Linear vs. Nonlinear Time Trials

When we’re testing products like batteries or cables, it’s easy to envy the editors of auto magazines who get to write stuff like:

The tires neighed and snorted as I slammed her through the slalom, but I twitched the wheel to rein them in.

While all we can do is sigh and murmur "vroom vroom" softly as we sit there swapping RCA plugs and dreaming of the fun we too could have if we tested products by hurling them through fabled road race courses.

What we can do, however, is use the metaphor of the race track to compare a typical linear editing system with a typical nonlinear system. Our Videomaker "test track" was a 10-minute video with 100 shots and numerous titles and effects (see Sidebar). The course was designed to push two thoroughbred vehicles to their limits and then some—a face-off between old and new, analog and digital, linear and nonlinear, and like that.

The linear system (see specs in sidebar) is a fabled workhorse of advanced amateur and entry-level professional videographers, with upgraded torque and horsepower added by a state-of-the-art digital mixer. Its nonlinear rival is the new kid: macho, assertive, rippling the muscles of its Pentium II processor under the hood.

To rate the two, we’ll push them into the pits and set them up for competition; then we’ll compare them, mano a mano, for speed through the course, handling characteristics, driver comfort, and overall quality.

So assume a look of breeding and old money, as if your other vehicle were a polo pony, while you stroll with us onto the test track.

Setup for Competition

To begin the trials, we stopwatched two top-of-the-line pit crews as they set up the competitors.

At the beginning, the linear crew began arranging decks and black boxes, while the nonlinear crew had the computer’s case open, preparing to plug a card or two into the motherboard’s PCI-bus expansion slots. By the time they had the computer’s cover back on, the linear crew had finished selecting their cables from an assortment of composite, S-video and FireWire options. As the linear crew checked to make sure all of the edit control cables matched up between the edit controller and the decks, the frustrated nonlinear crew was surfing the Net for the latest capture-card drivers.

The hookup phase was no contest as the nonlinear crew cabled S-video and stereo audio lines to the PC and then right back again. A similar run to the single monitor and the crew jumped back, arms upraised, to show they completed the first leg.

Meanwhile, the linear crew was running the feeds from two VCRs and a black box titling machine into the audio/video mixer and then out again to the record VCR, lashing up the four–count ’em–four required monitors (and trying to figure where the titler’s preview output went).

But once the lines were lashed and the lug nuts torqued, the advantage switched to the linear side because their setup was complete and they headed for the starting line.

In contrast, the nonlinear crew was still configuring the editing software, loading the capture card drivers, surfing the Internet for newer drivers, loading the newer drivers, configuring the software, surfing the Net again for tech notes from the hardware and software vendors, reconfiguring the software…To their credit, they were out there on the starting line, sweaty but triumphant, when the linear crew returned from a coffee break and the time trials were ready to start.

Speed Trials

Now it was time to push each beast, in turn, through a tough road course, negotiating the hills and curves of shot logging, data inputting, editing, revising, and outputting the finished product to tape.

In the first, shot-logging phase, the advantage shifted away from the linear machine as our driver juggled four source tapes, hand-writing shot names and descriptions. The nonlinear system had the advantage of an automated logging system, which aided in the task of searching the tapes and automatically saving in and out points. After defining each scene on the nonlinear logging system, the software even placed a tiny image called a picon at the beginning of each clip, to identify it visually.

Next, the driver selected the shots he would use. On the linear system, he did this by placing check marks near the shot names on the paper shot log. On the nonlinear system, he clicked a check mark into a box near the selected shot names on the screen. When the stopwatch stopped, the systems were even.

But when the digitizing commenced on the nonlinear system, the balanced tilted because linear editing skips this chore, assembling the show directly from the source tapes. The linear system was immediately ready to move into the editing process, jumping into a distinct (but temporary) lead. With the nonlinear system, however, the driver would have to wait for the computer to finish batch digitizing the clips before editing could begin.

But at the editing backstretch, the nonlinear system started using its random access ability, and it cut in like an afterburner. With all the shots on the hard drive, the driver could grab them instantly, stick them on an on-screen timeline, manipulate them, preview the effect, try out transitions and superimpose titles, all at once.

The traditional linear setup didn’t fire as well, with the driver searching tediously for every shot, making and re-making whole edits to revise the cut points, rehearsing and recording A/B-roll transitions.

When the driver had his shots lined up in the nonlinear system, he suddenly veered right off the track and headed cross-country. He was showing off nonlinear’s audio capabilities: layering and balancing multiple tracks–in stereo yet–cross-fading music, effects, and narration along a path where the more limited linear machine just couldn’t go.

