Expected Performance
3 min
emulsion agnostic profiles previous targets for film scanning were printed on photographic film in so doing they were bound spectrally by the three color layers of that specific emulsion, and color profiles made with them would only be applicable to scanning film of that emulusion for example a provia target could be used to calibrate a scanner for scanning provia film, but results for scanning ektachrome film would be poor the dt ntpt is designed using colorants that are not based on film emulusions the profiles it creates are not bound to any specific emulsion and should perform well across for the digitization of a variety of photographic materials hardware limited performance calibration can only maximize the performance within the capability of the hardware being used to digitize the film; no target can overcome low performance for example, using the target with a low quality light source (as measured in cri, or preferably cqs) will improve the color accuracy of the system, but absolute performance will still be subpar metameric failure some film emulsions use colorants whose spectral properties are tricky for the three colors of a modern bayer sensor camera, regardless of the quality of the illuminant that is, the camera sensor experiences metameric failure compared to the human eye in seeing the colors of these films while a good calibration based on a good target can improve the color performance for such film it cannot fully resolve this metameric failure for those looking for precise color with such film the dt fusion technology in the dt stellar captures color in six channels across two rapid captures for higher and more robust color quality for such materials flare one of the largest challenges and limitations to camera based digitization of transmissive materials is flare the target should be tightly masked in order to reduce the chance of flare note that patch a9 on the dt ntpt can be used as a flare detector – it should register as close to 0 luminance neutral black as your digitization system's noise floor allows flare should be evaluated using lab values without a custom profile applied, as some profiling systems will try to account for flare by raising the black clipping point or otherwise altering the tone scale of the system while this forces the target numbers to "look better" it is not a good solution for flare, the source of which should be found and eliminated rgb is not well suited for checking for flare as flare below 1 rgb value can still be problematic in high quality of high contrast film where meaningful subject matter detail can also carry tone values below 1 rgb value (and later made visible by dodging, exposure adjustment, hdr shadow adjustments or similar) the 0 255 range for rgb is a human friendly 8 bit scale, while the underlying data is 12 bit, 14 bit, or 16 bit so between "0 rgb" and "1 rgb" there are between 16 and 256 levels of tone for each of the three color channels for a total of up to 17 million unique colors as a point of reference, using the dt film scanning kit with the dt 4x5 magnetic film carrier on a dt atom with an ixh 150mp with no color correction profile applied the a9 patch registers as between l =0 00 and l =0 01 the same system with a naked backlight registers the black patch as around l =0 1 (roughly 2 rgb values) our https //heritage digitaltransitions com/product/digitization workflows transmissive pdf download/ discusses this topic in greater depth most camera systems that record 16 bit files produce nowhere near 16 bits worth of data per color channel see our article https //heritage digitaltransitions com/in depth on bit depth/

