Originally posted by beholder3 One needs to be wary that photonstophotos website is a hobby project of one single amateur internet guy - like you and anyone else - and shady with regards to robustness as well very questionable with regards to interpreting what is being presented.Dxomark still seems much better and "better" is coming from pretty low standards.
I disagree with the conclusion. The process on Photons to Photos is absolutely transparent and repeatable, errors, sample variation impact discussed and published. The PDR definition makes some assumptions, which make comparisons only applicable across similar sensors, definitely not extending to processed signals. And that's what's expressed in what luftfluss quoted from Bill: "Except between ISO 100 and ISO 160 there is very strong Noise Reduction (NR) baked into the raw data even with NR set to off.These results are not comparable to any other cameras." - PDR is not a universal measure of image quality, not even for the single aspect of dynamic range.
Originally posted by Kunzite Unfortunately, photonstophotos proves most useful to people always looking to bash Pentax (on the other site there are some new names "interested" in the K-3iii but always having bad things to say about it), and not at all or very little to Pentaxians.
Definitely not true for me, and things I have shared with others - you can definitely learn something about cameras and what settings to use for things that you cannot immediately judge visually - e.g. at what ISO to shoot astro shots, depending on how you intend to process them. Or explain how those 'empty' low ISO settings of other cameras may help to achieve low shutter speeds, but otherwise don't necessarily improve image quality.
Also the K3 III curve seems to indicate something interesting: Comparing the K3 III PDR with the Nikon D500 and Sony alpha 6400, I suspect there could be a similar underlying dual-gain design, boosting the ISO400 PDR higher than ISO100 by the combined effect of the extra gain stage and accelerator image processing. Of course, a switch in parameters/algorithms
could have the same effect, but shots at ISO400 may come out marginally better than ISO320 ...
I really like working with the 'cooked' raw files from my KP, less work in most cases than having to optimize noise post-processing myself and likely better optimized than I would achieve in the time I'm willing to spend on it. I can only see very niche cases where K-1 style, largely unprocessed raw files would provide an advantage - such as specific model-based astro processing. But in that case, other specialized cameras are available by now at reasonable prices, more suited to the job. For my use, they're mainly providing me a way to balance tones and colours without penalties, some level of noise reduction upfront isn't going to impact that. As I expect that Pentax has further improved processing over the KP, I'm looking forward to see results in difficult light from the K3 III.