From Pink To Color... Sort Of
Added 2021-11-25 21:41:46 +0000 UTCProfessional color correctors will point out the primitive and inaccurate means that I use to 'correct' pink films after transfer, but hey... I use Vegas and I never listed colorist in my resume. So just enjoy the films - they are free!
Comments
I have been enjoying these.
Steve Whitacre
2021-11-28 16:31:13 +0000 UTCColor Science consists of many dark rooms connected by twisty-turny passages. While doing ultra-high-speed digital video camera development in the late 1990s and early 2000s, I attended the intensive summer program of the Munsell Color Science Laboratory (MCSL) at the Rochester Institute of Technology (RIT). The first thing that amazed me to learn was that (at least at the time) there was no "closed form" model of human color vision! We'd had detailed models of human sound perception for decades, but not for vision. This meant there was still lots of Art in the field of Color Science: Making images that "looked good" to a human often required getting humans to look at them. (That is, we had no algorithm that, when given two images with only slight color differences, could correctly predict which would be perceived as better by a person.) This is part of the reason why audio compression is so much easier than video compression. We have a predictive audio model that lets a computer compare audio codecs, but no such thing for video codecs. Fortunately, human vision sensitivity and accuracy can be measured (Ishihara test and others). Everyone in the class was tested, and it turned out I had the best color vision. I took a set of the tests back to our company and tested everyone I could. It turned out there was just one person with better color vision than mine, and he "happened" to be the "Golden Eye" in our manufacturing QA/QC department who visually checked the calibration of every color camera we made. We also learned we had employees whose opinions on color should always be ignored. Some of them were engineers. Another was in our graphics department. Color Science did have mathematical models that, despite the lack of a full human vision model, still were able to do lots of the heavy lifting. These included color representations (HSV, RGB, CMYK, etc.), color spaces (CIELAB, NTSC, sRGB, Adobe, etc.), color transformations (quite separate from other image processing transformations), and many techniques for color content analysis. Most consumer software only does linear color correction, which, in particular, is poorly suited to fully correcting film scans. Film, being chemical, can exhibit not only fading, but also spectral shifts. And these shifts are often highly nonlinear. But they can be modeled, and the shifts present in any film can generally be reverse-engineered with a high degree of fidelity. I believe only Adobe Photoshop has the capability (via plugins) to make this accessible to advanced consumers. Specialized software, such as ImageJ, allow researchers to dig into the nitty-gritty. Which, of course, I had to do. Our super-duper camera created enormous torrents of video data, for which we built a video management and analysis system. Many of our customers had lots of video and film from other cameras, all of which they wanted to import into our management system. And, of course, also be made compatible with our analysis tools. Which meant we had to make it seem, to the greatest extent possible, as though it had come from our camera. The hardest piece of software in the entire system wasn't in the camera, and it wasn't in the PC-based user interface, nor was it in the video management and analysis system. It was in that blasted external data import system. Which we quickly split off into a separate product that sold for nearly as much as the camera itself.
BobC
2021-11-28 01:25:10 +0000 UTCI've been scanning some 35mm family slides from the 1970s. The Kodachrome seem to have held up well. The Ektachrome are pink, others are fading out completely. I was wondering if others could share their experience. Mine were stored indoors but not in cold storage.
Chris Watson
2021-11-27 01:42:31 +0000 UTCThere are some in the list...
Fran Blanche
2021-11-26 17:02:32 +0000 UTCDo you have any Jam Handy films? My father worked for them in the 1950s. They produced industrial films primarily for Detroit area businesses like General Motors.
RustedDucksDesign
2021-11-26 08:32:16 +0000 UTCI know from my attempts to copy faded viewmaster reels how difficult it can be to get a good colour balance. And that's only doing a single image at a time, not a continuous stream of them. Also the difference in fading of images from the same reel on the same film stock is incredible.
David Peaker
2021-11-25 22:46:33 +0000 UTCI've been loving these. On a similar note, I've been taking poor looking public domain movies on the archive dot org, "color correcting" them the best I can and re-uploading them. It's not as easy as it seems! Big kudos for improving them as much as you do.
Matt Lawhead
2021-11-25 22:38:52 +0000 UTCFascinating. These films take me back to my school days.
William Alsing
2021-11-25 22:31:26 +0000 UTC
