Digital Cameras - Sony Mavica FD91Test Images
(Original test posting: 3/27/2000)
|We've begun including links in our reviews to a Thumber-generated
index page for our test shots. The Thumber data includes a host of
information on the images, including shutter speed, ISO setting, compression
setting, etc. Rather than clutter the page below with *all* that detail,
we're posting the Thumber index so only those interested in the information
need wade through it! ;)
Note on Exposure Info for the Sony MVC-FD91 test images: We normally include exposure information for all our test shots, extracted from the JPEG file headers by Max Lyon's excellent "Thumber" program. In the case of the FD91 (and some other Sony cameras) though, the required information apparently isn't stored in the file headers in the first place. Thus, none of the images below have any of the usual shutter speed/aperture information associated with them. Sorry! :-(
This is a tough shot for many digicams, due to the extreme tonal range (which is why we set it up this way!). The trick is to hold highlight and shadow detail without producing a "flat" picture with muddy colors. We shot this image using the manual white balance setting because it gave us the best tonal values overall. Overall color balance on the MVC-FD91 is good, the very difficult blues of the flowers and the model's pants look good although slightly dark, and with just a hint of the common purple/blue problem (many digicams have a tendency to reproduce these with a distinct purplish hue). We did notice some blue tints in the shadow areas, as well as a moderate amount of noise but shadow detail still looks pretty good. Overall detail is a bit soft relative to higher-resolution cameras (judging by the small leaves next to the model's shirt and her hair), but surprisingly good when compared to other sub-megapixel models. We didn't require any exposure compensation adjustment on this shot for our main image, which means that the MVC-FD91's exposure system did a good enough job on its own. Just the slightest adjustment blew the highlights in the shirt and the white flowers. The table below shows the results of a range of exposure settings from zero to +1.5 EV.
Exposure Compensation Settings:
The MVC-FD91 does a nice job with this "portrait" shot, thanks to its 14x optical zoom lens. (Shorter focal length lenses tend to distort facial features in close-up shots like this. The availability of longer focal lengths is a key feature if you're going to be shooting close-up people shots). As is typical with this closer shot, our main shot required no exposure compensation at all, even the slightest adjustment made the highlight areas on the shirt and flowers too hot. Sharpness and detail seem a little crisper in both the highlight and shadow areas, with the shadows producing about the same level of noise as in the other portrait. The table below shows the results of a range of exposure settings from zero to +1.5 EV.
Exposure Compensation Settings:
portrait, flash: (102k) This shot is
always tricky because of the potential differences between the color balance
of the flash and the bright room lighting. Many cameras produce odd bluish
highlights here, which the MVC-FD91 did somewhat as well. The default
flash exposure setting (102k) did a good job of
illuminating both the foreground and background without overexposing the
highlights, although the overall exposure is a bit dark. We increased the
exposure compensation to +0.5, snapping this
(101k) only slightly lighter image, which we chose
as our main shot for this category. We also tried using the flash with the
camera's manual mode (sorry, didn't record exposure specifics, counting
on the file header info to provide it - discovered too late about the missing
exposure info!), capturing this (102k)
slightly lighter image.
portrait, no flash: (152k)
This shot is a very tough test of a camera's white balance capabilities, thanks to the strong yellowish cast of the household incandescent lighting it's shot under. The MVC-FD91 did a nice job with this difficult light source, with the best results achieved in manual white balance mode (152k), without too much warmth or odd, purplish hues. The incandescent (152k) setting left more of a warm, yellowish cast from the lights than we'd like (although fairly typical of many other digicams) and automatic (165k) came out a just a little bit pink. The camera's exposure system did a good job here as well, causing us to use only a +0.5 exposure compensation adjustment for the main image.
The table below shows the results of various exposure compensation settings from zero to +1.5 EV.
Exposure Compensation Settings:
NOTE that this is the "new" house shot, a much higher-resolution poster than we first used in our tests. To compare the image of the MVC-FD91 with previously tested cameras, here's a shot of the original house poster (275k).
