Sigma SD9 SLR Review
By Movable Type Admin
The SD9 is the first digital camera to incorporate Foveon's X3 image sensor and it's
Sigma's first digital camera. The SD9 brings the price of owning an interchangeable
lens digital SLR camera down to just $1799. I'm not going to go into any details here
about the Foveon imager, you can read about it
here. And if you have several hours to spare you can peruse DP Review's Sigma SLR Talk
Forum and read the pro and con comments from the user community. Just bear in
mind that 99% of them haven't even seen or used the camera as of the time of this
Suffice it to say that the Foveon sensor technology is new and this is Sigma's
first digital camera, we'll undoubtedly see future improvements from both Foveon and
All SD9 images require post-processing due to its raw-only image format. Sigma bundles their Photo Pro software with the SD9 and this gives the user a lot of control over the output image. It has slider controls with ±2.0 adjustment (in 0.1 steps) of the exposure, contrast, shadow, highlight, saturation and sharpness and a "color wheel" with eyedropper to make changes in the CMY color balance. The true image purists out there will applaud Sigma's use of the raw format but it will also be seen as an unnecessary "extra" step by potential buyers that are not that computer savvy. It's not always "convenient" to have to post-process EVERY image, other dSLRs give you the choice of capturing raw or finished JPEG (or TIFF) images in the camera. The Kodak DCS pro cameras used to only save in a proprietary raw format but Kodak saw the need for in-camera JPEG to be added later. They went so far as to create a whole new type of JPEG (JPEG-ERI) format to give you the convenience of a "ready to use" JPEG but with the added ability to post- process it with embedded raw data if desired. Compressed JPEG files also require a lot LESS storage space than raw image files. Sigma Photo Pro lets you output to 16-bit TIFF, 8-bit Exif-TIFF or JPEG format. A problem with the current JPEG compression algorithms has been identified, a solution and explanation for it can be found at: jpegclub.org/foveon/. The word is that Foveon will implement the suggested changes in a soon to be released upgrade of the software. This could potentially solve some of current image problems that we are seeing now.
The USB port is incredibly slow. A combination of the large file sizes (5-10 MB each) and a poor hardware/software routine makes it virtually unusable. The SD9 is not a Mass Storage compliant device so you must use their software for image downloads. My computer is a fairly new P4/2000Mhz with 1GB RAM and USB 2.0 running Windows XP Pro with all the latest patches. I plugged in the USB cable and fired up the software, after many minutes I thought it had locked up. The software first shows how many files are on the CF card and then builds thumbnails for each of them. It's painfully slow and the lack of a progress indicator will have you wondering if it's working or locked up the computer. The FireWire port is faster, you can visibly see the thumbnails being created and the transfer of selected images is accomplished at about 10 seconds per image. Sigma should have bundled a FireWire PCI card with the camera and just skipped the USB interface altogether. I highly suggest the use of a USB 2.0 or FireWire card reader.
The SD9 is a hybrid camera like the Fuji S2 and requires one set of batteries for the digital portion (four AA or two CR-V3 cells) and another set of batteries for the camera functions (two CR-123 lithium one-use.) I didn't really have the time to do any kind of technical battery usage analysis during our 2-week review period. We used freshly charged 1800mAh PowerEx and 2100mAh iPowerUS NiMH cells. These batteries are inexpensive but heavier than the proprietary lithium battery packs used by the D100 and D60. I'd say on average that a set of 2100mAh cells lasted us for a good two to three hour shoot with the camera on all the time and a 5-second preview on the color LCD after every shot. We were shooting with a 512MB SanDisk and a 640MB Delkin Pro CF card. We never did "kill" the CR-123 batteries during our testing, the indicator always showed them as "full" so I don't know how long they'll last. Be sure to have a spare set of CR-123 batteries with you as the camera won't operate without them.