Roaring back onto the test track, the driver turned the nonlinear system’s stopwatch lead into what looked like a rout, because he was now screaming into the revision phase. On this part of the course, the old-fashioned linear system looked plain pathetic because if the driver didn’t like a shot, he could only replace it frame-for-frame, or else re-assemble every blessed shot that followed it.

On the nonlinear system, the driver was pushing shots here and there, swapping footage instantly, trimming lengths, retiming transitions–even driving with one hand as if to show the competion how easy it was.

But at the last moment, the linear system bounced back. You see, the home stretch was output, where the arrogant computer-based system slowed to a glacial pace while the old-style edit system went zero to 100 in four seconds flat.

How come? Because it didn’t need any newfangled "output." When the driver finally did get through his revisions, he was almost across the finish line. The upstart nonlinear system, on the other hand, had to "render" the project, i.e. build a single file containing all the clips, effects and transitions, indicated in the timeline. All the while the white-knuckled driver hoped that the system wouldn’t crash. (The driver never would forget the time when the final compositing had quit because the computer’s power-saving feature cut in.)

So when all the dust had settled, there the two were at the finish line, with overall stopwatch readouts that looked nearly the same. (On this course, at least. For other kinds of speed trials, check out the accompanying sidebar.)

Handling Traits

And what was it like to drive them? Once again, each contender showed that they had both strong and weak points.

Though the linear system’s mixer actually offered plenty of flips and wipes and other digital video effects, the nonlinear software let the test driver vary motion speed, color temperature, and every other image characteristic you could think of.

As for titling, both systems offer a wide range of backgrounds, type faces and motion effects–more options than you could get through in a year.

In the win some, lose some department, the nonlinear system beat the pants off its older rival in sound management, as noted earlier.

On the other hand (as we keep saying), the storage abilities of the nonlinear system are laughable compared to the two-hour capacity of even a humble VHS tape. If you’re going to go nonlinear, you’d better stick to very short movies, or get yourself a very large hard disk.

Comfort and Convenience

The traditional linear system was easier to learn because the equipment was simple and intuitive. You could say that the wheel, shift lever, and pedals were right where the driver instinctively reached for them.

On the other hand, the computer-based system faced our driver with a learning curve as steep as its acceleration graph: a screen packed with multiple windows and menus within menus within menus, all supported by a lengthy instruction manual that can most charitably be described as a nice try.

But once our driver had the controls down pat, the nonlinear software proved equally intuitive, especially the timeline metaphor that lets you see a model of your movie as you massage it. (Some systems use a storyboard metaphor, which also works well.)

And to be fair about it, the nonlinear system also takes longer to learn because it has so many more features, which means that there’s much more you can do with it when it’s time to sit down and begin editing.

The bottom line? Depends on your personal style (and your types of programs, as the accompanying sidebar points out). Computerized editing is a far more detached, intellectual exercise than traditional cutting. If you feel more comfortable when you can get your mitts on actual knobs, sliders, cables and cassettes, then you may naturally relate better to the more traditional linear setup.

But if you routinely use computers, you’ve probably grown accustomed to mice, menu trees and all the little procedures that are now semi-standardized on PCs and Macs.

Overall Quality

So which car is better? Sorry, but there’s no escaping that wishy-washy old "it depends." In general, the linear system will outrun the nonlinear on long projects requiring long clips, few cuts, few special effects and transitions and simple audio. The nonlinear system will blow the doors off of the linear on short projects with many cuts, transitions, effects and sound tracks.

A second-generation master tape produced by the linear system looks better than the master tape made from the nonlinear system (though a nonlinear system could easily beat it if you forked over several thousand dollars more than the cost of our test setup). That’s because the nonlinear rig’s output suffers from visual artifacts.

Third generation tapes (dubs made from either master), will of course suffer additional generation loss–unless the masters and dubs are made on DV decks connected by FireWire.

How about reliability? Here the nod goes to the traditional vehicle. If you’ll permit a personal note, Videonics mixers and Panasonic VCRs have chugged away in my teaching lab five days a week for seven years without even a single breakdown. Desktop computers, on the other hand…well, you already know about their reliability, don’t you?

Besides, linear editing commands a mature, slow-moving technology in which seven-year-old hardware is maybe 95% as capable as the very latest stuff. On the nonlinear side, products are improving so fast that seven-year-old hardware is practically worthless. (Try video post production on a 386 machine with an 80 megabyte IDE hard drive!)

And now, if you’ll excuse me, I have to resume testing RCA cables.

LEAVE A REPLY

Please enter your comment!
Please enter your name here