The MVC-FD91 did the best job on the manual white balance (113k) setting, although the automatic setting (113k) came in extremely close with just a hint more warmth. We shot a sample in the daylight (114k) setting, which came out much warmer, despite the carefully daylight-balanced lighting we use. Resolution and detail look just slightly soft in the top tree limbs and the bricks on the house, and the roof shingles are a little noisier than average (actually, the entire image shows a moderate amount of noise). In-camera sharpening is evidenced by a small halo on the dark and light edges, but not too bad on the whole. Despite the soft resolution and the amount of noise, the MVC-FD91 does a nice job of capturing the image. The table below shows our standard resolution series.
This image is shot at infinity to test far-field lens performance. NOTE that this image cannot be directly compared to the other "house" shot, which is a poster, shot in the studio. The rendering of detail in the poster will be very different than in this shot, and color values (and even the presence or absence of leaves on the trees!) will vary in this subject as the seasons progress. In general though, you can evaluate detail in the bricks, shingles, and window detail, and in the tree branches against the sky. Compression artifacts are most likely to show in the trim along the edge of the roof, in the bricks, or in the relatively "flat" areas in the windows.
The MVC-FD91 did a pretty good job on this shot as well. Detail and sharpness still look a little soft, especially in the tree limbs, but the image is nice overall, and compares favorably with other sub-megapixel cameras. The telltale roof shingles do show a moderate amount of noise though. The MVC-FD91 did pick up the usually difficult highlight details in the white painted area of the bay window (a tough area for many digicams since the area was painted with a bright white paint that makes tonal handling very difficult). The in-camera sharpening does produce a heavier black outline between the white trim and the darker shingles, and makes the tree limbs a bit heavier-looking than they'd be otherwise. We chose auto white balance on this one, which gave us the best color balance overall, although there was little difference between auto and daylight settings. The table below shows the full resolution series
||Lens Zoom Range (new)
We've received a number of requests from readers to take shots showing the lens focal length range of those cameras with zoom lenses. Thus, we're happy to present you here with the following series of shots, showing the field of view with respectively, the lens at full wide-angle and the lens at full 14x telephoto. Unfortunately, these images were shot at the MVC-FD91's larger file size, which will add to the download times. Overall, the MVC-FD91 performs very nicely in this category and we noticed that sharpness and noise seemed to improve with the full, 14x optical zoom enabled. - If you intend any telephoto shooting at all, it's hard to overstate the impact of the 14x lens on the FD91! The level of detail captured by the FD91 far exceeds that we've achieved with any other digicam tested to date! (March, 2000).
We shot samples of this using auto,(86k) daylight (85k) and manual (86k) white balance options, leading us to choose the daylight setting for our main shot (204k) due to the much cooler casts of the auto and manual settings. Daylight white balance gave us the best skin tones and overall color saturation, particularly noticeable in the blue robe which is just about right. (Some cameras are tricked by the extensive amount of blue in the image. Additionally, choosing the correct white balance on this shot is a bit of a toss up because of the predominant blues in the image, so we look at the skin tones to decide). The MVC-FD91 also did a nice job with resolution relative to other ~1 megapixel cameras, reproducing good detail in the bird wings and the tiny silver threads on the Oriental models robe. The table below carries links to our standard resolution/quality series.
||Macro Shot (174k)
The MVC-FD91 performs very well in the macro category, with a minimum area of only 1.48 x 1.11 inches (37.69 x 28.37 mm). Good detail, sharpness and color, except for the left side which appears a bit blurry. The very top of the brooch and the coin seem a bit soft as well, possibly due to them being raised up a little out of the focal range. Unfortunately, the MVC-FD91's flash was blocked by the lens and therefore was ineffective in this shot.