Physically, the SD9 is only slightly larger than the
Nikon D100 or Canon D60. Like these two cameras, the SD9 also has an optional "battery
booster" that can be attached on the bottom. Sigma was unable to supply us with one of
these during our review period so I can't comment about it one way or the other. The SD9
is a little "boxy" in appearance compared to the D100. I liked the way it felt in my
hand, the grip is adequate and secure. The body is constructed from a polycarbonate
material that makes it both light and durable. When fully outfitted with the 20-40mm zoom
and batteries the SD9 was a little heavier than the D100 with a 24-85mm zoom. The
fit-n-finish of the camera is very good and the controls are clearly marked and located in the
proper places for easy operation. It has the same look and feel of a 35mm SLR and for
good reason, it was designed around Sigma's SA9 film camera. The lenses are easy to put
on and take off, the Sigma SA lens mount is a bayonet type requiring only a 1/4 of a turn
to affix the lens. The viewfinder is large and bright with a good eye relief and lots of
dioptric adjustment. The viewfinder's coverage is about 98% and is unique in the way that
it shows more than just the active capture area.
This is a SLR type camera so its 1.8-inch color LCD display cannot be used as a
live viewfinder. It does let you immediately preview the image just captured or review
any previously captured images. It also lets you access the camera's menu system.
We found the color LCD to have very good visibility in all but the brightest outdoor
conditions. The image rendition is very good and a playback magnification of up to 400%
lets you check for critical focus as well as composition.
Pressing the INFO button toggles between a full screen image and the Info screen with a smaller image window and the exposure info strip across the top as well as time and date, metering mode, exposure compensation, image size, exposure mode, focus mode, focal length, white balance, image dimensions and exact file size. It also gives you a histogram display with separate curves for each of the color channels. The histogram is keyed directly to the playback image, even when magnified, so it can be used to accurately analyze smaller portions of the image.
Some performance statistics: The SD9 is ready to take a picture in less than a second from flipping the power switch on (except when using a Microdrive which slows the startup process to just under 3 seconds.) Using the 50mm lens it exhibited an average autofocus lag time of just over half a second. Shooting in single frame mode in HI quality the shot to shot time is .5-.6 sec. per frame up to six frames. It took an average of ten seconds before the next frame could be captured as the buffer contents were being processed. If you filled the buffer it took about fifty seconds to one minute to write the buffer contents. We tried an "average speed" CF card and one of the fastest pro cards on the market and noticed little to no difference in processing times so the camera clearly does not have an enhanced CF interface. In continuous mode the frame rate for HI was a little less than 2fps with a depth of six frames and a 10 sec. delay before the next frame could be captured. To write the full buffer was the same as in single exposure mode, about a minute. The frame rate in MED was slightly faster than 2fps with a depth of 14-15 images and a 5-6 sec. delay before the next frame. Writing out the buffer required a minute and a half. The LOW frame rate was close to 3fps with a depth of 30 images and a 5-6 sec. delay before the next frame but it took two and a half minutes to write the buffer.
Sigma manufactures a wide variety of excellent prime and zoom lenses in their SA mount as well as popular Canon, Minolta, Nikon and Pentax mounts. Sigma sent us a 50mm EX f2.8 macro and the 20-40mm EX f2.8 zoom lens. Because the image sensor has a capture area smaller than 35mm film, the SD9, like most other dSLRs, has a focal length multiplier. The SD9's focal length multiplier is 1.7x so this means that the 50mm lens is equivalent to a 85mm lens and the 20-40mm zoom covers the 34-68mm focal range. This is great for the telephoto lovers out there because it's like having a 1.7x teleconverter that doesn't affect the aperture value. However it also means that doing true wide angle photography is difficult, more expensive and sometimes impossible.
