Sigma SD9

I include links in our reviews to a Thumber-generated index page for the 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!

For an interesting side by side comparison between the SD9 and three other competing digital SLRs (with a variety of subjects) see my Digital SLR Shootout pages.

Outdoor Portrait: Hmm - Workable, but high default contrast, takes a bit of fiddling in the Photo Pro software to get a decent-looking result. The extreme tonal range of this image makes it a tough shot for many digicams, which is precisely why I set it up this way, and why I shoot it with no fill flash or reflector to open the shadows. The object is to hold both highlight and shadow detail without producing a "flat" picture with muddy colors, and the SD9 had a good bit of trouble with it. By default, the SD9 seems to have a fairly contrasty tone curve, particularly apparent when dealing with a high-contrast subject like this one. The Photo Pro software is a big help in dealing with the consequences, dropping the contrast, bringing up the shadows, and controlling the highlights. Unfortunately, there seems to be very little headroom in the raw files for overexposure, and boosting exposure to bring up shadows increases image noise there a fair bit.. I played with this shot quite a bit, working with different original exposures, and varying amounts of adjustment in Photo Pro. The shot I felt came out the best was one with +0.5EV of exposure compensation, which left the image rather dark overall, but didn't lose quite as much detail in the highlights, and didn't result in as much noise in the shadows after processing. The "main" shot above shows the results of this image as it came from the camera, using the defaults in the software to convert it to a JPEG. The images below compare the default version with the software's "auto" processing, and my own manual manipulations of the image. The image right out of the camera was rather contrasty. This shot was taken with +0.5EV of exposure compensation, to try to open the shadows on Marti's face a little, so I wouldn't have to tweak the image so hard in the software. The highlights in her collar are all but lost (even in the RAW file), but if I'd gone with less exposure, I'd have ended up with tons of noise in the shadows after tweaking in Photo Pro. This is the result of Photo Pro's "auto" adjustment. The "auto" setting only seemed to help with images that were pretty close to being OK to begin with. It didn't handle anything with extreme contrast or marked color casts very well at all. This is the result of my manual tweaking in PhotoPro. I dialed down the contrast a fair bit, which pulled in both the shadows and highlights quite a bit. With the highlights toned down, I boosted the exposure, which opened the shadows. (Even though the highlights here are well below being saturated at 255,255,255, there's very little shape left in them.) Settings in Photo Pro were as follows: Exposure: +0.6 Contrast: -0.5 Shadow: 0 Highlight: -0.1 Saturation: 0 Sharpness: 0 Color: 0 I was a little surprised to not find more headroom in the SD9's RAW-format files. While I haven't subjected them to my usual tests, I've seen demonstrations of Kodak's RAW file format, where it was relatively easy to recover excellent image data from shots that were overexposed by a full f-stop. In the next set of "shootouts" I do, I'll take a look at how effective different RAW file formats are at preserving data in extreme highlights and shadows for post-exposure manipulation. In terms of color, the SD9's shots of this subject came out looking pretty good, although Marti's skin tone was a little yellowish, and the blues and reds in the flowers were moderately oversaturated. At the same time, the greens in the leaves are rather flat and dull-looking. The SD9's white balance system seemed to do the best with the daylight setting - Auto seemed a little cool-toned, and manual came out quite yellowish. Auto white balance, +0.5 EV exposure, default settings in the software. Daylight white balance, +0.5 EV exposure, default settings in the software. Manual white balance, +0.5 EV exposure, default settings in the software. Resolution is very good here, with the image looking nearly as sharp as those from competing 6-megapixel cameras. (See my comments in the resolution test section below for more detail on this.)

