Who Will Wear the New 3-CCD Camcorder Crown?
Posted Jan 20, 2005

Planning a new 3-CCD DV camcorder purchase? Wondering which one best suits your videography needs? You're in luck—EventDV's first camcorder shootout compares three new 3-CCD models built with videographers in mind. They're all leading contenders for the current pro/prosumer crown: Canon's XL2, Panasonic's AG-DVC60, and Sony's HDV-capable HDR-FX1. Read on for the tale of the tape.

Few things in life are more fun than testing camcorders, because when you have great camcorders around, well, at least around here, it seems that more life happens. Canon's great new XL2 comes in and it's time for my daughter's first-grade class to sing for the Veterans at a local church. Seems like a great way to write off some mileage and DV tape, and butter up the first grade teacher, which always seemed to help when I was in school.

Then Panasonic's highly functional AG-DVC60 appears and my wife says that she's been promising some of the kids in her ballet class a DVD they can use to practice at home. Hmmm, sounds like a great idea for a multiple-camera shoot, and another great way to test out that cool multicam feature in Pinnacle Edition.

Then Sony's fantastic HDR-FX1 comes in, and the local Tae Kwon Do instructor asks me to film his seven year-old daughter's self defense routine. Seems she just won the world championship in Myrtle Beach, S.C., and wants to send off DVDs to Oprah, Jay, and Dave to see if she can get on their shows. No pressure there—fail and I can't show my face in town for the next two years.

Of course, one thing I've learned over the years is that weeks of laboratory testing can't replace the lessons you learn in one live shoot, especially when your marriage, physical well-being, or elder child's first-grade education are on the line. Now, if only I could convince my editors of that and dispense with all the grueling, time-consuming standardized tests.

Oh well, maybe I'll get to do that in the next camcorder shootout, but not this time. You've met the players; let's set the scene.

What's Inside
As you probably know, Sony's HDR-FX1 is a hybrid camcorder that can shoot in both DV and HDV mode. For more detail on the current state of HDV, refer to Eric Schumacher-Rasmussen's "HDV Goes Mainstream" and Stephen Nathans' "From HDV to DVD" in EventDV's November and December issues, respectively (I've learned later in life that it pays to butter up editors as well).

However great the promise of HDV, my take is that about 1% of the initial buyers of the FX1 will use it solely for HDV, and that most will use DV at least as much as HDV, if not more. Not only right now, I'd argue, but possibly as long as they use the FX1 as their primary camera, given the 2-3 year turnover in technology products at this level. For this reason, this article will focus exclusively on the camera's DV performance, with HDV performance and workflow examined in a follow-up article in the February issue.

Similarly, I'm thinking that few in our audience are concerned with a camera's ability to produce video they can later transfer to film (mostly an indie filmmaker application), and I'm confident they're pretty well-divided on the issue of whether you can actually make digital video look like film based on the technology you use to shoot it—or the effects you apply in post, for that matter. For this reason, I didn't compare the cameras based on the XL2's ability to shoot in 24 frames per second progressive mode (24p) or the FX1's CineGamma feature that reportedly makes your video look like film. (Although Panasonic pioneered prosumer 24p with the AG-DVX100, the DVC60 doesn't support it.)

Instead, I ran a battery of tests—some standard, some a bit off the beaten path—to illuminate the differences between these camcorders. When I could, I included the Sony DCR-VX2000 in these tests, just to see whether folks who own the VX2000, Canon XL1, or Panasonic AG-DVX100 need to upgrade to achieve additional quality.

All tested camcorders are 3-CCD camcorders. Just for the record, I checked prices at Express Cameras.com and found the FX1 at $2,899, the DVC60 for $1,799 and the XL2 at $3,799 with 20X zoom lens. The Sony DCR-VX2100—successor to the VX2000—costs $1,589.

I'll start by describing our formal tests, weaving in the real-world tests when they add value to the discussion.

