Get Pentax 645Z at B&H Photo.
This is a shot discipline and technical execution article that should be assimilated by anyone shooting raw on any brand camera.
This discussion of post-shot and Play histograms on the Pentax 645Z also incorporates the RawDigger histogram and a discussion of color space and gamut and recommended practices for raw shooters.
While this is a Pentax 645Z image, the discussion is useful for any digital camera. And while it is not a field image (landscape or similar), the genesis for this page was observing histogram variances during field shooting; this histogram variance is discussed and shown for this image, but occurs with all images.
The image shown below is perfectly exposed yet the histogram suggests that the red channel is slightly blown; in fact the green channel is most at risk by about half a stop. The discussion explains the reasons and the fundamental algorithmic flaws in virtually all camera implementations of the histogram (for the raw shooter).
Lawrence B writes:
Thank you for your extremely useful article ‘Interpreting the 645Z Histogram’. I believe this is the first time I have ever seen in print reference to this most disturbing discrepancy between the ‘before and after’ histograms as displayed on most digital cameras.
Though I do not own the Pentax 645Z (I use a Nikon D800E), the differences observed have been most confusing, and though one can always check later (via RawDigger) histograms based on the RAW data, this helps one little when out ‘in the field’.
Unfortunately, you didn’t offer an explanation as to why the post-shot histogram differs from the one shown during live view (the ‘play’ version). Both are regrettably based on the camera’s JPEG settings. Shouldn’t their ‘inaccuracy’ compared to the RAW data based histogram be identical? Why is the live view histogram somewhat less inaccurate than the post-shot histogram?
I don’t understand why the industry has been so reluctant in offering a histogram based on RAW data. Photographers have been requesting such an option for as long as I’ve been shooting digital (probably longer). In any case, I am most appreciative that you tackled this disturbing phenomenon of the differing histograms. The tips you offered are indubitably the best one can do under the given circumstances. Many thanks!
DIGLLOYD: Yes, other cameras exhibit similar behavior.
As with science, an observation must come first, but an observation does not produce an explanation. Saying “I don’t know” is often the reality. It’s on my “why” to-do list.
The 645Z was configured to shoot DNG only (not DNG + JPEG), so it cannot be the result of the embedded (within the DNG) JPEG versus a full size companion JPEG.
That leaves a camera processing algorithm, and only Pentax can say for certainty, but a reader out there might have a credible explanation. My speculation is that the Play variant is based on the JPEG embedded in the DNG (since it is clearly in the color space with which the camera is configured, AdobeRGB), and that the quickie post-shot variant is based somehow on the image processing pipeline as it “flows through” and/or on every other sensor line, or some other efficiency optimization.
See also true raw histogram.