Closer Portrait: Excellent resolution and detail, very contrasty file from the camera though. Overall results are similar to the wider shot above, with very good detail, but again more contrast than I'd like to see. The shot at right is the result of the default exposure setting on the camera, and a fair bit of tweaking in the PhotoPro software. Although it resulted in very dark shadows, I went with the default exposure to try to preserve at least a little detail in the highlights. As a result though, the shadows show more noise than you'd see from the camera with a less-contrasty subject. The default shots from the camera looked so contrasty on this shot that I "cheated" a little and used the custom-manipulated one for my "main" example, even though that's not entirely fair to other cameras I test, which don't have the benefit of manual tweaks being applied to their images. The table below shows the results of default, auto, and manual adjustments in Photo Pro, all based on the default exposure from the camera. Default exposure in the camera, default processing in Photo Pro. Default exposure in the camera, "auto" processing in Photo Pro. Default exposure in the camera, manual adjustments in Photo Pro. Exposure: +0.7 Contrast: -0.4 Shadow: +0.3 Highlight: -0.1 Saturation: +0.3 Sharpness: 0 Color: 2M+1Y

Indoor Portrait, Flash: Since the SD9 has no built-in flash, and I didn't have one of Sigma's dedicated units to test with it, I didn't shoot this test.

Indoor Portrait, No Flash: Auto White Balance Incandescent White Balance Manual White Balance Mediocre handling of a difficult but common light source by the SD9's white balance system. This shot is always a very tough test of a camera's white balance capability, given the strong, yellowish color cast of the household incandescent bulbs used for the lighting. This is a very severe test of digicam white balance systems, but it's a very common light source here in the US. Digicams responses to it range all over the map, with many having great difficulty. - At the price point of the SD9 though, you'd expect the camera to have greater white balance flexibility. Both auto and incandescent white balance settings produced rather warm-toned results, while the manual setting resulted in a distinct greenish cast. (All images at right are the result of default processing in Photo Pro, showing the images more or less as they came from the camera.) I chose the manual white balance setting as the main selection for this section, feeling that it was closest to accurate. For some reason, the SD9's exposure system *really* wanted to underexpose this shot. - The images at right were all captured with an exposure compensation of +1.5 EV.(!) The light background tricks most digicams into underexposing this shot, but +1.5EV is more than most pro models require here. Here again are the results of default, auto, and manual processing in Photo Pro. The Auto setting here did better than with the outdoor shot above, but I felt still left things a little dark-looking. I also felt that the best color balance was one with a bit more magenta in it than the auto setting provided. Default exposure in the camera, default processing in Photo Pro. Default exposure in the camera, "auto" processing in Photo Pro. Default exposure in the camera, manual adjustments in Photo Pro. Exposure: -0.1 Contrast: 0 Shadow: +0.2 Highlight: +0.5 Saturation: 0.1 Sharpness: 0 Color: 5C+5M   ISO Series: The SD9's ISO options are limited to 100, 200, and 400, a somewhat restricted range when compared to the more expensive SLRs it competes against. All shots are taken with Auto white balance, and +1.5EV of exposure compensation, except for the one on the right. That image on the shows something I found on this test with the SD9, namely that the color balance seemed to shift a fair bit with the exposure compensation setting. At lower ISO settings, the shift between +1.5 and +2.0 EV was more subtle, but at ISO 400 with this light source, the camera seemed to figuratively throw up its hands, producing the bright yellow image at far right. As far as noise goes, the SD9 is a bit noisier than competing models at ISO100, with the noise increasing steadily as you move to 200 and 400. While I don't have any quantitative measurement to apply to this shot, my impression is that these noise levels are more reminiscent of consumer-level cameras than pro SLRs. (This is far from a "night" shot, the light level corresponds to pretty bright household lighting, a good 400 watts+ of light in the room, yielding a light level of about 40 foot-candles, 440 lux, give or take.) ISO Series ISO 100 +2.0 (!) EV ISO 200 +2.0 EV ISO 400 +1.5 EV ISO 400, +2.0 EV (See text for note on color cast)

House Shot: Auto White Balance Daylight White Balance Manual White Balance Great resolution, detail, and color. The SD9's Auto white balance setting produced the best results here, with the most accurate white value on the house trim. The Daylight setting was nearly accurate, though slightly warm, and Manual white balance resulted in a yellow/green cast. Resolution is very high, as the tree limbs and shrubbery show a lot of fine detail, quite a bit beyond what you'd normally expect to see from a 3.4 megapixel sensor. The detail is exceptionally crisp, due to the 3-color pixels of the Foveon sensor, as well as the lack of an anti-aliasing filter over the sensor array. A great job.

Far-Field Test Excellent resolution and detail, with a good dynamic range. 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. This is my ultimate "resolution shot," given the infinite range of detail in a natural scene like this, and the SD9 did a very good job with it, at least in the resolution arena. The tonality and color could use a little help in PhotoPro (the shots here were imported with the default settings), but the details in the fine limbs and foliage above the house are super-fine and crisp. Definitely impressive for a 3.54 megapixel sensor. Resolution Series: 20-40mm EX zoom lens (Quality 12 JPEG from Sigma/Foveon software.) 2268 x 1512 SD9FARLF 1512 x 1008 SD9FARMF 1134 x 756 SD9FARSF

Lens Zoom Range - Since the SD9 uses interchangeable lenses (and hence the zoom range is a characteristic of the lens, rather than the camera), there are no shots in this category.

Musicians Poster Auto White Balance Daylight White Balance Daylight White Balance Good color with auto white balance, although a slightly cool tone. Odd underexposure problem though. This shot is often a tough test for digicams, as the abundance of blue in the composition frequently tricks white balance systems into producing a warm color balance. The SD9's Auto white balance setting did a very good job with it though. Resolution and detail are both excellent. (Click on a thumbnail at right to see the full-sized image.) The strangest thing about this shot is that the SD9 wanted to significantly underexpose it for some reason. The shots at right have a full f-stop (1EV) of positive exposure compensation dialed in, and the result is just about right. I'm really puzzled as to why the SD9 would have had such hard time exposure-wise with this shot, since it's not at all high-contrast, and other cameras seem to handle it just fine.

Macro Shot Standard Macro Shot Small macro area with great detail, but auto white balance had some difficulty. Well, if I avoided shooting the zoom range shot because the SD9 uses interchangeable lenses, I probably should have skipped this shot as well. I'm a bit of a macro nut though, and the 50mm f/2.8 macro lens that Sigma shipped along with the SD9 is a really nice lens for macro shooting. Here again, the SD9 wanted to underexpose the subject, but not as badly as before. The shot at right was snapped with the default exposure setting, and default processing in the Photo Pro software.

"Davebox" Test Target Auto White Balance Daylight White Balance Manual White Balance Generally very good color, hue-accurate and properly saturated. The Auto white balance produced the best color here, with the most accurate white value in the mini-resolution target and large, white color block. Manual white balance was rather yellow, and the Daylight setting a bit less so, but still warm. Color is generally accurate and well-saturated, although the red is slightly oversaturated, and the green slightly undersaturated. Image noise is only slightly higher than competing SLRs in the neutrals and most colors, but gets quite high in the red and magenta swatches. (I'm working on a more complete noise analysis, including noise at various ISOs, will post some comparative numbers as soon as I can get to it.)   ISO Series: ISO Series ISO 100 ISO 200 ISO 400

Low-Light Tests An notable weak point. Image noise and short exposure times limit performance. One of the weakest areas for the SD9 is its low light capability. At ISO 100, shutter times to 15 seconds permit shooting at light levels as low as 1/8 foot-candle (about 1.3 lux), corresponding to light levels about 3 stops down (1/8) that of typical city streetlighting, although at that level, the images are quite noisy. At ISOs of 200 or 400, the shutter time is limited to a maximum of one second, which severely limits the low light capabilities. Image noise under low light conditions appears to be quite a bit worse than that of competing SLRs. Colors are also rather muted. Not the camera to purchase if you need to do a lot of after-dark work...   1fc 11lux 1/2fc 5.5lux 1/4fc 2.7lux 1/8fc 1.3lux 1/16fc 0.67lx ISO 100 3 sec F/2.8 ISO: 100 4 sec F/2.8 ISO: 100 10 sec F/2.8 ISO: 100 15 sec F/2.8 ISO: 100 ISO 200 1/1 sec F/2.8 ISO: 200 1/1 sec F/2.8 ISO: 200 ISO 400 1/2 sec F/2.8 ISO: 400 1/1 sec F/2.8 ISO: 400 1/1 sec F/2.8 ISO: 400   Love high ISO photography? Hate noise? Check out Fred Miranda's ISO-R noise-reducing actions for Photoshop. Incredible noise reduction, with *no* loss of subject detail. (Pretty amazing, IMHO.) Check it out!

Flash Range Test - Since the SD9 uses only external flash units, this test doesn't apply to it.

ISO-12233 (WG-18) Resolution Test Moderately good resolution, about 1,000 lines per picture height, before aliasing obscures detail. Resolution with the SD9 and its Foveon X3 image sensor is going to be a controversial issue. Foveon has claimed that the X3 technology provides twice the resolution of a conventional striped array, on a pixel-for-pixel basis. In the case of the SD9's 3.54 megapixel sensor, that would equate to resolution equivalent to that of a conventional 7 megapixel CCD or CMOS. In my testing though, I found that the SD9 resolved a bit less detail than competing cameras using six-megapixel sensors, although the difference wasn't dramatic. - Overall, impressive for a 3.54 megapixel array, but not quite up to the levels claimed by Foveon. "Resolution" is an area open to a lot of interpretation though. When most of us talk about "resolution," we're probably thinking in terms of how "sharp" an image looks. - And on that score, the SD9 does very well. Technically, "resolution" refers to how much detail a camera can see or "resolve," while sharpness more accurately describes how abruptly brightness values change across an edge. As you'll see in the image crops below, the SD9's X3 chip delivers very "crisp" or sharp images, but the resolvable detail isn't quite up to the levels delivered by competing cameras like the Canon D60, Fuji S2, or Nikon D100. (I included the D60 and D100 in the comparison, but deliberately left out the S2. (Which does slightly better than any of the three cameras compared.) - I didn't want to pry open the unequal-file-size can of worms any further than I already had in upsampling the SD9's image to match the size of those from the D60 and D100. The S2 really delivers its best performance with its interpolated 12-megapixel output file size, which would have required another level of interpolation.) This first shot above (click on the image to see the full-sized version) compares the horizontal-resolution component of the ISO12233 res target across the three cameras mentioned. More detailed text is attached to the image file. - Note the tradeoff between aliasing, "sharpness," and resolvable detail. This second shot shows another look at the sharpness/aliasing tradeoff. (I also divulge the explanation for the spurious dots that sometimes appear in D60 res-target images, but apparently rarely in "real world shots. ;-) As before, click on the thumbnail above to see the full-sized image. This is the point in these reviews where I usually report on optical distortion in a camera's lens. Since the SD9 uses interchangeable lenses, there's no point in talking about distortion here. - Other than perhaps to say that the professional-grade Sigma lenses seem to be both quite sharp and relatively distortion-free. Resolution Series 50mm f/2.8 prime lens (Stopped down) (Quality 12 JPEG from Sigma/Foveon software.) 2268 x 1512 SD9RESLS0 1512 x 1008 SD9RESMS0 1134 x 756 SD9RESSS0

Viewfinder Accuracy/Flash Uniformity Excellent accuracy from the "sports" viewfinder. I have to say that I'm not wild about the SD9's "sports" viewfinder, with its small active area. To its credit though, the finder is very accurate, with the cropped central area showing almost exactly 100% of the image area. (Shot with 50mm prime lens)