These days, if you want to install some sort of security camera system, you can just run
to your favorite big box store or wholesale club, buy a kit with 4 cameras and a DVR,
and within hours have a working CCTV system capable of recording days or weeks or even
months worth of HD, or sometimes 4K security footage.
Now we can even access them remotely over the Internet, or put them in doorbells, or
backup the footage to the cloud, or any number of things which, depending on your point of
view, you may find interesting or…
concerning.
But…
Not that long ago, we only had the world of analog video.
Sure, videocassette recorders were widely available, but even with the most generous
of tape lengths and the slowest of recording times, you're not going to be able to get
more than 10 hours or so out of a tape.
That's not great for security cameras, particularly if you want to be able to record footage over
a weekend when no one is at the office.
But this video is running at 30 frames per second and for surveillance footage, that's
probably more than we need.
If there were a way that we could reduce the frame rate, we could get more time out of
one tape.
If we recorded, maybe, only two frames per second, we'd still have a pretty good record
of what happened, and we could get 15 times as much footage.
Well, that idea did come to fruition in the form of the time-lapse VCR.
I've known about these things for years, but there have always been some nagging questions
in my mind about how they work.
For example, most VCRs just kinda…
suck at showing a still image.
If you hit pause on this one, you'll inevitably get some sort of snowy garbage at the top
or the bottom.
In later years, a digital frame buffer would grab a snapshot of whatever you were playing
so that there could be a nice clean still frame, but a time lapse VCR that's RECORDING
footage, how does that work?
Now I could do some research and try to find out exactly how a time lapse VCR is doing
all the magic stuff it does, but it sure would be great to just have one that I could poke
around with.
Oh right!
This is a Sony SVT-3050.
You can tell that this is a legitimate piece of A/V equipment because of its two-tone beige
and brightly colored transport buttons.
Oh yeah.
Matte black?
Heh, this is where it's at.
Seriously, why was so much professional A/V equipment from the this time period made in
this particular color scheme?
It's like it's trying its best to be plain.
From the front, this looks like any normal VCR save for its relatively few buttons
(most are hidden behind the door)
and surprising omission of the VHS logo even though this
uses VHS cassettes.
Are we bitter about something, Sony?
We'll get to that in a minute because there is some mystery as to why the label's missing,
but what gets weirder is when you come 'round to the back.
For starters, you won't find standard coaxial jacks for a television tuner.
Nor will you find standard RCA jacks for composite video.
Instead, you'll find BNC connections, which were and are to this day very common in
security camera applications.
You'll also find a bunch of screw terminals for various things like timers and alarms.
For example, the unit can send out an alarm signal if the tape becomes jammed or it runs
out of tape.
And it can also accept an alarm input to start recording.
When I got my hands on this, the first thing I did (as would anyone, I presume) was to
open the case.
I wanted to see just what's inside this thing!
Oh.
There's a circuit board in the way.
No matter, we'll just pop that up, and look at that!
There are little pegs made specifically for holding the board up and out of the way for
servicing.
That's neat.
Under the board was a completely normal looking VHS transport.
Real talk, if you are into old VCRs like I am, you actually should open them up before
you test them to make sure nothing is stuck which might cause it to eat a tape.
This one indeed had something preventing the take-up spool from turning, and after exercising
it a bit the problem went away.
But, had I not checked, it may have ruined a tape.
These are the sorts of life changing tips you can find only here on Technology Connections!
One of the more curious things in here is the use of a modular power supply.
You probably noticed the standard IEC C13/14 connector on the back, and it's clear that
it wouldn't be very difficult to swap out this power supply for another.
This not only makes repair easier, but it also means that the same chassis could be
used in many markets.
OK, so with it operational, I popped a tape in and starting messing with the playback.
On any speed but the normal 2 hours, the tape simply juddered forward one step at a time.
Is that really it?
I had imagined that maybe the head drum would spin at a slower speed, or that the machine
would behave in some other strange way, but it appeared that this was simply stepping
forward one frame at a time.
Lots of VCRs can do that… so what makes this special?
Well, keen eyes may notice that the quality of that still frame is perfect.
There's no snow, there's no jitter, and there isn't even any sign that that's
a digital frame buffer.
To be sure that this wasn't playing tricks on me, I touched the spinning head drum while
watching the screen of the TV, and sure enough the picture was distorted along with my touch.
So, this picture is coming live from that tape.
That alone isn't so surprising, as many high-end VCRs with four heads can do that.
But what is a little more interesting is that it not only has to do that when playing a tape,
but it also has to do it when recording.
After much searching, I found a great source describing how time-lapse VCRs work, and one
of the more remarkable things about them is how precisely they can move the tape itself.
See, in normal recording and playback, the tape is continuously moving at a rate of about
1.3 inches per second, or 3.3 centimeters per second.
This means that with each sweep of the recording or playback heads, which happens 60 times
a second on an NTSC machine, the tape has actually moved roughly half a millimeter.
