Originally posted by Smeggypants Nope, the ISO improvement from K20D to K-5 was just under three stops.
Even the DXO measurements almost back up my real world experience ( showing nearly 3 stops @ ISO100 ), although I claim better performance on my models at higher ISOs.than they do.
K-5 v. K-20D - see Dynamic Range Tab Originally posted by Uluru From DXOMark:
Dynamic range of the K7 (landscape) = 10.6 Evs
Dynamic range of the K5 (landscape) = 14.1 Evs
14.1 – 10.6 = 3.5 EV
1EV = 1 stop.
"Maximum dynamic range is the greatest possible amplitude between light and dark details a given sensor can record, and is expressed in EVs (exposure values) or f-stops, with each increase of 1 EV (or one stop) corresponding to twice the amount of light."
Originally posted by Rondec The question is how much dynamic range the cameras have at a given iso.
Signal to noise ratio and dynamic range are different concepts. SNR is the difference, measured in decibels, between the signal you intended to capture, and the grunge that the sensor added that you do not want in your photo. A doubling of the signal to noise ratio is 6dB, which corresponds with one stop. As I stated previously, the SNR ratio of the K-5, i.e. noise performance, is one stop better than the K20D's.
The Dynamic Range spec does not measure noise. DR is the amount of signal captured, measured in EV stops between the brightest recorded portion of the signal and the darkest.
Capture a raw image, at ISO, of a dark mountain and bright sky using a K20D and K-5. Expose so that the highlights are at the same point in each image, just below clipping. Noise levels are about 1/3 stop difference, not significant. The K-5 will still have recoverable signal (detail) for three stops below the K20D. Let's say you take advantage of the K-5's DR and boost the shadows by 3EV. You will see more of the mountain, but also more noise. Noise will now be at ISO 800 level on the K-5 and ISO 100 level on the K20D. You used the extra DR and it was beneficial, but the unboosted K20D photo now has lower noise, by approximately 1.2 stops (32dB vs. 39.3dB = 7.3dB).
Smeggypants' original statement, that noise is 3 stops better, is overstated by two stops. You can't interchange DR stops and SNR stops. If you push the shadows, you push the noise by an equal amount.