Sony Cyber-shot DSC-F828Sony extends their high end to encompass 8(!) megapixels and a sharp 7x Zeiss zoom lens. - And Sony's new RGB+E sensor technology for more accurate color!
<<Video, Power, Software :(Previous) | (Next): Reference: Datasheet>>
Page 12:Test Results & ConclusionReview First Posted: 08/15/2003, Updated: 02/05/04
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 Sony DSC-F828's "pictures" page.
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 F828's images compare to other cameras you may be considering.
- Color: Very good color, excellent handling of highly saturated reds and yellows. Normal color a tad oversaturated for my tastes, "Real" color mode very natural-looking. The F828 handled my test lighting quite well, and produced accurate, pleasing color in most cases. It handled the difficult incandescent lighting of my Indoor Portrait test well with both its Incandescent and Manual white balance options. Color was more accurate than average among the cameras I've reviewed throughout my testing, and I particularly noticed the benefit of the RGB+E sensor technology in the level of detail that was preserved in the highly saturated red color of the flowers in the Outdoor Portrait shot. The camera has two color modes, "Normal" and "Real", the former producing more vivid and saturated colors than the latter. I chose the "Real" color setting for the majority of my test shots, as I felt that it produced a more accurate and natural-looking rendering of the subjects' original colors.
Phil Askey of DPReview.com has reported seeing an odd color shift from green to cyan in overexposed regions. I looked for this effect on my unit, but found it only to a very slight degree, not nearly to the extent that Phil saw it. Phil himself pointed out that it only affected a single image out of the hundreds he shot, so given that my unit showed the effect even less, I'd have to say that this is likely a non-issue in actual use.
- Exposure: Better than average exposure accuracy. Contrast adjustment works exactly right. The F828's metering system performed quite well on the majority of my shots, handling even the harsh high-key lighting of my Outdoor Portrait test much better than average. The camera's default tone curve tended to be slightly contrasty, losing highlight detail and plugging shadows on the deliberately harsh Outdoor Portrait torture-test. The contrast adjustment worked exactly as it should though, pulling in both highlight and shadow areas without shifting midtone brightness, color, or saturation. With the contrast adjustment set to "Low," the F828 did a decent job with harshly-lit scenes. (I'd really like to see an even broader range of adjustment though, with at least five steps instead of three.) On my "Davebox" test, the F828 distinguished the subtle pastel tones on the Q60 target without any trouble, yet still showed good detail in the shadow areas.
- Resolution/Sharpness: Exceptional resolution, with very few artifacts. 1,600 - 1,650 lines of "strong detail. "The F828 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 both horizontal and vertical directions, but I found "strong detail" out to at least 1,600 lines vertically, and perhaps 1,650 lines in the horizontal direction. "Extinction" of the target patterns didn't occur until about 1,950 lines. (This is the best performance to date (Late January, 2004) of any prosumer digicam I've tested.)
- Image Noise: Higher image noise than the best 5-megapixel cameras. We'll have to wait to see if this is a uniform characteristic of 8-megapixel digicams, once other units hit the market, from other manufacturers. Much has been made of the F828's image noise. There's no question that its images show more noise than those from the best 5-megapixel cameras, and it isn't even in the same category as digital SLRs like the Canon Digital Rebel and Nikon D100. That said, I didn't personally find its noise at low ISO settings to be particularly objectionable -- Noise is very much a subjective matter, for each prospective user to decide for him/herself. As you increase the ISO setting though, the noise does increase to what I would consider objectionable levels by the time you get to ISO 400. Its noise also tends to have a fairly large grain structure, which makes it more noticeable than it would be if it had a smaller, tighter pattern.
Since the F828 is the first 8-megapixel camera to hit the market, we don't know what its competition is going to look like in terms of noise performance. It's possible that we're now seeing the point at which we begin to seriously trade off noise for increased resolution. On the other hand, it could be that 8-megapixel cameras from other manufacturers will do better. - Stay tuned to this site, as there are going to be several other 8-megapixel models coming fairly soon in 2004, that'll beg comparison to the F828.
- Closeups: Very good macro performance, but a fair bit of barrel distortion and the lens blocks the flash at closest approach. The F828 performed very well in the macro category, capturing a minimum area of only 2.11 x 1.58 inches (54 x 40 millimeters). Resolution was very high, and detail was excellent in the dollar bill. The coins and brooch were soft due to the very short shooting distance, but that's not the camera's fault. There was a lot of softness in the corners, particularly on the left side of the frame, and there was a fair bit of barrel distortion too, even though it wasn't too visible in my standard macro test shot. The F828's long lens barrel blocked the flash, making it ineffective for this close-up shot, so plan on using external lighting for your closest shots.
