Resolution Chart Comparison (JPEG and RAW)
Images on this page are of our standard resolution chart which provides for measurement of resolution up to 4000 LPH (Lines Per Picture Height). A value of 20 equates to 2000 lines per picture height. For each camera we use the relevant prime lens (the same one we use for all the other tests in a particular review). The chart is shot at a full range of apertures and the sharpest image selected. Studio light, cameras set to aperture priority (optimum aperture selected), image parameters default. Exposure compensation set to deliver approximately 80% luminance in the white areas.
What we want to show here is how well the camera is able to resolve the detail in our standard test chart compared to the theoretical maximum resolution of the sensor, which for the charts we shoot is easy to work out - it's simply the number of vertical pixels (the chart shows the number of single lines per picture height, the theoretical limit is 1 line per pixel). Beyond this limit (when talking about line pairs usually referred to as the Nyquist frequency) the sensor cannot faithfully record image detail and aliasing occurs.
This limit is rarely attained, because the majority of sensors are fitted with anti-aliasing filters. Anti-aliasing filters are designed to reduce unpleasant moiré effects, but in doing so, they also reduce resolution (the relative strength and quality of these filters varies from camera to camera). In theory though, a sensor without an AA filter, when coupled with a 'perfect' lens, will deliver resolution equal to its Nyquist limit. Therefore, even though it may be effectively unattainable with normal equipment in normal shooting situations, an understanding of a sensor's theoretical limit provides a useful benchmark for best possible performance. Nyquist is indicated in these crops with a red line.
On this page we're looking at both JPEG and Raw resolution. For a (more) level playing field we convert the latter using Adobe Camera Raw. Because Adobe Camera Raw applies different levels of sharpening to different cameras (this confirmed) we use the following workflow for these conversions:
- Load RAW file into Adobe Camera RAW (Auto mode disabled)
- Set Sharpness to 0 (all other settings default)
- Open file to Photoshop
- Apply a Unsharp mask tuned to the camera, in this case 100%, Radius 0.6, Threshold 0
- Save as a TIFF (for cropping) and as a JPEG quality 11 for download
|JPEG (4928 x 3264) 4.4MB||RAW (4928 x 3264) 4.2MB|
There is a clear difference in detail resolution between the D5100's JPEG and RAW output. Whereas in the JPEG files, the D5100 cannot accurately describe the 9 lines on our test chart much beyond 2600LPH (roughly), the RAW file still shows all nine lines distinctly at 2600LPH, and they only begin to merge at around 2800LPH. This is still some way off the Nyquist limit (equal to the number of vertical pixels in the D5100's output - i.e. 3264LPH) but impressive nonetheless.
Also worthy of note is the fact that even beyond Nyquist, some line detail is still visible in the D5100's RAW file. Of course this isn't 'accurate', (and represents merely how good the camera's demosaicing algorithms are at 'guessing' what lies beyond Nyquist) but it enhances the impression of fine detail. At the same point on our chart the JPEG file shows only a monotone gray mush. Comparing results from the D5100 against those from the higher-end D7000, the two cameras give close to identical performance - not surprising given that they employ the same Sony-manufactured CMOS sensor.