Quality Tests: Hard Focus

They say that beauty is in the eye of the beholder, and we would argue the same for the quality of DV camcorders. That's why we captured two scenes for our reviewers to behold—a talking-head scene and an action clip—and presented the resulting video clips in as close to an unadulterated DV format as possible, in this case rendered AVI.

The talking-head clip was a typical instructional or business presentation. In this static scene, a subject sat behind a desk giving an oral presentation. Recognizable key elements were inserted to aid evaluation. The high-action sequence was produced under the same studio lighting conditions. Various props and subjects animated the scene, taxing the cameras' auto-focus capabilities and their ability to capture fine details in motion.

We shot the two video clips in contributing photographer Thom O'Connor's studio at PC Magazine. We mounted the cameras side by side, approximately 15 feet from a neutral paper backdrop, and we ran them all simultaneously, so the circumstances would be identical.
To simulate good interior conditions for filming, we used tungsten lighting with a lux reading of approximately 375 and a 2,500-degree color temperature. We dimmed the lighting at the end of our talking-head clip to evaluate the auto-exposure response. Cameras with low lux ratings compensated best when filming under this kind of "distressed" condition (not to be confused with infrared settings for night shoots). We were impressed with the Sony DCR-VX2000's ability here.

We set all the cameras for auto-focus and auto-exposure. We selected each camera's highest-quality audio setting, either 12 kHz or 16 kHz. We also grabbed frames from the resulting video clips to isolate key elements for inspection.

Our test-bed viewing system included a 1-GHz Dell Dimension 4100 with 512MB of SDRAM and a 40GB Western Digital Caviar 400BB EIDE hard drive. We used an ATI Radeon 8500 graphics adapter. The jury viewed the clips on a 21-inch Sony GDM-F520 CRT, a high-end monitor that we calibrated and optimized with an X-Rite DTP92 CIE colorimeter and ColorShop 2.5 software. To judge audio quality, we used the Creative Labs Audigy Platinum sound card and Monsoon Multimedia MM2000 speakers. Pinnacle Micro provided a video-editing station with a Calisto video capture card and Pinnacle Studio 7.09.15 software.

We, the Jury

The jury viewed both clips and still captures to rate their quality. Each clip could get a maximum of 10 points based on such categories as clarity, dynamic range, color, noise level (or artifacts), and audio. The cameras that achieved the top three video quality rankings were those that had three CCDs instead of one. Our jurors rated these three units equally high on clarity and focus during our talking-head scene.

We looked for any visible signal noise and interference, deducting points for any artifacts. All jurors agreed that noise was easily detected in the clips from the Sharp Viewcam VL-NZ8U and the JVC GR-DVP3.

Stop and Go

We appraised each camera's snapshot capabilities using both an action shot and a still life. The action shot revealed the differences between interlaced and progressive-scan techniques. All the cameras stored these photos on separate removable media cards except for the Canon units, which stored them on tape (we had to extract shots from the videos through software for the latter). We shot all the photos indoors, under fluorescent D50 lighting conditions. The Sony DCR-TRV30 had 1.5-megapixel images, and it showed, topping the still-image category.

A Word or Two About Words

Unless you're standing right on top of your subject, you should get an external mic rather than relying on the built-in one. At the very least, you'll want a manual gain control and audiometers like those on the DCR-VX2000.

We used the built-in condenser mics on all cameras except the JVC model, which comes with an external unidirectional boom mic. Because of the distance from our subjects, all the units captured considerable ambient noise. Our jurors rated the audio simply as acceptable or unacceptable, and ambient or signal noise was a big factor. One sobering fact: In real-world use, a very sensitive omnidirectional microphone may pick up much unwanted ambient noise, depending upon a location's acoustical character. —Analysis written by Glenn Menin

Interlaced Vs. Progressive Scan

The NTSC television signal consists of 29.97 frames per second, with each frame comprising two fields: The first field contains odd horizontal lines, the second even ones. To create NTSC video, a camera actually shoots about 60 frames a second to enhance video smoothness. This type of video is called interlaced because of the way the two fields are combined to make a frame. Interlacing works well for video that's viewed from a distance, like television, but presents a problem for still images, especially when there is motion. That's because the two fields are 1/60-second apart, and a lot of motion can occur in that short time.

In our interlaced image, the camcorder combined two fields into a frame (left). The yo-yo has moved in that fraction of a second, and the two fields don't match up. No significant artifacting is visible in the still arm or hand.

By contrast, progressive-scan images are shot from top to bottom, which is how digital cameras capture images. DV cameras in progressive-scan mode actually shoot only 30 times a second, dividing each frame into the two fields required for NTSC compatibility. The progressive-scan image (right) is very sharp, because the two fields created from the same frame seamlessly merge back into that frame to create the still image. Progressive scanning was first introduced to create video that smoothly converted to film, which has 24 fps. It's also useful when you're converting to frame-based digital formats like MPEG-1 or streaming formats like Real, QuickTime, or Windows Media. Today, DV cameras strive to serve dual roles—for video and for digital stills—and progressive scanning provides much better still-image quality.

Capture the Moment

We captured some frames from the video clips shot by each camcorder in the roundup to show you the difference in each camcorder's ability to produce videos.
View the MiniDV Camcorder video clip slide show.

Copyright © 2002 Ziff Davis Media Inc. All Rights Reserved.
Originally appearing in PC Magazine.