Formal Tests
A quick note on how I ran the tests: I manually white-balanced for all indoor tests and optimized exposure, gain, and iris parameters for the resolution, color chart, and one chromakey test. I ran other indoor tests with manual white balance and all other parameters automatically set. After shooting, I captured all video and extracted comparative test frames in Adobe Premiere Pro, pulling together my composite figures in Ulead PhotoImpact.

In each test, I'll rank the cameras (including the VX2000) in reverse order, giving the top camera four points, and so on. Then we'll tally the points and see how the camcorders stack up.

I started by testing each camera's ability to capture detail by shooting the EIA Resolution Chart 1956. For each camera, scan down the vertical line near the middle that starts with 400 and works down towards 800. The point where you can no longer distinguish the converging lines is the lines of resolution for that camera.

It's kind of a fuzzy metric, but what's clear is that the FX1, DVC60, and VX2000 all show some white lines between 500 and 600, while the XL2 doesn't. Here I'll give first place to the FX1 for the smoothness in the gaps between vertical lines. Second place goes to the DVC60 for the same reason, with third place going to the VX2000. In the meantime, the obvious value of these synthetic tests is to predict how well the camera performs in real-world tests; hold that thought until we get to these results below.

Note that I shot these first tests at an aspect ratio of 4:3. Now let's see how the cameras perform at 16:9, which is the aspect ratio you would use when shooting for widescreen television sets.

16:9 Resolution Tests
Now let's look at the results of tests shot in 16:9 resolution. Here we saw a huge difference between the FX1 and XL2, both of which shoot in true 16:9 mode, which means that each has sufficient pixels on its CCDs to capture the complete 16:9 image and then squeeze it down for display on the DV tape. Note the clarity of the horizontal lines produced by both cameras, and how crisp the 200 (for lines of resolution) is on the upper left of both charts.

In comparison, the VX2000 uses a "letterbox" mode that cuts off the top and bottom of the frames, and tells the display device to stretch the video to fit a 16:9 display. However, as you can see in the comparative chart, this technique doesn't produce the same vertical quality as either the FX1 or XL2.

The DVC60 offers both letterbox and a "stretch" mode that supposedly simulates true 16:9. However, Figure 5 shows the results, and the quality doesn't compare. If 16:9 video is important to your customers, this is a negative for the DVC60 and perhaps a reason to upgrade from any camera that doesn't produce true 16:9 video. It's close, but in this test I rate the XL2 image slightly higher, with both the VX2000 and DVC60 rated even in last place.

Color Fidelity
Our next tests related to color fidelity, using a chart called the GretagMacbeth ColorChecker Color Rendition Test. Briefly, the chart contains 24 colors representing natural objects, like human skin, foliage, and blue sky. When comparing results, consider three characteristics of the images. First is the accuracy and vividness of the colors, and in this regard, all the cameras did very well, as you can tell by comparing their respective images to that shot with a Kodak still-image camera under the same lighting conditions. Though the FX1 initially looks a bit muted, it appears that its colors most closely match the Kodak image. From there, it gets too close to call. 

After comparing accuracy, look at the smoothness of each color in the image, and the crispness of the edges between the colors and the black spaces in between. Oftentimes, when comparing single-CCD camcorders, plates within the image will be marred by artifacts, and colors will bleed onto the black. Here, all images look very similar to me in both regards, so I'll award one first-place score (four points) and three second-place scores (three points). 

Let's quickly look at the results of the ballet images, the only real-world scenario where I used all three new cameras. Before chatting about the images, you should know that when shooting ballet, it's critical to get both the hands and the feet in the picture, which means shooting from way back.  

For this reason, the face that you see (that's no ballerina, that's my wife!) isn't as crisp as other images discussed later in the chapter. Still, you'll notice that there's no great disparity in detail between the cameras, though the XL2 face is framed slightly larger and consequently appears a bit more crisp. The only major difference I see is that the Panasonic and Sony images are slightly yellow, which I'm guessing relates to faulty white balancing which I'll attempt to fix in post. Pray, don't tell the wife.