The SD9's autofocus system is generally good but not as "robust" as the D60 or D100 cameras when using the latest Canon or Nikon lenses. The SD9 has no user-selectable focus points. You have the choice of Single-AF where the camera focuses only after the shutter is pressed halfway or Continuous-AF. The Continuous-AF mode is always adjusting the focus and incorporates an AF Predict function that tracks a moving subject and predicts where it will move to next. In low light there is frequent AF hunting and more than once the lens (50mm) got stuck at one limit or another and required some manual assistance. The SD9 lacks a focus illuminator and as Sigma failed to send us the EF-500 DG Super SA speedlight, I don't know how well its focus illuminator works. The SD9 has no built in flash so you're going to need an external flash, only the EF-500 DG Super SA or EF-500 DG ST SA flashes will integrate with the camera's TTL exposure system. Other flashes can be used in manual exposure mode. The fastest flash x-sync speed is 1/180 sec. which limits its use for daylight fill flash.
In most cases the SD9's images are amazingly sharp and defined with very good color rendition. In fact, they're some of the most sharpest images I've ever seen from a camera with a 3.4 megapixel sensor. Most of this is due to the fact that the Foveon sensor does not employ an anti-aliasing filter which tends to blur fine details. On the flip side, when examined at 200% (or greater) there are lines or curves that exhibit prominent stair- stepping (jaggies.) Red colored objects can be oversaturated at times to the point that they look to be painted in. There is an abundance of red channel noise visible in all images. More than a few of my sample images have unnatural and over-saturated blue skies. We've also seen chromatic aberrations (purple fringing) on contrasting highlight areas. There seems to be less than half a stop of exposure latitude before highlights get "blown out." Long exposures are problematic to the point that shutter speeds slower than one second are not even available at ISO 200 or 400. Long exposures (beyond 2 seconds) at ISO 100 lose sharpness and color saturation, possibly due to the effects of an undocumented noise reduction process.
The image quality issue is always very subjective, potential buyers should checkout our sample images and those found at the other review sites - see our "Second Opinion" links below. If you need a more technical analysis of the pros and cons of the SD9's images then read Imaging-Resource's and DP Review's SD9 reviews. Dave and Phil have expended a lot of time and effort to fully cover the technical aspects of this camera and sensor. I have always tried to give folks just the "short version" from a more non-techy point of view. This may not be "fair" to a new technology like the Foveon sensor but the bottom line is how do the images look to YOU? Download a variety of sample images, look at them in a graphics program with a properly calibrated monitor (not a web browser window), print them out, ask other people what they think about the pictures... Don't get caught up in a feeding frenzy over a new product just because people in discussion forums think it's the greatest thing since sliced bread. It's like buying a stereo due to its advertised (and over-hyped) lower wow and flutter or higher dynamic response numbers rather than just listening to how it sounds with your own ears.
A manufacturing quality control problem has been identified by all reviewers. My camera
had an abundance of dust or dirt on the sensor and this caused visible dark spots
on some images. These "spots" become more pronounced as the aperture gets smaller and
are very visible in blue sky areas of outdoor photos.
Dave and Phil also commented about their cameras having the same problem.
I was not comfortable with cleaning the sensor as the number of SD9's
available for review were very limited and if damaged we could have been
"out of business" for finishing our review. I don't feel that a user should be
required to clean the sensor on a brand new camera, especially one that employs a dust
protector to eliminate these problems in the first place. Like I said at the
beginning of this conclusion, the Foveon X3 sensor technology is new and this is only the
first camera to use it. For a "first effort" I think that Sigma has put together a very
capable camera that will only get better in the future. Do I recommend it? With
the Nikon D100 and Canon D60 being
available for only $200-400 more, I'd pass on the SD9 until it matures a bit.
Sigma Recalls SD9 Battery Cases
12/25/02 - Sigma Japan has posted a recall notice for the 'battery case' (the cartridge which slides into the base of the camera) for cameras with a serial number between 1000251 and 1000459. This is to correct a fault where the amount of remaining power is incorrectly reported when using NiMH batteries, this problem does not occur with Lithium CR-V3 type batteries. The notice states that only the battery case need be returned and new cases will be supplied in return. It's not clear yet if this range of serial numbers affects any cameras outside of Japan.
Want a second opinion?
Return To Steve's
Visitors of Steves can visit the stores below for real-time pricing and availability. You can also find hot, soon to expire online offers on a variety of cameras and accessories at our very own Camera Deals page.