I can make a file. Defocused colorchecker or flat area under Tungsten any use to you ?
[a href=\"index.php?act=findpost&pid=167653\"][{POST_SNAPBACK}][/a]
If you want to accomplish a lot in one shot, make sure it has OOF near-black areas, and meter the OOF color-checker with an ambient meter or grey card so that we can get an estimate on what the real base ISO is.
I am looking at your ISO 800 of the woman that you used to show the striations in your first P45+.
Everything seems scaled low for a 16-bit RAW. The highest value is only 9339 out of 65535, and that's the center of the bulb just above the woman's head. I have a feeling that the RAW may actually be just like an ISO 50 or 100 RAW, and that high ISOs actually reflect absolute exposure, just with a metatag to tell the converter to brighten during conversion. The RAW histogram shows most of the pixels less than 4000 and a near-flatline from about 4000 to about 7000 with a few direct lights and hot pixels up there; a single isolated area of specular highlights up to 9339, and nothing above that (empty up to 65535).
I am assuming that the camera leaves the full DR of the sensor in the RAW data at all ISOs. If it is starting at ISO 50, and 50 is indeed what I am looking at in this RAW data, then the read noise is about double the P30 at the same exposure, but then again, I have no way of knowing what the real sensitivity of the RAW data is. It may be that the RAW really reflects ISO 100, and not 50 (IOW, 50 implemented as an extra, low-headroom ISO), in which case the pixel-level read noise would be about the same as the P30 or slightly better, relative to absolute signal.
Anyway, the ratio of max signal to pixel read noise in this RAW is about 2000:1, compared to about 3000:1 for the best case cameras (1D(s)mk2/3, D3, Fuji S*. 2000:1 is about what you get for the Canon 5D at ISO 100. On the other extreme is something like the D2X with about 1000:1 at ISO 100. So, at the pixel level, the DR seems to be good, but not among the very best. Of course, with all those pixels, the image-level DR performance should be excellent. I don't subscribe to the notion that pixel DR directly limits image DR.
Relativistic characteristics like DR are much easier to determine from RAW data than absolute SNRs and base ISOs, as you can see.
The actual exposure index of the camera's lowest ISO is usually not that important to me; I use it mainly because of its DR. When you get to higher ISOs, however, the game changes and absolutes become more important, as apertures and shutter speeds start to become harsh limiting factors and "exposing to the right" is not really a possibility all of the time.
Your camera seems to exhibit a real case of what I have theorized in the past; a scenario where ISO means nothing except a hint for the RAW converter, and any review image the camera may give. The whole idea of exposing to the right for RAWs breaks down with this model, of course, as it is totally meaningless at any ISO other than the base one. The "right edge" of the RAW histogram is the same at all ISOs, apparently.