Low-light Performance
Next up is low-light performance, as in how well the camcorder performs under shooting conditions where the available lighting is less than ideal. By way of background, most camcorders have a "lux" rating that measures the minimum lighting necessary to produce an acceptable image. Note that acceptable is not optimal; for example, the XL2 has a minimum lux rating of .7, while the recommended rating is 100 lux. Just for the record, the DVC60 has a lux rating of 4, the FX1 a lux rating of 3, and the VX2000 has a lux rating of 2.

Generally, lux rating is a good indicator of low-light performance, but it isn't definitive. For years, I've tested by shooting an image of Lance Armstrong that sits conveniently beneath my desk. To standardize the results, first I try shooting in fully automatic mode, which generally produces optimal results. Then I'll go manual, setting the F stop as low as possible, and boosting gain as high as possible, but never going below 1/60 for shutter speed, since slower than that makes the picture unusable.

All four cameras performed very well here, though I'll give the edge to the VX2000 over the XL2, primarily due to the sharpness of the cords and text on the router. I'll rate the DVC60 as superior to the FX1, though the color of the jersey was slightly distorted. All images were remarkably noise-free given the gain levels used to boost the dark image. This test is noted in the table as Low-light I.

Let's look at a real-world image that combines several factors, including low-light performance, color accuracy, and sharpness. I shot these in full automatic mode, with no white balancing or adjustments of any kind. The only light in the room is daylight streaming in from the windows. The video cameras are presented up top, with a still image shot of my daughter's pajama shirt and the couch taken with a Kodak still-image camera on the bottom.

Let's start with the most obvious: that the Panasonic's automatic brightness adjustments didn't produce as bright an image, so the DVC60 places last in this fully automatic test. Next, the XL2 produced a slightly purplish pajama top (as did the DVC60) when the true color was blue, as accurately reflected by the FX1 and VX2000. In addition, the XL2 produced a slightly softer facial expression, perhaps reflecting the lack of resolution seen in the EIA Chart tests. This drops the XL2 to third position.

Though it's subtle, the hair color in the FX1's shot is closer to my daughter's real hair, so it wins by a strand over the VX2000. Though the FX1 doesn't have the sheer lux rating enjoyed by the VX2000, in decent light, it sure does make the best of what it's got. This test is noted in the table as Low-light II.

Chromaker Tests
One of the most challenging tasks for any camera is shooting video for chromakeying, since any noise in the signal will prevent a clean overlay. In this group test, we shot all four cameras simultaneously, and overlayed each video over the same background in Adobe Premiere Pro.

The FX1 produced the best result, with very clean lines on the shoulders and neck, and relatively few artifacts in the hair. The XL2 trailed only slightly, with faint greenish fringes on several edges, but still outperformed the Panasonic, which showed fairly consistent fringes of green on the neck and face. The VX2000 placed last, with slightly more jagged edges and similar touches of green. For the record, the results were much closer when watching the videos play in real time, even side by side. Use a slightly more powerful tool, like the plug-ins from After Effects, and you can produce an excellent green screen with any of these cameras.

On-Board Microphones
Since we shot this chromakey sequence simultaneously, it was a great opportunity to test the quality of the cameras' on-board microphones. Sure, most of the time you'll use an external microphone—probably a shotgun or lav, depending on the nature of the shoot—but it's definitely good to know what you have on board, just in case.

To assess comparative quality, we exported the WAV file produced by each camera, imported it into Sony Pictures' Sound Forge to view the comparative waveform. The results didn't look good for the FX1; the on-camera microphone produces a disquietingly weak signal compared to the other cameras. T

he other waveforms look fairly close, but when you actually listen to the files, you can hear significant differences in the results: the XL2 wins hands down. Though the waveform doesn't document the sound as being quite as loud as the DVC60, the audio exhibits the least background whine and the voice sounds crisp and clear. I also experienced great results with audio on the Veteran's Day shoot as well as the Tae Kwon Do demonstration.

