Overall, I think DPR published a decent review report here.
My main critics would be that they still base their DR measurements on JPG which means that they measure the non-linearity of the applied tone curve only.
(linear RGB is limited to 8.0 EV, linear sRGB to 11.5 EV; a non-linear curve is an arbitrary choice by the vendor as the sRGB
standard does
specify the tone response curve).
Others and my own lab tests have shown that the current best breed of APS-C sensors are capable of 13 EV DR, something DPR cannot measure anymore using their approach.
It is therefore particularly bad when they write
With our dynamic range test setup we could not trigger the shadow expansion functionality and therefore we can unfortunately not provide any measurements for this feature.
The K-x shadow expansion functionality is fully useable even in 3/3 setting and with highlight expansion enabled. If a DR test is only measuring the tone curve then it is a joke not to use the tone curve meant to provide the highest DR.
IMHO, they have to redo the "Expanded Dynamic Range Function" paragraph with highlight
and shadow expansion enabled. I don't see why they "could" not trigger the function. It is a simple menu option.
I expect several stops more DR coming from enabling the option. It can have a pretty strong effect.
Originally posted by Erik (K-x vs D5000 at ISO 3200)
These are accompanied by a text that says "In the RAW comparison the K-x and Nikon D5000 are quite close in terms of both noise and detail (although the Pentax output is better by a small margin)" How is this a "small margin?" At all? Nikon is getting MURDERED here.
My own tests indicate that DPR is correct in what they write. Don't know why their sample images don't show it more clearly.
Moreover, both DPR and my own lab tests show that Pentax RAW have NR applied for ISO 3200 and higher which D5000 doesn't seem to apply. Without this additional NR, DPR would have measured identical black/gray RAW noise for D5000 and K-x, not only up to ISO1600 but for the entire range.