The first song videos I ever saw were made by people who owned one VCR and had a friend with another. They got together, cabled their machines and took out their stop watches. First, they would figure out how many seconds a phrase of music was, then they would begin to make an assembled, normal video tape consisting of clips of those lengths. When finished, they would copy the assembled tape using the video cable from the play machine and the audio cables from a tape player. The result, of course, depended on when the music began to go in.

There were glitches on the tape and scenes that changed out of synch but, on the whole, I found it amazing that they were able to make them at all. Again, it speaks to the level of obsession this hobby engenders, that people would NEED to make them, even with inadequate equipment. It also speaks to the obsession of song video viewers, who were able to see past technical problems to what the songmaker was saying.

The next level of videomaking takes place with VCRs with audio dub. They still built the assemble tape the same way, but now they were able to put the music directly on the tape. Not only did it save a generation of video quality, it was also easier to position the music so that the start of the song, at least, had some semblance of synch.

Songmakers quickly came to understand that what they needed was a VCR with video AND audio dub. Let's stop a moment to define some terms. When you hit record on your VCR, the tape shows a glitch at both the start and the end of recording. You can avoid the glitch at the beginning by playing the tape, pausing it, and switching into a paused record mode. This is what I'm calling assembling. The start is clean but the end is still dirty. Dubbing is the only way to get both a clean start and a clean end to a clip. It records material into a preexisting recorded track, replacing the audio and/or video.

Dubbing is a word VCR manufacturers love to overload. One common usage is the way I am using the word. Another term used is "insert editing." The second overloaded definition of dubbing is a type of copying from one machine to another (e.g., I'll dub a tape for you.). The problem with this use is that it comes from the professional world where dubbing means to copy using a special wide cable that produces the very best quality picture between two machines. This is NOT what any consumer document means when they use the term.

My first editing machine was an RCA VHS recorder. It didn't have flying erase heads and therefore every edit created a stretching of the record tape which is seen as "rainbows" of color in the video. Until I had it repaired, it was able to redo the same edit was very little positioning error. Consumer VCR's only know where they are by "counting" tics, a very imprecise method of working. Professionals call this working with control track.

My next editing machine was a Sony Betamax 1000, which recorded in SuperBeta I High Band. This machine was worth it's weight in gold and I have had people go home from working with me and scour the country for used ones. The claim to fame of this machine, in my opinion, was not just the quality of the picture, it was also the fact that the Beta 1000 was rock stable in positioning for an edit. It was ALWAYS off by one frame, and always in the same direction. This meant that I could always compensate and get an almost perfect edit. It also had flying erase heads so there were no rainbows.

What it did do over time was to begin to cause a rip in the first frame of the edit (I'll explain frames later), a problem that produces a visual flash as if you left in one frame of old material. That machine must be well known by now within the Sony repair community because the head engineer searched everywhere for information on the cause of that problem. They were kind enough to let me help set up test cases but, in the end, the frame still ripped. I could compensate by working my edits backwards, but it wasn't ideal.

There is a new Sony betamax editing VCR out, but it lacks the precision of the earlier model so I won't replace my old ones with it.

There was an intermediate consumer/commercial Sony beta machine out called the ED-Beta VCR. This gave a picture as clear as a laser disk and had the stability of the Beta 1000. Unfortunately, it was also incredibly expensive and reproduced noise almost better than it reproduced signal. In a choice between recording on the old Beta 1000 or the ED-Beta, I would still choose the ED-Beta, but not joyfully. I have one of these and wouldn't try to get another.

Panasonic makes a low-end editing SuperVHS machine that has a good picture but lacks the stability of positioning that makes song videos possible, in my opinion. I use one to collect TV shows, but not for editing.

A videotape can be thought of as having a video track, two audio tracks, and a timing track, or control track. When you record, you lay down all four. When you assemble, you lay down all four. When you dub, you leave the control track untouched and replace the video track and/or the audio tracks.

In the United States, video is recorded every 1/30th of a second by scanning across the picture 525 times from left to right. This 30th of a second snapshot is called a frame. In the "old" days, you could pause your VCR and get a beautiful still picture. This was because the machines didn't use digital technology and showed you the entire analog frame.

But frames could have a problem. If you dug deeper, you found that a frame was made by scanning only the ODD numbered lines first, then the scanner goes back up to the top of the picture and scans the EVEN numbered lines. This means that a frame is actually composed of two parts, or fields, snapped 1/60th of a second apart with only half the resolution you expect. You lay one on top of another and, voila, you see a whole frame.

Of course, if that frame consisted of fast moving video, it really did move over that 60th of a second, so what you saw was jittery. Another problem that could crop up was that the VCR just grabbed two fields that were next to one another and might give you the last field of a snow bank and the first field of a fireplace.

