Thank you for your attention to my question.
I am looking for color reproduction.
Now I'm shooting on IQ 280 and H5D-50c
They have a great color, but often I need more accuracy in this
Will I get a more accurate color with the H4D-200 MS?
It's a myth that multishot helps with general/global/overall color accuracy.
Multishot
can help with the color accuracy of specific subject detail in the case that the detail is 1) stochastic and 2) nearly exactly one pixel in size. So for example in a tightly woven fabric with randomly colored threads a multishot camera will more accurately render each fiber if you shoot it such that each thread is ~1 pixel in size. If you are not using a multishot camera, in that scenario, you can simply move a bit closer (such that the detail is more than one pixel in size) and stitch to capture the entire object. We have a free tool called
DT Batch Stitch to aid with that if you are doing it for a batch of different objects. Each of those two approaches (moving the sensor as in a multishot back, or moving the subject as in stitching) has pros and cons.
Multishot also does gain some resolution. You should measure yourself with an
ISA Test Target or other FADGI/ISO/METAMORFOZE compliant target what amount of actual gain it provides, as even the slightest vibration of the camera or subject, or any mis-calibration of the camera can lead to considerably less resolution gain than advertised. Many comparisons you'll find published on the internet are comparing the multishot-mode and single-shot mode of the
same camera, but forgetting that the manufacturers of multishot camera have significantly less-advanced raw processing math for single-shot captures. A more salient comparison would be a best-in-class single-shot raw to a best-in-class multi-shot raw; this will absolutely show advantages (assuming zero vibration,
very good lens, ideal aperture, etc) – just much less advantage than you'd think based on the stated resolution of each.
In the last decade nearly all US institutions of cultural heritage have moved away from multishot cameras, which were, in the 2000s and into the early 2010s, quite popular in those settings. In that time improvements to single-shot resolution and improvements to single-shot raw file processing have made and potential advantage of multishot are nearly always far less important than other considerations.
If you do decide to get a multishot back, especially (though not exclusively) if you're buying it used, make sure to test for the proper calibration of the back and look, at 200-400%, carefully for any artifacts or banding, when using multishot mode.
If you're struggling with color accuracy with your IQ280 or H5D-50c I would suggest reviewing our
DT Digitization Guide for Reflective Materials for a step-by-step workflow for both understanding, calibrating, and empirically evaluating color accuracy. I would also suggest
BasIIColor which is referenced in those guides and the
DT NGTv2 target developed with the Library of Congress. Elements of good color include a good sensor (you have two; the 80mp sensor was a hair better than the 50mp sensor, but both were very good by nearly any measure, meaningfully surpassed only by the 100mp, 100mp Trichromatic and 150mp sensors), good color calibration (BasIIColor or the art reproduction profiles in Capture One CH), and good illumination (strobe is good, better still is a flat spectrum high-CRI, high-CQS LED like the
DT Photon XL). You should get *excellent* color results with your current gear if your illuminant is of high quality and your color calibration process is sound; if you aren't then I'd focus on that rather than a new camera, to improve your color accuracy.
Notably, color itself is a fickle beast. We did a free webinar recently with Dr. Wyble, on the nature of color and color measurement, specifically in the context of art and other CH institutional digitization; you can
rewatch the recording here. The point being that some "problems" with color are inherent in what color is: a sensation of the viewer, rather than an objective characteristic of the object. Objects have spectral reflectivity, which our eyes/brains/minds process into the experience of a color; it's like if you could only hear an overall effect/feeling of a musical chord, and not the individual notes that combine to form those chords. Short of
multispectral-based capture and spectrally matched physical reproduction no single-shot or multi-shot camera can resolve all "color problems" you might run into. Though, to be clear, 99% of the time the answer is not more complex than a basic workflow issue like white balance, even fielding, exposure, color profile, or method of proofing.
You may also find our
DT Digitization 101 Online Class helpful, as this is a class that over 500 staff at Cultural Heritage institutions have taken as professional development. The 201 in-person class can also be recommended. Both of those classes focus more on library/museum/archive staff and the service bureaus that do work for them.
So in summary, I can reduce my advice above to
1) Test for yourself before purchase; do not take my (or anyone else's) word on what results you should expect based on the spec sheets
2) Investigate whether a change in your current workflow (profiling, illumination, even-fielding, proofing methods) can address your color issues using your current camera, before investing in any new camera
Bias Disclosure: DT chooses not to sell any multishot cameras, so must be considered biased.
