How do digital cameras fare? We find out by taking pictures
How do digital cameras fare? We find out by taking pictures
Lab tests 10 cameras for color fidelity, resolution and more
- By Carlos A. Soto
- Aug 16, 2001
Digital cameras, whether professional-grade or point-and-shoot, have surpassed conventional film cameras in functionality. At the high end, digital is catching up with film in quality and price.
But professional-grade cameras still have significant drawbacks that have frustrated users from the beginning: short battery life, slow image processing, and confusing software and hardware interfaces.
To judge the overall quality of 10 leading digital cameras, the GCN Lab magnified and scrutinized their output on three categories of subjects in Adobe Photoshop 6.0. The lab staff also printed out all the images for close study.
As a further measure of quality, I solicited GCN graphic artist Michael J. Bechetti's perspective on which pictures would need the least Photoshop color correction for publication.
Finally, I looked at the convenience of each camera'its media choices and battery life, how easily it connected to a PC and how simple it was to operate.
The first test pictures were of a colorful red hibiscus flower on a sunny patio. Although this image presented no challenge at all to the professional cameras, which showed identical output, it did stump the point-and-shoot cameras.
A second, tougher test was accurately recording skin texture and color under tricky lighting. I tested point-and-shoot cameras by photographing Laura Currey, a GCN marketing specialist, partly in shade with the light at an angle and using fill flash.
I tested each professional camera by photographing Currey in complete shade with skylight behind the camera.
A third battery of so-called neutral color tests showed each camera's color fidelity in a white, dark or gray portion of each picture.
Knowing the percentages of cyan, magenta and yellow in a neutral area of an image helps a graphic artist tune the color using graphics software with the necessary controls.
In general, for proper color rendition, an image should contain equal amounts of yellow and magenta, with cyan about 4 percent more. Note that the CYM percentages for the cameras don't add up to 100 percent because each color is measured separately (see charts, pages 52
The lab's subject for the neutral color test was challenging: a bowl of fruit against a white backdrop under hard-to-render fluorescent lighting. For each camera, I took readings of three areas of the fruit image and averaged them.
It might seem logical that the higher a camera's resolution'expressed in millions of pixels, or megapixels'the better the image. But that wasn't always the case.
Some high-resolution cameras in the review produced lower-quality pictures than others with fewer megapixels.
Why? Light sensitivity is a big factor. Digital cameras have a poor track record under low lighting because the light-sensitive surface, either a charge-coupled device or a complementary metal-oxide semiconductor chip'cannot process an image with too little illumination.
Recent advances have made CCD and CMOS chips somewhat more light-sensitive, however.Fisheye to wide angle
Professional-grade cameras are distinguished from the point-and-shoot variety mainly by higher resolution.
They start at more than 3 megapixals and can emulate up to 5 megapixels.
They also have interchangeable lenses with aperture controls so you can control depth of field. The aperture setting is half of what determines how much light enters the lens.
Professional cameras also let you control the shutter speed, the other variable that affects the amount of light admitted into the camera.
Professional cameras also feature extensive, if generally confusing, built-in software controls for sorting, sizing and manipulating images.
Keep in mind that the large files they produce can swamp a PC used to process them unless there's plenty of RAM'say, in the 256M to 512M range.
Priced without a lens at $2,400, the FujiFilm FinePix S1 Pro
was by far the best buy in the professional category and my Reviewer's Choice for professional cameras.
Although I would have liked to test a high-end Kodak or Nikon camera, neither company chose to submit professional cameras for the review.
The S1 Pro would have given them a run for their money.
The 3.54-megapixel, single-lens-reflex FinePix S1 Pro is compatible with most Nikon F-mount lenses and can accept either CompactFlash or SmartMedia memory add-ons.Makes sense
The interface is logical and simple, and the camera's weight and shape are satisfying.
The only negative I observed was that images tended toward yellow in poor lighting.
The FinePix scored highest in the neutral color test and the skin tone test. In the latter, it far surpassed the other two professional cameras for rich color and texture.
Much of the Fuji's ability to capture high-quality images could be attributed to its 3.54-megapixel CCD, capable of emulating up to 3,024- by 2,016-pixel resolution.
It was tough to tell the Fuji S1's pictures from those of the Olympus Camedia
. The E-10 had 4 megapixels and 2,240-by-1,680 resolution. Its depth of field was slightly better, whereas the Fuji produced slightly crisper and more textured fruit images.
These merits and a relatively low $1,300 price tag earned the Olympus Camedia
E-10 the Bang for the Buck designation in the professional category.
Conversely, it was easy to tell the Fuji pictures apart from the Olympus ones in the skin tone test.
Although the Olympus came in second, its bright and shadowed contrast levels facial segmentation distorted skin tone and texture. As a result, I found the Fuji's output truer to life.
The E-10 is not a single-lens-reflex camera, and it could not accept other lenses. That makes it less versatile than the Fuji, which can do short-distance as well as telephoto work.