To get away from both of these problems, VCR manufacturers took advantage of the digital technology and kept a piece of memory that could hold one field, and this is what the VCR shows you these days when you still, or pause, a picture. This "still-frame" is actually a field, perfectly stable but with only half the resolution you see when the picture is running and you're seeing frames.

This standard of 525 scan lines and 30 frames per second is known as NTSC. All VCRs that you will come across, or TVs, or laser disk machines, etc., handle NTSC. This is why you can cable any machine together with any other. You can copy between VHS and Beta, or laser disk and VHS, or SuperVHS and VHS, etc.

In Europe, on the other hand, they don't use the American standard. Their video is captured only 25 times per second, but it is scanned in 625 lines. This means that European TV is not as smooth for action, but has a more resolved picture. It also means that you can't cable a NTSC machine to a PAL machine and expect to get a copy. You have to do a conversion, which takes a black box, or a special VCR which is designed to convert and has the black box built in. The other obvious corollary is that you can't record a TV program here and send it to a friend in England and expect them to be able to watch it if they don't have a conversion machine. And visa versa, of course.

I should add that there is another standard which is not as common, the French SECAM standard. This produces the very best pictures. Not only do they take 30 snapshot frames per second, they scan each one in 625 scan lines. The same conversion problems apply here as they do to PAL.

The reason why you often can send an NTSC video to friends in Europe is because Europeans are far more likely than Americans to buy equipment that can read more than one format. But your vanilla TV or VCR won't be able to read the tape your European friend sends you.

In the dead space on the tape that happens during the time when the scanner is repositioning itself, horizontally back to start scanning the next line or vertically back up to the top of the tape to start scanning the next field, there is more than enough room to store vast quantities of information, such as closed captioning. It is also possible to place a number into the tape which will uniquely identify the particular frame. These numbers are called "time code."

Time code can be placed within the video (Vertical Interval Time Code (VITC)), or along or in the audio track (Longitudinal Time Code (LTC)). It is these numbers that allow professional video machines to be absolutely reliable and absolutely repeatable. LTC is not accurate at slow speeds, where VITC is. Think of it as playing a tape on which you've recorded a baseball game. When you pause the tape, you can see the scene but you cannot hear anything. In the same way, if the timecode is LTC, as the tape slows down, it begins to approximate the time code number. In other words, it switches from reading an accurate number to counting tics to guess where you probably are now. The only safe accuracy is found by recording VITC. If you record BOTH VITC and LTC and you don't make them record the same number, then you see one number when playing and a DIFFERENT NUMBER when slowing down - a major potential source of problems.

I can strongly recommend the Panasonic SuperVHS 7750, which is an older model of professional editing VCR. It has a built-in Time Base Corrector (TBC) which rebuilds a frame to try to clear up normal analog degredation. It also has a time-code board which can handle VITC. To use the VCR, you need an edit controller which is able to read and write VITC. Luckily, my husband wrote one, so we can run the edit controller on a PC we attach to the VCR.

In the best of all possible professional worlds, edit information is maintained in an Edit Decision List (EDL) which keeps track of what came from where.

I was able to convince an IBM Vice President of Research that I would be far better off making videos than working on computer language design. He gave me an international multimedia magazine and went away to let me figure out how to do it. Luckily I had my own private expert - my husband, Paul. Paul came over from the other IBM Research building at night and discovered two 1" broadcast machines (that is, the tape width is 1" rather than the 1/2" of VHS) in various offices. Since they weren't in current use, and they did ultimately belong to my VP, we got permission to put them together. Paul hooked them up and read the instruction manuals. So I learned to edit off the console of the 1" machines in exactly the same style in which I edited off my home Super Beta 1000's. Of course, I had better posture at IBM, because if I slumped in the closet where the VTRs (video transport recorders) were kept, I got poked in the back by the cameras we also "discovered."

I got sent to a video editing week to learn to edit to voice, and I was on my way. Eventually, a group formed up around my magazine, and Paul came over to join. I still reported my project to the VP, who wasn't interested in managing me closely, and only my body belonged to my manager. You can't ask for a more lovely work environment. Eventually, Paul built a half million dollar studio around me. We had effects equipment, gorgeous cameras, wondrous lighting, and great audio equipment. Paul wrote an edit controller program that ran everything, and then wrote a router that let you connect various pieces of equipment together through the computer. He gave a paper on that router at the Society of Motion Picture and Television Engineers, a group to which I belonged.

