Canon EOS 20DBy: Shawn Barnett and Dave EtchellsSlightly smaller and lighter upgrade brings greater speed and ease of use along with higher res and lower image noise. <<Video, Power, Software :(Previous) | (Next): Reference: Datasheet>> Page 12:Test Results & ConclusionReview First Posted: 08/19/2004, Update: 11/19/2004 |
Test Results
In keeping with my standard test policy, the comments given here
summarize only my key findings. For a full
commentary on each of the test images, see the EOS 20D's "pictures"
page.
Not sure which camera to buy? Let your eyes be the ultimate judge! Visit our Comparometer(tm) to compare images from the EOS 20D with those from other cameras you may be considering. The proof is in the pictures, so let your own eyes decide which you like best!
As with all Imaging Resource product tests, I encourage you to let your own eyes be the judge of how well the camera performed. Explore the images on the pictures page, to see how the EOS 20D's images compare to other cameras you may be considering.
- Color: Excellent color, hue-accurate, with good saturation.
(These comments relate to the 20D's default color mode.) The EOS-20D's
overall color was very good, generally hue-accurate, with appropriate saturation.
Like many digicams, including most d-SLRs, the 20D oversaturates strong
reds a fair bit, but other colors are for the most part appropriately saturated.
Bright blues and greens are slightly oversaturated, and a range from yellow-green
through yellow to orange are slightly undersaturated. That's the technical
description at least, but it turns out that photos from the camera look
much more accurate and natural than a strictly numeric analysis of its color
characteristics would indicate. I suspect that this is because the 20D boosts
the brightness in the same yellow-green to orange range through which it
drops the saturation. The net result is very natural-looking and pleasing
color rendition across the spectrum. All in all, excellent color. (For
those interested, the illustration above from Imatest shows the color error
for each swatch on the MacBeth chart. If you mouse over the image, you'll
see the same plot for the EOS-10D, letting you compare the color response
of the two cameras. - They're pretty similar, as you can see. See the 20D's
Imatest page for more details.)
- Exposure: An accurate exposure system. EOS-20D's
exposure system accurately exposed most of my test targets, both indoors
and out, generally requiring less exposure compensation on tough subjects
than did other cameras I've tested. The "Sunlit" Portrait required
only a +0.3 EV exposure adjustment to brighten the shot, resulting in bright
midtones with good detail. The camera's contrast adjustment option worked
very well, allowing the 20D to do a much better than average job of holding
onto both highlight and shadow detail under the deliberately awful lighting
of the Sunlit Portrait shot. The one hiccup I encountered (and a minor one
at that) was the slight overexposure it delivered at its default exposure
setting on the Far-Field test. The exposure error there was very slight
though, I'd estimate it to be about 1/3 EV. All in all, very accurate exposure,
and excellent dynamic range, particularly when the contrast adjustment was
dialed down.
- Resolution/Sharpness: Very high resolution, 1,500
- 1,650 lines of "strong detail." The EOS-20D performed very
well on the "laboratory" resolution test chart. It started showing
artifacts in the test patterns at resolutions as low as 1,200 lines per
picture height in the horizontal direction, and around 1,000 lines in the
vertical direction. I found "strong detail" out to about 1,650
lines/picture height horizontally, 1,500 lines vertically. (Some reviewers
will doubtless argue for higher numbers than these, but I tend to be more
conservative in how I "call" the resolution of cameras with this
target. My rule of thumb is to not count lines of resolution beyond the
point at which artifacts start to obscure the target lines, even though
one's eye may see signs of the target lines beyond that point. Here, the
target lines in the horizontally-oriented resolution hyperboloid (which
tests for resolution along a vertical axis) show considerable aliasing beginning
at about 1,500 lines, even though there's evidence of subject detail at
least as far as 1,600 lines. Consequently, that's where I called the vertical
resolution.) Using its MTF 50 criteria and correcting to a standard 1-pixel
sharpening radius, Imatest reported the 20D's resolution as 1513 lines/picture
height along the horizontal axis, and 1707 lines/picture height vertically,
for an average of 1610 lines. "Extinction" of the target patterns
didn't occur until right at 2,000 lines. Very impressive.
- Image Noise: Really excellent noise characteristics, especially
at high ISOs. Low image noise is one of the real standout features of
the EOS-20D, particularly at very high ISOs (1600 and 3200). The chart above
compares the 20D's image noise levels against those of other d-SLRs, but
as I've noted in the past, graphs like this tell only part of the story,
because they don't reveal what the image noise actually looks like.