- Night Shots: Excellent low light performance, with framing and focusing possible even in total darkness. Higher image noise than 5-megapixel models though. The F828 does very well at low light photography, offering full manual exposure control and maximum exposure time of 30 seconds. The F828 also offers Sony's outstanding Hologram Autofocus and infrared-based Night Framing options for excellent framing and focusing even in total darkness. The F828 produced clear, bright, usable images down to the 1/16 foot-candle (0.67 lux) limit of my test, with good color at all five ISO settings. The F828's Noise Reduction system did an good job of keeping noise in check, but there was clearly more noise visible than I've been accustomed to seeing in top of the line 5-megapixel cameras.
- Viewfinder Accuracy: An excellent LCD and EVF viewfinder system. The F828's electronic "optical" viewfinder (EVF) was very accurate, showing 99+ percent frame accuracy at both wide angle and telephoto zoom settings. The LCD monitor was also very accurate, since it shows the same view, just on a larger screen. Given that I like LCD monitors to be as close to 100 percent accuracy as possible, the F828's LCD monitor is essentially perfect in that regard.
- Optical Distortion: Lower than average barrel distortion, particularly for a wide-angle zoom. Average pincushion distortion. Higher than average chromatic aberration, some tendency to produce "purple fringe" around areas of light overload.
Optical distortion on the F828 was lower than average at the wide-angle end, where I measured approximately 0.7 percent barrel distortion, very good for a 28mm zoom, but still more than I like to see. (Barrel and pincushion distortion are pet peeves of mine. Most digicam lenses show 0.8-1.0 percent barrel distortion at the wide angle end, so it's not fair to single out any one manufacturer's cameras, but I'd really like to see all lenses improve in this area.) Interestingly though, the F828 gets better quickly as you zoom away from the full wide-angle setting. At an equivalent focal length of only 33.5mm, the barrel distortion drops to only 0.25 percent, an excellent figure by any measure. As usual, the extreme telephoto end did a little better, as I measured a 0.32 percent pincushion distortion there.
Chromatic aberration looks to be higher than average, showing about nine or ten pixels of fairly strong coloration on either side of the target lines. (This distortion is visible as a very slight colored fringe around the objects at the edges of the field of view on the resolution target.)
In my test shooting, I found some evidence of the infamous "purple fringe" problem, but seemingly less than some other reviewers did. (This is an optical phenomena that produces bright purple fringes around the edges of very bright objects in the image. It's apparently caused by a combination of conventional chromatic aberration and diffraction by the microlenses on the surface of the CCD itself.) I was able to trigger the purple fringe problem in the studio, particularly with very strong light sources, at relatively close range, and with the lens set to its wide angle position and maximum aperture. For whatever reason though, I saw little or no evidence of it in most of my test images. Not to say that the problem isn't there, just that the range of shooting conditions I worked under didn't seem to trigger it to the extent seen by some reviewers. For what it's worth, it does seem to require a pretty severe light overload to trigger the phenomena.
- Battery Life: Excellent battery life from the InfoLITHIUM system. While I still strongly recommend purchasing a second battery along with your F828, Sony's InfoLITHIUM system provides excellent run times, with a worst case run time (in capture mode with the rear-panel LCD illuminated) of over 3 hours. (188 minutes)
|Free Photo Lessons|
That said, it also has its limitations, although each user will have to decide how critical they are for their own applications. As many (myself included) had expected, it does show more image noise than the best 5-megapixel cameras. This was anticipated because the smaller pixels required to cram 8 million of them onto a single chip make for lower signal-to-noise ratios. I won't go so far as to suggest that this is what we'll see from all 8MP cameras though, as various manufacturers have surprised me in the past by beating what I thought were absolute limitations of the sensor processing. (Stay tuned, this is definitely going to be "the year of the 8 megapixel" at the high end of the field, with several models due from various parties.) The F828's optics also appear to be subject to the infamous "purple fringing" phenomena, producing magenta-colored fringes around the edges of very bright objects against dark backgrounds. For whatever reason, I didn't see this characteristic nearly as much in my testing as some reviewers did, but the effect is definitely present.
So overall, the F828 is a bit of a mixed bag, in my mind coming out ahead of the pack of current high-end digicams when all factors are considered. (At least as of this writing in late January, 2004) There's a great deal to like, particularly for users who want loads of resolution (it has that in spades), superb color, ultra-fast shutter response, and Sony-only technologies like Hologram Autofocus and NightShot/NightFraming. On the other hand, if you're bothered by higher than average image noise (relative to current high-end prosumer cameras at least, all of which were still 5 megapixel models as I write this) and the dreaded "purple fringe" problem, it may not be the camera for you. If most of your shooting is outdoors (e.g., low ISO environments), and you don't have a lot of burned-out highlights in your images, chances are you'd never encounter the camera's primary negative points. Bottom line, I'm going to give it a provisional "Dave's Picks" status, awarding that accolade for the resolution category and Sports usage (thanks to the exceptional responsiveness), but not for the general "Enthusiast" or "Long Zoom" categories.
Top 3 photos this month win:
1 Canon PIXMA PRO-100
2 Canon PIXMA MG6320
3 Canon PIXMA MG5420