That means that the end point of the sweep is about half a millimeter to the right of
the beginning, imparting a slight angle to the track.
That's why a two-head VCR can't display a nice still frame.
Without getting too deep into explaining analog video, each head is responsible for one half
of the frame, or one field.
With a two head VCR, if the tape stops, then the heads are suddenly misaligned a little bit.
One head will start showing the correct thing, but it will drift into the adjacent track.
The same thing happens when the next head sweeps.
In essence it's always reading the wrong thing for at least some part of the frame,
and the result is a noisy, incomplete picture
But a 4 head VCR actually has two heads passing by with each sweep.
It's capable of reading either the odd or even fields at any time, and can switch between
them on the fly.
This allows it to perfectly display either the odd or even field without noise bars.
There's a great link in the description if you'd like to learn a little more about
the specifics here.
Anyway, that's all fine and good, but what's most interesting here is how the machine records
onto the tape.
It needs to get that correct angle otherwise the signals will be completely out of spec
from VHS, and it won't be able to play them in fast motion for the purpose of searching
for the moment that your restaurant was robbed.
Or whatever.
And that is what makes this more advanced than your average 4 head VCR.
When recording, the heads are completely inactive except for when they write that single field
onto the tape.
Most of the time they're simply spinning for the fun of it.
And here's the kicker.
The writing action is perfectly synchronized so that it happens just as the tape is being
advanced to the next position.
This means that the machine actually is recording in a completely normal fashion.
But, it is able to start and stop the tape with such precision that although it's recording
up to 86 times slower than normal, the signals on the tape are unaffected.
Take a look at this slow motion footage.
I've placed a marker on the tape so that we can see it advance.
The tape is completely stationary, but the moment that head is aligned with it, it moves.
And as soon as the next head is in position, it stops.
This extremely precise control of the tape's movement, along with the idling of the recording
heads until the moment that next frame is to be recorded, enables the machine to make
intermittent, single-field recordings onto the tape.
Pretty neat.
Now here's where things get a little weird.
Keen viewers will have noticed that there is an audio input on here.
If you're familiar with how audio is stored on a non-HiFi VHS tape, you'll know that
it's recorded linearly on the edge of the tape, just like a normal audio cassette.
How would audio work if the tape is moving in steps?
Well, the machine can only record audio in the normal 2 hour mode, and the 12 and 24
hour time lapse modes.
It can't record audio in the slower speeds.
But, this means that when recording in the 12 and 24 hour modes,
the tape isn't advancing in steps.
Instead, it's continuously moving very slowly.
The images are still stored in a reduced frame rate, but the continuous motion of the tape
enables sound, albeit very very terrible sounding sound.
So, the same thing happens.
The heads only write in individual bursts for each frame, but we get an odd consequence now.
See, if I want to play the tape back with sound, it can't play the tape back very well.
It's certainly not awful, but there are noise bars and other anomalies.
If we want perfect quality, we need to switch the sound off, and then it returns to the
stepping mode that it was using before.
And this may give us a clue as to why the machine is missing a VHS badge.
Technically, the format specifications aren't being broken if it is recording in steps like this.
It is recording a normal SP signal on the tape, it's just doing it intermittently,
one field at a time.
But this weird extra super duper slow speed doesn't conform to that standard.
Is this THE reason?
I can't say for certain.
The S-VHS variant of this machine, the Sony SVT-3050P, does have an S-VHS badge on the
front, and it, too can record audio in 12 and 24 hour modes.
So… the missing logo may remain a mystery.
But we're not done yet.
Did you notice that "Faroudja laboratories" mark on the back?
Yeah, that got me curious.
Who are they and what do they want with my VCR?
Well, Faroudja Labs was quite famous for analog video processing technologies, in particular
devices called line doublers.
These could take a low resolution image, and double the number of horizontal lines of resolution,
to give it an apparent resolution bump.
This device would really benefit from a line doubler, because it records just one field
at a time.
Each still frame is only half the resolution of a normal television signal, and for something
like security cameras where you might want to be able to identify who is stealing your
stuff, that's not good.
Now I'll admit, you can't get more detail out of an image than was already there, so
the extent to which a line doubler may help with tasks like identifiability is up for
debate, but it's clear that this technology is helping.
Part of why these still frames seem remarkably clear may simply be due to Faroudja's video
processing prowess.
Now I can't say for certain that there is a line doubler in here, but knowing what Faroudja
had done in the past, it seems very likely.
At the very least, something of theirs is making this a very sharp picture for what
the raw signal is.
So, that pretty much wraps up how a time-lapse VCR works.
It's a lot less magical than I thought it might be.
More or less it's just a standard VCR.
There were some special considerations to using it, however.
With a normal video tape, the tape is only touching the spinning head drum for a few
seconds at a time.
But with this machine, on the slower recording speeds, it might be touching that drum for
a few minutes at a time, and this could quickly wear out the tape.
For this reason, manufacturers would recommend that you discard your tapes after a certain
number of uses.
