just assumed that VHS casettes were more or less a larger version of the same thing, but this is actually very wrong, in interesting ways. Unlike a tape deck, the insides of a VCR are rather more hidden from view - okay, you can lift up the flap to peek inside, but it's dark in there, and the slot is narrow! And it turns out there's neat stuff going where you can't see it.

The difference between these two tape formats ultimately comes down to information density. There's a lot more information in one second of video than in one second of audio. With the simple linear reading solution used by audio casette tapes, the only way to read more information off the tape in an given unit of time is to increase the speed at which the tape moves across the head. This imposes various practical limitations: ultra-fast tape is liable to wear out quickly, requires larger, heavier, noisier motors, and probably most important of all, means that a 2 hour film would require such a great total length of tape that the spools to hold it all would be enormous and unwieldy.

The solution to this problem used by both VHS and Betamax is called helical scanning. The tape head itself actually rotates against the tape at the same time as the tape is drawn past it, and furthermore the tape head is inclined at an angle to the tape. As a result, the signal isn't visualsed as a straight line drawn across the tape, but rather as a series of closely spaced, roughly diagonal (but actually slightly curved, hence the name "helical") lines. That's probably quite unclear in words - check out the Wikipedia article on helical scan for a diagram which hopefully makes it clearer. With this approach, the relative speed of the head against the tape is much greater than the actual speed at which tape is drawn across it, resulting in a high information rate without any of the drawbacks mentioned above, making the system practical for home use. It's seriously clever and I'm really impressed that somebody thought of this! It would never have occured to me.

There's a cost to this, of course, which is increased mechanical complexity compared to the relatively simple situation in a tape deck. The VCR actually extracts quite a length of tape from outside of the casette, and uses a series of rollers on moving arms to press it up against, and almost wrap it around, the rotating tape head. There's a really nice video of this at Wikipedia, showing the process from multiple angles:

I had no idea this was happening inside my VCR! To my memory, audio casette tapes were a lot more liable to get chewed up by the tape deck than VHS tapes were liable to get chewed up by the VCR, but now knowing what happens under the hood this seems like a perfectly backward state of affairs. Is my memory bad, or was I lucky enough to grow up with an unusually high quality VCR?

Alex over at "yet another tech gemlog" made a response, of sorts, to this post, discussing various other technical aspects of both audio and video tape, which you should definitely check out if you are at all interested in tape tech:

The stuff about different kinds of tape chemistry is super interesting. I was totally oblivious to this back when I actually used tapes. I bought my blank tapes from the supermarket and, being pretty young at the time, cost would absolutely have been the driving factor. I was not unaware that different tapes could potenitally yield better or worse results - I'm pretty sure at least one of the brands in my market used little RPG-style pentagonal attribute score graphs to try to convey what different tapes were best for - but any kind of objective technical on the variation was totally absent, both from the marketing and my head. I only became really cognizant of different tape types some months back when I started to entertain the idea of using physical media to readjust my music listening habits.

Alex mentions the concept of "azimuth", the angle between the tape head and the tape itself, which should be as close to 90 degrees as possible for maximum accuracy, and also mentions high quality Nakamichi tape decks. During the reading I did which lead to make this post, I read about Nakamichi's "Dragon" tape deck, considered by many to be the best deck ever made. It can use a servo motor to slightly rotate the head to adjust this angle, and is able to detect and automatically correct for azimuth error on tapes which were recorded on poorly aligned equipment. Incredible stuff!

Video tapes and helical scan

Responses to this post