In our studio tests, both the Panasonic and the VX2000 exhibit a very similar background whine, though the Panasonic sounds slightly muffled, while the VX2000 sounds faintly metallic. They'll split second- and third-place votes, while the FX1 takes fourth. I'll talk a bit more about the audio features of each camera in the mini-reviews below. We retested all camcorders using manual gain control with the FX1 and auto with the others and the FX1 produced comparable volume. In the retest, the XL2 produced less noise and a higher-quality signal than the FX1, while the slight metallic whine of the VX2000 dropped it into last place. The Panasonic still sounded a touch muffled but produced less noise than the FX1, which produced more accurate sound but slightly more noise. For this reason, we rated the FX1 and the Panasonic a tie in our retests.

The Lens
All cameras perform well outdoors, where the light is plentiful and the white balancing fairly simple, so we don't have specific outdoor comparisons. However, we did want to see how each lens performed at extreme optical-zoom ratios. When my wife isn't teaching aspiring prima ballerinas, she performs surgery at a hospital about a mile from our house as the crow flies. So we zoomed in on the hospital from our backyard.

All the images are quite clear, but the XL2 and DVC60 are clearly bigger. That's because the XL2 has a 20X zoom, and the DCV60 a 16X zoom compared to a pair of 12X zooms for both Sonys (all optical). Also helping the long-range fidelity for the Canon and Panasonic is the fact that both cameras are shoulder-mounted and have excellent optical image stabilization.

In the Veteran's Day shoot, in which I shot with the XL2 exclusively, I used the zoom to great effect. I set up in the back of the church, which was over 100 feet from the front of the church. Even from this distance, I was able to zoom in and catch my daughters in the front row with mom-pleasing fidelity. This is one great piece of glass. Now let's spend a few minutes on each camera's controls and usability, which I'll score separately.

Canon XL2
The XL2 is the largest and heaviest camera in the review, and you'll have to take it apart to store it between shoots, so it's not the best for impromptu shooting. Once you've got it rolling, you'll find a lot to like, including the lens and shoulder mount discussed above. During the Veteran's Day shoot, I went off-tripod for many close-up shots, and the shoulder mount—combined with great optical image stabilization—produced very stable results. On the other hand, the XL2 was often the worst performer in automatic settings, as experienced throughout our tests, particularly when shooting under incandescent lights.

If you buy this camera, spend time up front getting to know how to set and store white-balance settings for your most common locations. Note that the XL2 doesn't have a dedicated LCD panel like the other camcorders tested here; instead, you can unlatch the viewfinder and see that panel directly. This worked well in most tests, including when shooting in directly sunlight, but you can't completely turn the panel around so that the subject can see it. For example, when I was shooting my wife's ballet class, the FX1 was on a tripod shooting the big-picture background while I operated the XL2 manually for close-ups. To help keep my wife within the extremes of the FX1 shooting area, I turned the 3.5" LCD panel around so she could gauge her position. This isn't possible with the XL2; the screen simply won't swing that far around. If you're shooting a reporter in the field who likes to look at herself while talking, this could mean that you have to carry a separate monitor around. For these reasons, the XL2 rates second behind the FX1 in terms of usability.

On the other hand, in terms of controlling your camera functions, the XL2 is a dream. Virtually all relevant controls are available on the camera body, including adjustments for aspect ratio, shooting mode (60i, 30fps, 24fps), and image stabilization, which are buried on some other cameras. The paradigm for adjusting parameters like gain and white balance while shooting is automatic mode is quite clear, which is great considering how often you'll have to go manual. Audio capabilities are excellent, with dual XLR connectors on the back with separate volume controls for each channel and even level controls for the headphone.

During our tests, however, we ran into one potentially serious hiccup of which any potential XL2 purchaser should at least be aware. Specifically, you control the on-board neutral density filters via a ring on the lens. When fully engaged, a status signal displays in the LCD panel. However, when you're turning off the ND filter, the ring can get stuck between the on and off position. The filter stays engaged but the status signal goes out. This caused us to lose a test or two before we figured out why the XL2 suddenly became our worst performer in low-light conditions. This probe aside, however, working with the XL2 is a dream.

Sony FX1
The FX1 is the first camera that made my wife say "wow," both in terms of its sleek black lines and the color and accuracy produced by its 250K 3.5" LCD panel on top, which worked very well even in direct sunlight. It's much smaller than the XL2, so you can carry it with you much more easily, and it works better in automatic modes, though I would definitely white balance manually when shooting under incandescent lights.

