Originally posted by ElJamoquio I never found their test procedure. I will continue to 'diss' IR as long as they have results that don't match up to the rest of the world.
OK I found some more info on how they test.
They discuss most of their test procedures here, inc resolution tests:
Imaging Resource Digital Camera Test Methods
and here:
Digital Tips - The Imaging Resource Finding the Right Digital Camera Article
I think a relevant excerpt is:
"At its most extreme, you can use the ISO 12233 target to extract a full "spatial frequency response" (SFR) curve for a camera/sensor system. While this is absolutely the most accurate and comprehensive test of a camera's resolution performance, it's also by FAR the most time-consuming: Computing the SFR requires capturing 9 separate images of another, gray-scale target, measuring and averaging the gray-scale tonal response of the camera, then using that information to modify the tonal balance of the resolution target itself. Finally, a Photoshop plug-in can be used to extract the SFR data, which in turn can be loaded into a spreadsheet program for graphing and display. We figure this would take us at least another 2-4 hours per camera to execute, and just don't have the time to devote to it.
Fortunately, your own eyeballs can tell you quite a bit about how well a camera does with this target, just by looking at the output. The number values next to the resolution wedges refer to resolution in units of line pairs per picture height. This is an excellent, consistent measure of camera resolution, but the term undoubtedly requires a little explanation for the uninitiated.
The dilemma for the standards-makers was to define a resolution measurement that would apply equally well to cameras with a variety of sensor sizes and different height-to-width ratios. The goal was to express resolution relative to the total image area, not pixels, since the number of pixels involved could change significantly from camera to camera. Rather than expressing resolution as a number of pixels, the standards committee decided to measure resolution in terms of the number of pairs of black/white lines across the image area that the camera could distinguish. To avoid confusion with cameras having different width/height ratios, the "fineness" of the line pairs was expressed in terms of how many of them would fit across the picture from top to bottom. Thus, the term "line pairs per picture height," or "lp/ph." (Note though, that the reference to picture height only refers to the size of the lines, not to the direction the measurement is being made in. That is, even when resolution is being measured along the long axis of the camera's frame, the results are still expressed in lp/ph.) When the target is properly framed in the image, the numbers adjacent to the resolution elements indicate the pattern "pitch" in hundreds of line pairs per picture height. (That is, "5" on the target means 500 lp/ph.)
Confused? Look at the adjacent (to the right and below) resolution target clips: Both show the same resolution, the one on the right showing how well the camera does in the horizontal direction, while the one below shows the results in the vertical direction. (Note that we're concerned with how many lines the camera can resolve across the pattern. As a result, the "tall" pattern shows resolution in the horizontal direction, and the "wide" one in the vertical direction.)"
tl;dr - they just look at the images on their ISO 12233 test target and visually read the resolution off the resolution target clip indicators.