Test Target (138k)
The MVC-FD91 produced very nice color and tonal results on this test with just a slight weakness in the magenta, yellow and cyan color blocks. The green block also looks slightly weak, but the colors look very good overall. The MVC-FD91 did a great job of distinguishing between the red and magenta color blocks on the middle, horizontal color chart as well (many digicams have trouble here and try to blend the two together, so we're always happy to see this demarcation). The subtle tonal variations in the Q60 chart also came out nicely, with the camera able to distinguish color blocks up to the "B" range on the pastel end. The usually difficult shadow area on the briquettes also turned out nicely, with a good bit of detail. The noise level looks a little high overall, but not too bad. We shot with the daylight (62k), auto (61k) and manual (60k) white balance settings. Daylight produced a very warm cast while auto and manual looked nearly the same. In the end, we chose the auto setting for our main shot (138k) as the whites and overall tones looked the best in that setting. The table below shows the usual range of resolution/quality settings.
Well, the MVC-FD91 is an exceptionally versatile camera, but it's probably not going to be your first choice if low-light shooting is a top priority: Hampered by a maximum exposure time of 1/60 of a second (?!), the FD91 is really only barely usable at the upper end of where we start testing low-light performance. At a light level of 8 footcandles (88 lux), the image delivered by the FD91 is bright enough, but a large amount of image noise combines with the JPEG compression to produce obtrusive artifacts across the image. Bottom line: Definitely not the camera to use for night-time photos. :-( The table below contains links to sample files shot at light levels from 8 down to 2 footcandles.
||Flash Range Test (New)
(This test was added in August 1999, so cameras tested before that time won't have comparison pictures available. As we go forward though, all the new models will have similar tests available). Sony rates the MVC-FD91's flash effectiveness out to a maximum of 8.3 feet (2.5m) in normal, wide angle mode with the flash set at normal intensity. We found this to be pretty accurate with a brightness loss starting immediately at eight feet and images getting progressively darker on out to 14 feet. The table below shows results obtained at a range of distances from eight to 14 feet.
(WG-18) Resolution Test
While not measuring up to the latest 2+ megapixel cameras, the MVC-FD91's resolution is quite good for an 0.8 unit, competing very well with cameras up to about a megapixel of resolution. We called the resolution here as being about 575 lines per picture height in both vertical and horizontal directions. (We rated some 0.8 megapixel units a bit higher in the past, but have been more conservative with these numbers in recent testing. Relative to previously-tested digicams in its resolution range, the FD-91 does quite well.) There's sum blurring of the image in the lower right-hand corner, with the lens set to its widest-angle zoom setting. This effect isn't present at all at the telephoto setting, but we don't know where along the lens' zoom range the artifact goes away. Telephoto resolution is about the same as wide-angle, albeit with slightly less edge sharpness. (Some resolution falloff at telephoto settings of zoom lenses is common. The FD91's lens probably does a bit better in this respect than most.) The tables below contain links to all our test shots, taken at both wide-angle and telephoto settings.
Wide Angle Resolution/Quality series
Telephoto Resolution/Quality series
||Viewfinder Accuracy/Flash Uniformity
We found the MVC-FD91's LCD monitor to be very loose, showing about 86
percent of the final image area in wide angle
(50k) and about 85 percent in telephoto
(59k). To give you an idea of what we're looking
for, we generally like to see LCD monitors as close to 100 percent accuracy
as possible. Using the optical viewfinder, we found exactly the same measurements
as the LCD monitor in wide angle (51k)
and telephoto (64k). Not
a surprise, since the MVC-FD91's optical viewfinder is essentially a smaller
version of its LCD monitor display. We did notice better accuracy on the
MVC-FD91's smaller resolution (640 x 480) size, catching about 92 percent
of the final image area on the wide angle end with both the LCD
(23k) and optical viewfinder
(24k). We shot as close to the full 14x optical
telephoto as we could get in our studio, snapping this
Back to Sony Mavica FD-91 Review
Jump to Comparometer(tm) to compare with other cameras
Or, up to Imaging Resource Cameras Page.