Overall, however, in terms of portability and ease of use, this camera is number one. On-board controls are greatly expanded, including separate white balance and three position gain presets, though I really preferred the infinite gain settings available on the VX2000. However, the iris adjustment knob on the bottom left is a great improvement, as are the three custom setting switches. As with the VX2000, you can manually set all exposure and other controls and then freeze the settings, which is very useful.

Other nice touches include an "expanded focus" button, which zooms into the picture by 4x to assist your manual zoom, and a button you can push to have the camera auto focus while in you're in manual focus mode, which is great way to quickly get focused and then take over manually. However, you have to go inside the camera's menu structure to turn on image stabilization, which is a pain when taking the camera off of a tripod for more candid shots. What's more, its accessory shoe is not powered, so you have to power all lights and microphones separately.

The great downside of the FX1—which it shares with the VX2000—is audio connectivity and control. You don't get XLR connectors on the body, so you'll need to buy a converter box from BeachTek, and you can't control left and right volume separately; there's one dial for both. The FX1 is the only camera of the three new models that doesn't display a volume meter when shooting with automatic gain control engaged; the meter only shows up in manual mode, which is silly. Overall, this pushes the camera behind the XL2 when it comes to on-board controls.

Panasonic AG-DVC60
Panasonic knows how to build a sexy camera, as we saw with the AG-DVX100, but they assigned a different set of designers to the DVC60, which is big, bulky, Spartan, and spare. We liked the shoulder mount, which helped stabilize the 16X zoom, but found color issues in most shoots in automatic mode, and the smallish LCD doesn't work well in direct sunlight, which made usability an issue.

Controls were the most confusing of the tested group. For example, where all other cameras have two or three position switches for manual focus on and off, the DVC60 has one on the bottom that says "focus." Shutter and iris share a single ring, which also gets confusing, and you can't access the menu system when recording—a problem, for example, if you want to adjust headphone volume.

On a positive note, the camera does offer XLR connectors on the back panel, with straightforward controls for switching between the front and back audio sources and controlling both audio channels. When comparing the camera's price to that of the VX2100 or FX1, remember that you'll have to spend $200-300 for XLR connectivity for both Sony cameras.

Summary and Conclusion
The XL2 is our overall winner, edging out the FX1 by one point, though if you take on-board audio out of the picture, the FX1 wins, and costs about $1,500 less. That said, the XL2 has many unique strengths, including its uniquely high powered lens, great onboard audio and awesome onboard configurability; if you like to tinker and try wildly different settings and combinations of settings, this is your camera.

However, it's a camera that requires lots of care and feeding, from a three to four minute setup to get all the pieces together, to de rigueur white balancing to ensure proper optimal color. It's a great camera for working in a studio or for carefully planned travel, but not the best camcorder for an investigative journalist.

As a DV camera, the FX1 is obviously a great performer, and that doesn't even factor in the ability to produce HDV. Excellent performance in most automatic modes (watch for browning under incandescent lights, though) makes it a great point and shoot camcorder, even for beginners. It's the smallest of the three new camcorders, and feels rugged enough to throw in a bag and carry around. When you buy the camera, however, be sure to buy a BeachTek DXA-8 ($369 at B&H) or similar unit for XLR connectivity and a good shotgun microphone. Or, you can wait until February, when Sony is expected to ship the HVR-Z1U, with XRL connectivity and 24p mode for $4900.

The DVC60 was our least expensive camcorder, especially if you consider the extra equipment you'll need on the audio side to use the Sony cameras professionally. Like the XL2, it's too big to throw in a bag and go, and performance in automatic mode wasn't stellar. In addition, it's obviously not optimal for those producing in 16:9. Probably the most appropriate use for this camcorder is in a studio or office operating under fixed, controlled environments where you can optimize the settings and all quality adjustments.

Finally, is it time to throw away your VX2000-class camera? If 16:9 performance and top-notch chromakeying are critical, you probably need to upgrade. But for most other uses, that old gray mare stacks up fairly well.