In the case of the 20D, its noise pattern is particularly fine and tight-grained,
which makes it much less visually apparent and objectionable than that of
many cameras. At the same time, the 20D's noise-suppression technology does
an excellent job of holding noise in check, without losing subject detail
in areas of subtle contrast. It's easy to have low image noise, if you don't
mind giving up a lot of subtle subject detail, but the EOS-20D seems to
avoid this tradeoff almost entirely. Most impressive is the extent to which
Canon has managed to reduce image noise at ISO 1600 and 3200, relative to
every other camera on the market, including even their own 1D Mark II. At
ISO 1600, the 20D's image noise looks as good as or better than that of
many competing cameras shooting at ISO 800, and even its ISO 3200 images
are entirely usable for many applications. All in all, a really excellent
performance. (See the 20D's Imatest page
for more details.)
- Closeups: (No info here, as macro performance is entirely
dependent on lens selection.)
- Night Shots: Excellent low-light performance, with
bright exposures, good color, and very low image noise. Thanks to several
recent improvements in Canon's CMOS sensor technology (several of which
made their first appearance in the 1D Mark II), the EOS-20D's image noise
levels are just amazingly low, particularly for an 8-megapixel chip. As
a result, the 20D's low light performance is arguably the best I've yet
seen from a digital camera, at any price point. The EOS-20D produced clear,
clean images down to the 1/16 foot-candle (0.67 lux) limit of my test, with
good color at all six ISO settings. Color balance turned slightly warm as
the light level decreased, but the color shift was pretty minor. Because
image noise was so low to begin with (there were essentially no "hot
pixels" to be found anywhere), enabling or disabling the camera's Noise
Reduction system didn't make a lot of difference, at least at the light
levels that I cover in my tests. Even at ISO 3,200, image noise was relatively
low, and didn't interfere with color balance or detail.
- Viewfinder Accuracy: A pretty accurate (94%) optical
viewfinder. The EOS-20D's digital SLR design proved fairly accurate,
showing about 94 percent of the final frame. I prefer SLRs like the EOS-20D
to have as close to 100 percent accuracy as possible, so the EOS-20D fell
slightly short here, although it's important to note that most other d-SLRs
have viewfinders that are only 95% accurate as well. Typically, only very
high-end models like Canon's own EOS-1D Mark II and EOS-1Ds Mark II have
true 100% viewfinders.
- Optical Distortion: (Here again, no comments in this
category, as distortion will depend entirely on the lens in use.)
- Shutter Lag and Cycle Time: Excellent shutter response and cycle
times. Shutter lag in an SLR will obviously depend on the particular
lens in use, since different lenses can adjust focus more or less quickly.
Regardless of the lens we hung on it though, the 20D was very responsive
to the shutter. With the "kit" lens, it showed shutter lag ranging
from 0.15-0.16 second. In Manual focus mode, it showed an average lag of
0.114 second, and prefocused (by half-pressing and holding down the shutter
button before the exposure itself), it showed an average lag of only 0.080
second. Shot to shot cycle times were also very good, an average of just
0.48 seconds/shot in single-shot mode, with the buffer holding 31 large/fine
JPEG images, or 6 RAW or RAW+JPEG frames. In continuous mode, it can capture
up to 31 large/fine JPEG images or 6 RAW or RAW+JPEG ones at a rate of 4.80
frames/second, and clears the buffer of all 31 JPEG frames in just 19 seconds
with a Lexar 80x memory card, or all 6 RAW frames in just 12 seconds. (The
20D takes good advantage of fast memory cards, so a fast card is a good
investment with it.)
- Battery Life: Excellent battery life, although I can't measure it directly. Because it lacks an external power terminal, I couldn't perform my usual direct measurements of power consumption on the EOS-20D. I can attest to its excellent battery life though, as I should shoot literally for days without exhausting a fully-charged battery. Canon rates battery life at 1000 shots without the flash, or about 700 with 50% flash use, at 20 degrees C / 68 degrees F, which certainly seem like reasonable ratings given my personal experience with the camera.
Conclusion
Free Photo Lessons | |
|
Reader Comments! --> Visit our discussion forum for the Canon EOS 20D!
Follow Imaging Resource: