Sony A900 Imatest Results

Sony A900 Imatest Results

We routinely use Norman Koren's excellent "Imatest" analysis program for quantitative, thoroughly objective analysis of digicam test images. I highly recommend it to our technically-oriented readers, as it's far and away the best, most comprehensive analysis program I've found to date.

My comments below are just brief observations of what we see in the Imatest results. A full discussion of all the data Imatest produces is really beyond the scope of this review: Visit the Imatest web site for a full discussion of what the program measures, how it performs its computations, and how to interpret its output.

Here's some of the results produced by Imatest for the Sony A900:

The Sony A900 showed very good color accuracy overall. Hue accuracy was quite good, with most of the hue shift occurring in the cyans, sky blues, reds and oranges. Average saturation was 105% (oversaturated by 5%, mostly in the deep blues and reds and some greens). Average "delta-C" color error was only 5.27 after correction for saturation, which is quite low, albeit very slightly higher than some of the competition. All in all, a very good color response for an SLR. Mouse over the links below the illustration above to compare results with competing models.

Using the Adobe RGB color space (which provides a much wider gamut, or range of colors that can be expressed), the Sony A900 delivers more highly saturated color, with an average saturation of 108% and average saturation-corrected hue error of 5.08 "delta-C" units, which in this case is slightly more accurate than the competition. Again, mouse over the links below the illustration above to compare results with competing models.

This image shows how the Sony A900 actually rendered the colors of the MacBeth chart, compared to a numerically ideal treatment. In each color swatch, the outer perimeter shows the color as actually captured by the camera, the inner square shows the numerically ideal color, after correcting for the luminance of the photographed chart (as determined by a second-order curve fit to the values of the gray swatches), and the small rectangle inside the inner square shows the numerically ideal color, without any luminance correction. This image shows the A900's very good hue accuracy, as well as a gamma curve that results in an overexposure of some colors, notably cyans, reds, orange and some greens. (We've seen this in most cameras, increasing the "punch" of images by brightening saturated colors a little, while actually keeping both the hue and saturation level very close to technically accurate values.)

There's a lot in this particular graph, a lot more than we have room to go into here. (This set of plots has also changed a little in the more recent versions of Imatest. Some of the plots that were shown here previously are now shown in other Imatest output. Since we largely focus on the Noise Spectrum plot, we'll only show the graphic above, which includes that plot.)

In comparing these graphs with those from competing cameras, I've found that the Noise Spectrum graph at lower right is the most important. Cameras that manage to shift their noise spectrum to higher frequencies have much finer-grained noise structures, making their noise less visually objectionable. In the graph above, this would show up as a noise spectrum curve that remained higher on the right side, representing higher noise frequencies. The champion at this was (and still is) the Canon EOS-1Ds Mark II, which produced remarkably fine-grained image noise, even at very high ISOs.

Here, we see the results at ISO 100, which is as a low extension of the normal ISO range for the A900. The luminance curve is quite flat (meaning luminance noise is very fine-grained), however the color channels exhibit higher noise values, especially at lower frequencies. We also see the now familiar spikes (though to a much lesser degree), which are a result of how Sony's JPEG engine processes and assembles each block of the image (see the Sony A200 or A350 review for details). This graph reinforces our own visual observations of the A900's images, in which we saw that the luminance noise was relatively low, but that the chroma noise was much higher.

Here are the results at the A900's base ISO?of 200. Again, the luminance curve is quite flat and the A900 does a good job of keeping plenty of the noise energy at high frequencies. The blue and green channels follow the luminance channel more closely than at ISO 100, but the red channel still shows quite a bit more noise energy at lower frequencies. Bottom line, the Sony A900 noise levels are low at low ISO, with a frequency spectrum that has more energy on the higher frequency (right-hand) side of the curve than do many cameras in the blue, green and luminance signals, but less so in the red. This blotchiness in the red channel can be seen in the sky area of our Far-Field shot of the house. In that same shot, you can also see some red noise in the shingles of the roof that doesn't correspond to their actual color.

Above is the same set of noise data at ISO 3,200. Here, the Noise Spectrum graph has a dramatic shift toward the left-hand, lower-frequency side than at ISO 200, coarsening the "grain" of the image noise patterns. The red channel also continues to have dramatically more noise energy at lower frequencies than the others. This can be seen in the A900's ISO 3,200 images as very noticeable blotches of chroma noise.

Here's the same set of noise data at ISO 6,400. Very similar to ISO 3,200, with slightly more shift to the left indicating even coarser "grain" caused by the A900's noise reduction attempts to blur out the noise. Again, the red channel continues to have much more noise energy at lower frequency than the others. This seems a little odd, as most noise reduction algorithms reduce noise in all channels at high ISOs.

This chart compares the Sony A900's noise performance over a range of ISOs against that of other current, full frame cameras, using default settings. While we continue to show noise plots of this sort because readers ask for them, we each time point out that the noise magnitude is only a small part of the story, the grain pattern being much more important. Also, this is a plot of luminance magnitudes, where the A900 does relatively well: A plot of chrominance noise would tell a different story. In the case of the Sony A900, the magnitude of the image noise starts out slightly above most of the competition at ISO 100, and remains higher at all ISOs until the A900's high ISO?noise reduction kicks in aggressively at ISO 1,600. This brings the curve back down to the A900's ISO?800 level, and in-line with the Canons, but higher than the Nikons, which are "only" 12-megapixels. There's quite a large difference at ISO 6,400, but still a pretty good performance for a camera with 24-megapixels. Keep in mind these are at default settings, so the shape and positions of the curves will be influenced by your settings.

As evidence of the difference between what these graphs show and the appearance of actual images, check out the Sony A900 High-ISO NR page of this review for direct comparisons against the Canon EOS-1Ds Mark III.

Sony A900 Dynamic Range Analysis

A key parameter in a digital camera is its Dynamic Range, the range of brightness that can be faithfully recorded. At the upper end of the tonal scale, dynamic range is dictated by the point at which the RGB data "saturates" at values of 255, 255, 255. At the lower end of the tonal scale, dynamic range is determined by the point at which there ceases to be any useful difference between adjacent tonal steps. Note the use of the qualifier "useful" in there: While it's tempting to evaluate dynamic range as the maximum number of tonal steps that can be discerned at all, that measure of dynamic range has very little relevance to real-world photography. What we care about as photographers is how much detail we can pull out of the shadows before image noise becomes too objectionable. This, of course, is a very subjective matter, and will vary with the application and even the subject matter in question. (Noise will be much more visible in subjects with large areas of flat tints and subtle shading than it would in subjects with strong, highly contrasting surface texture.)

What makes most sense then, is to specify useful dynamic range in terms of the point at which image noise reaches some agreed-upon threshold. To this end, Imatest computes a number of different dynamic range measurements, based on a variety of image noise thresholds. The noise thresholds are specified in terms of f-stops of equivalent luminance variation in the final image file, and dynamic range is computed for noise thresholds of 1.0 (low image quality), 0.5 (medium image quality), 0.25 (medium-high image quality) and 0.1 (high image quality). For most photographers and most applications, the noise thresholds of 0.5 and 0.25 f-stops are probably the most relevant to the production of acceptable-quality finished images, but many noise-sensitive shooters will insist on the 0.1 f-stop limit for their most critical work.

The image below shows the test results from Imatest for an in-camera JPEG file from the Sony A900 with a nominally-exposed density step target (Stouffer 4110), and the A900's settings at their default positions, where Dynamic Range Optimization is set to Off.

These are good results for in-camera JPEGs, near the results of other recent Sony SLRs we've tested, but not as good as recent results from other manufacturers. The tone curve shows excellent gradation in highlights, but the shadow end trails off a bit more abruptly.

Interestingly, the results with DRO set to Off were slightly better at the highest quality level (7.31 f-stops), however overall dynamic range remained the same at 10.6 f-stops. This is likely due to the fact that DRO standard increases shadow levels, which also makes noise slightly more evident, knocking the highest quality level down a bit to 7.26 f-stops.

Selecting DRO Plus (Auto) didn't have much of a result on the dynamic range numbers either, and also resulted in a warning about "indistinct steps" from Imatest.

Our standard RAW converter (Adobe Camera Raw, version 4.6 beta) was able to extract quite a bit more dynamic range (over 2 full stops), with a score of 9.36 f-stops at the highest quality level. These are excellent results, as the Sony A900's dynamic range score really benefits from ACR's default noise reduction. (Do note though, that the level of manipulation done to achieve the results above will result in severe color artifacts on color images: This level of dynamic range is really only achievable on monochrome images.)

Dynamic Range, the bottom line:

The net result was that the A900 JPEGs came in below those from most other DSLR models we've tested recently, but results from ACR processed RAW files were much better, only surpassed by the Nikon D700 at the highest quality level. An excellent performance.

To get some perspective, here's a summary of the Sony A900's dynamic range performance, and how it compares to other digital SLRs that we also have Imatest dynamic range data for. (Results are arranged in order of decreasing dynamic range at the "High" quality level.):

The results shown in the table are interesting. One of the first things that struck me when I initially looked at test data for a wide range of d-SLRs, was that here again, purely analytical measurements don't necessarily correlate all that well with actual photographic experience. There's no question that the Fuji S3 Pro deserves its place atop the list, as its unique "SR" technology does indeed deliver a very obvious improvement in tonal range in the highlight portion of the tonal scale. I was surprised to see the analytical results place the Olympus E-300 as highly as they did, given that our sense of that camera's images was that they were in fact noisier than those of many other d-SLRs that we looked at. In the other direction, I was quite surprised to see the Nikon D2x place as low on the listings as it did, given that we found that camera's shadow detail to be little short of amazing.

One thing that's going on here though, is that we tested each camera at its lowest (base) ISO setting, which should produce best-case noise levels. This is in fact what many photographers will be most interested in, but it does perhaps place some of the Nikons (like the D40) at a disadvantage, as their lowest ISO setting is 200, as compared to the ISO 100 settings available on most other models.

Sony A900 Resolution Chart Test Results

The chart above shows consolidated results from spatial frequency response measurements in both the horizontal and vertical axes. The "MTF 50" numbers tend to correlate best with visual perceptions of sharpness, so those are what I focus on here. The uncorrected resolution figures are 2,340 line widths per picture height in the horizontal direction (corresponding to the vertically-oriented edge), and 2,298 lines along the vertical axis (corresponding to the horizontally-oriented edge), for a combined average of 2,319 LW/PH. Correcting to a "standardized" sharpening with a one-pixel radius increased both vertical and horizontal resolution significantly, resulting in an average of 3,397 LW/PH, one of the highest resolutions we've seen yet.

To see what's going on, refer to the plots below, which show the actual edge profiles for both horizontal and vertical edges, in both their original and corrected forms. Here, you can see that there is fairly conservative in-camera sharpening applied (only slight bumps at the top ends, and no noticeable bumps at the bottom ends of the black edge profile curves). Imatest reports that the horizontal direction (vertical edge) is "undersharpened" by 15.4% while the vertical direction (horizontal edge) is undersharpened by 14.5%. Professionals and serious amateurs prefer this to oversharpening, and the A900's images respond very well to the use of strong/tight sharpening post-exposure in Photoshop or some other image editor. (That said, you should be able to extract still more fine detail if you begin with a RAW file, rather than a JPEG.)