Furthermore, the Olympus controls were nowhere near as simple or logically constructed as those of the Fuji. That made it harder to change settings such as aperture size or shutter speed.
Like the Fuji, the Olympus could accept either CompactFlash or SmartMedia.
The Minolta Dimage RD3000
accepted only CompactFlash and required a SCSI connection to a PC. A fast SCSI connection seemed pointless, however, because the in-camera processing rate was so slow.
The Dimage itself was enormous at 2 pounds, 13 ounces. It also chewed up batteries faster than any of the other cameras.
It went through about eight AA batteries per day taking an average of 15 superfine-resolution pictures and processing them on the camera for about 10 minutes each. That amounted to about eight batteries for 150 minutes of use, which was twice the consumption rate of the Fuji or the Olympus and four times that of the point-and-shoot cameras.
The Minolta processed images so slowly, especially in superfine mode, that it took on average eight to 10 minutes to see an image on the camera's LCD.
All the other cameras in the review took less than a minute to generate an image on the LCD even at highest resolution.Texture, perspective
Now let's move to the point-and-shoot contenders.
Overall, the $300 Toshiba PDR-M61
created the most textured and color-perfect images in this category. It had the best depth of field in good lighting. The PDR-M61 also came in first in the skin tone test.
The 2.3-megapixel PDR-M61 not only deserved the Reviewer's Choice designation for its accurate emulation of skin tones, even in low lighting, but its $300 price also merited a Bang for the Buck honor.
It did feel flimsy and lacked crosshairs for accurate viewfinder framing, however. The zoom control was in an awkward spot where it could be hit accidentally, and the viewfinder was a bit obstructed by the lens.
The PDR-M61 used SmartMedia memory only. I would have preferred to see it pop up as plug-and-play removable storage under Microsoft Windows 2000. Nevertheless, compared with other cameras in the review, it was easy to use.
The $500 Toshiba PDR-M65
was nearly identical in format and color capability, but with a resolution of 3.3 megapixels. Thanks to its extra pixels, the camera did a better job of rendering texture but not as well rendering the red hibiscus.
Like its brother, the PDR-M65 required software to load images and accepted only SmartMedia.
For the extra $200, I would have liked to see an additional port for CompactFlash, which has advantages over SmartMedia, such as integration with firmware to speed image processing.
Another Reviewer's Choice, the $500, 2.2-megapixel Kodak DX3600
, proved that a sub-3-megapixel camera could indeed take great pictures. It came with 8M of internal memory and had an expansion slot for CompactFlash cards.
But it wasn't perfect.
Glare from the sun reflecting off windows of a building behind our human subject produced a magenta skin tint. No other camera in the review was fooled by the sun's reflection.
Photographing the red hibiscus in bright sunlight, the DX3600's lower resolution manifested itself in colors that bled together. Nevertheless, I was impressed by its superb optical sharpness and vibrant color.
Kodak's image software could use more sophistication instead of playful icons and childlike text, which made the camera seem toylike. The software was complex and took a long time'about five minutes'to load.
All the images automatically went into a folder, leaving the user with little control over the data.
On the plus side, the Kodak software did indicate battery life, which is a guessing game with most digital cameras.
The second-best scorer on the neutral color test was the FujiFilm FinePix 2300
, which incidentally scored lowest on the flower test.
This $200, easy-to-use, 2.1-megapixel camera accepted only SmartMedia. It was the easiest in the review to connect to a PC.
Under Win 2000, it plugged into the Universal Serial Bus port and immediately popped up on-screen as a removable storage device with device drivers.
Like the Toshiba PDR-M61, the Fuji FinePix 2300 merited a Bang for the Buck designation. It was somewhat awkward to use, however, because the LCD screen tended to wind up pressed against the user's nose.
The $450 Olympus C-2040 Zoom
made indoor shots too dark and outdoor shots too bright. It felt like a professional camera, but at $450 for only 2.1 megapixels it was priced too high for its output quality. It too uses only SmartMedia cards.
It was, however, as easy to connect to a PC as the Fuji FinePix 2300.
The $800, 3.3-megapixel Epson Photo PC 3100Z
did well in the neutral color tests, but the pictures were rather dark. I noticed that the bottom left of its images looked blurred, as if somthing was misaligned.
The Epson was the most expensive camera in the point-and-shoot category, and it felt that way. It had the feel and heft of a professional camera. Quick on the draw
It took consistently good pictures and, like the Kodak and the Minolta, used CompactFlash media only. Unlike the Kodak, the Epson was easy and fast to install with a PC. Win 2000 saw it as a removable drive.
The Minolta Dimage 2330
reproduced the fruit bowl setup accurately, if a bit on the dark side.
But the skin tone test separated the Minolta from the rest of the point-and-shoot devices. Flesh took on a distinctly orange tint.
The Minolta uses CompactFlash media and connected easily to a PC as a removable drive. Like the Toshiba cameras but unlike the other cameras, it had no crosshairs and showed an obstructed view through the viewfinder.