Some would say to only use a tape twice if it was recorded with the slowest speed.
When these were in common use, you'd need to balance the desired detail of the recording
with your business hours and when the tapes could be changed, and how long of an archive
you need to keep (for legal purposes, usually at least a month of recordings).
So, you'd need to have something like 10 tapes on hand, use them in a rotation, and
discard them after they'd been used perhaps a dozen times.
For a 72 hour recording, this would let you use 10 tapes for perhaps a year before needing
to replace them.
And of course, a 24/7 duty cycle for a VCR isn't the kindest thing.
It was recommended that you replace the video heads on this at least every 10,000 hours
(which works out to pretty much every year).
You can still find the drum, the 2N4N-Q, for sale from people who specialize in security
equipment.
And it looks like this machine has had its heads changed at least once.
Poking through the menu shows that it's been powered on for 25,716 hours, but the
video heads only have 314 hours on them.
Now, it could be that this machine was set up to only record in an alarm condition.
This may be the original head, and it was simply idle the vast majority of the time.
There doesn't seem to be a way to determine if that counter has been reset, so I suppose
I'll never know for sure.
But before I go, I haven't answered one particular question about these things.
What happens when you have multiple cameras?
Well, that's a fairly deep rabbit hole so I'll be brief.
One option was to use what's called a quad.
This splits the incoming signal of four cameras into a single display showing all four at
one time.
The major disadvantage here was the very poor resolution of the recorded video, as each
camera only got one quarter of the screen.
Another option was a switcher.
This would simply cycle through each of your cameras one at a time, showing Cam 1,
then Cam 2,
then Cam 3
and so on.
You'd get a full resolution image, but you couldn't see what was happening on any of
the other cameras when one was being shown.
It also makes reviewing recorded footage a little…
nauseating.
The coolest solution, albeit the most expensive one, was a multiplexer.
These would combine the signals from all of the cameras, and send them over a single wire
by rapidly switching between them, and using the vertical blanking interval to signal which
camera is which.
When you want to see just one, the multiplexer would use a frame buffer to show only that
one camera at a reduced frame rate.
The beautiful thing was that this also worked with a time-lapse VCR like this one.
But, you'd have to synchronize the multiplexer with the VCR (using the switch out terminal),
and you'd also have to deal with an effective frame rate that's even worse than what the
VCR is natively recording.
See, if the VCR is recording 4 cameras, and it's recording at a rate of 4 frames per
second, those four frames are split between the four cameras, meaning that each camera
only gets a recording once per second.
So, while you did get to record four cameras on one tape with the full resolution, you
would get a seriously reduced update frequency.
That was a significant trade-off, but it was cheaper than having a dedicated VCR for each camera.
If I ever get my hands on a multiplexer, I'll revisit this video because I'd love to see
how it all works.
I'd be curious to know what exactly it stores in the blanking interval, and it may be fun
to set up a CCTV system using vintage equipment like this.
But for now, I hope you enjoyed this video on the time-lapse VCR.
It's a lot simpler than I thought it might be, although it ended up being far more mysterious
than at first it seemed.
As I'm sure you know by now, I end my videos with a thank you to my amazing patrons, but
this time I have more than one reason to thank them, and someone else to thank, too.
This video topic was a viewer suggestion!
A friendly gecko named Torrance reached out to me and wanted to know a little bit about
these time-lapse VCRs.
I've always been curious myself, and thanks to the people who support the channel on Patreon,
I was able to go and buy one and poke around with it.
And in addition to their generous financial support, this time around the Patreon crew
came to my rescue by helping me figure something out about this machine.
I was under a very mistaken assumption about how this thing was making such a great still-frame,
and thankfully you sent me some great resources (including one linked below) that made me
realize I was really overthinking this!
So, if you'd like to send in a video topic suggestion, please do!
You can leave them here in the comments, you can reach me on Twitter @TechConnectify, or
you can check out the Technology Connections subreddit linked below.
Down there you'll also find a link to my Patreon page.
With direct viewer support, this channel has really taken off and there are very exciting
changes coming pretty soon.
If you'd like to make a pledge of your own and get perks like early video access, occasional
behind the scenes footage, and general updates before anyone else, please check out my Patreon page.
Thanks for your consideration, and I'll see you next time!
♫ violently smooth jazz ♫
...because of its two-tone beige and brightly colored transport (unintelligible babbling)
Hu huh!
Huh huh!
[And other weird noises]
...to make sure nothing is stuck and might cause it to eat a pa…
(laughs)
Eat a pastry!
Did you notice that Faroudja Laboratories mark on that back?
Well that got me…
Is it laboratories or technologies?
It's both!
OK…
With a normal videotape, the tape
(pauses)
Yeah, dang it that was right.
Mmhmm!
Say the lines as written, please.
It's a lot simpler than I thought it might be, although it ended up….
(sigh)
[very high pitched squeaky sounds]
No comments:
Post a Comment