Every year we got to go to SMPTE and National Association of Broadcasters conventions to keep up on the state of the editing art. As a result, I had one of the most bizarre accumulations of video knowledge you can imagine. Yes, I did know that the control panel with which they destroyed the Star Wars death star was a Grass Valley but, mostly, I had a weird conglomeration of computer and video knowledge. I could explain 3/2 video pulldown from 24 frame film, or field dominance during video recording, but I had never edited on a standard video editor, like everyone else in the video world. On the other hand, I did have a 25 year computer history, and had gotten my masters degree and taken the courses for my PhD (never had the guts to take the exams). I was the chair of our international computer community for programming languages, and ran conferences and workshops, and edited newsletters on computer languages. It was sort of like being neither fish nor fowl. I was too arty for the scientists and too scientific for the professional video people. Luckily, I'm perfectly happy being a griffin, so I had a ball.

And because Paul wrote the software through which I edited, everytime I wanted to have a little more help in some project - a variation I'd want in logging, for example - I could go to Paul and he could rewrite the program to make it easier for me to do something.

So what can I tell you about using professional equipment? It produces an almost perfect picture, with minimal degradation over copying. When it breaks, you mortgage your soul to have a repair person in to fix it. And you don't fix it yourself. You'll always have a record of every edit that you do, so if you want to make a change, you can get the data to tell you where you took your source from and where you put it.

But you're still working with analog tape, and you still can't make the video longer or shorter without copying it - a very tedious procedure, and one that you don't do by choice. There really was only one future, and the video world was already seeing it in 1988 when I was making videos for IBM Research. That future was non-linear editing.

Non-linear editing means that you don't make a video -- you make a computer description of what the video would look like if it really existed. When you say "preview this edit", the non-linear editor shows you what the edit would look like, and when you say "do this edit", the non-linear editor just logs the information into the edit decision list, or EDL.

When you're done, you can make a real tape copy of your program by telling the non-linear editor to create it. Since all the clips are stored digitally, you don't even have to stay around to change tapes. Just come back when it's done and take the tape away.

Non-linear editing has been around in various forms since 1969. It came into its own professionally around 1984 and, today, is the way things are done professionally.

We were very slow to pick up non-linear editing because our home video studio was still run by Paul's edit controller, and going to non-linear meant letting all of Paul's work die. I was finally able to get a digital setup when Paul needed to bribe me into leaving home for a trip in our new motor home. He got to travel west. I got a digital editing studio in the bedroom.

Since everything I knew about non-linear came from a white paper I'd written while attending video conventions, I wanted an editor that would put me into the video world on whose edges I'd existed, rather than the computer world. For me, that meant Avid. When I was first writing my paper on non-linears, Avid was just another hole-in-the-wall, mostly known for great marketing and so-so product. Their brilliance was coming up with a common connection and publishing it. After that, they took off.

I've only been editing with this equipment for 7 months, as I write, so my style owes much to my years of experience on my old Beta 1000's and the 1" broadcast equipment. We have a computer connected to a DVD player, a DVD recorder, occasional other input devices (laser disc, SuperVHS, audio cassette, etc.) and, most importantly, 7 huge external storage drives, ranging from 200 Gigabytes up to 400.

I start by deciding which disk drive will hold which fandom. Star Trek is so huge it sits on two. I go through the DVDs and capture into the external drive all scenes I remember having used in videos over the last 20 years, and whatever incredibly attractive scenes I can't resist. Each clip is named with the name of the episode from which it comes, so that the list of clips can be organized for visual scanning alphabetically. In Avid Express I can define separate projects, so every song is, for me, a separate project. The clips I capture as source from a particular TV show becomes its own project, too.

Once I have enough source with which to work, I pull an old copy of my video into a timeline and, after making sure I have a clean audio track, begin replacing the video, scene by scene. Since I can see the audio wave file while I edit, I am usually making different decisions this time around as to where I want clips to start or end. I can actually see the 1/30th of a second before the new music phrase starts. Truly glorious. And if I don't like the color or the brightness, I can select that piece and change it, just as I could in the professional studio at IBM. All of these changes are being remembered in the computer list.

When I've replaced all of the video in the song, having made new versions as I go so that I have backup versions in case I really screw up, I save the video in multiple ways. I make one version of the video that can be played on a DVD, and I link that video into a growing DVD description. I make another version of the video that can be uploaded to the Internet. But both of these are degraded and wouldn't be of any use to me if I needed to work without access to the original source. So I then make a backup DVD for each song separately. The backup includes the project information, as well as copies of each clip I used, with 10 seconds of slop on either side of the actual in and out points. These clips get copied onto a Read/Write DVD so that I can keep changing the saved version as I make changes to the video.

We've had external drives fail. We've had backup disks fail. The more ways you duplicate your material, the more likely you are to be able to recover. All things considered, I'm thrilled that we've gone to non-linear, and I'd have a lot of trouble going back to analog.