Joseph Owens

Members
  • Posts

    9
  • Joined

  • Last visited

Everything posted by Joseph Owens

  1. The S-Curve. Up until about 15 years ago, everything or about 80% of it was stock negative-origination, exposed for digital scan. Most cinematographers still wanted a "print" look as a final result, but the hand that was supposed to go in that glove was the transfer characteristic of the electronic scan - CCD array or photo-multiplier tubes. Eastman and Fuji did provide targets for us -- I suppose that was our first "LUT" that was what the manufacturer thought that a certain stock would do if printed '25-across'. Which would be where we tried to place one-lights for dailies so that the dP could get a fix on where their exposure was, until instructed otherwise. So whatever density was there in the negative wended its way through the matrices that dropped out the other end as if it was printed to 2386 stock or similar. The relationship is not one-to-one, because video transducers just don't have the horsepower or range of a theatrical projector and the aim is to have a bright picture and opaque blacks -- after all it is a shadow-and-light process. You are right it is a big challenge to synthesize that experience with a completely different technology -- an emissive screen on top of all that. A counterfeit-currency enforcement officer was once asked if they spent a lot of time looking at bogus money. The reply was that it was exactly the reverse, and they spent most of their time looking at the real thing. The fake ones are easier to spot when you know what you are looking for. I spent so much time looking at camera neg that my native "look", if colorists have a "fingerprint" winds up being *filmic* -- more often than not, when I am left alone to deliver a grade, the comment will come back that "it looks like film." I don't do anything special at all... yes, very careful contrast placement, *rich blacks* *graceful fall-off*.... if you want harsh, we can do that, too. Very often trade comments around so-called "LUTs" -- really, every correction is a look-up-table. You take a set of RGB values and re-map them to another set of numbers. That's all they do, that's all we do. Artistically. jPo, CSI
  2. If you subscribe to an "apprentice" training approach, think about how colorists were introduced and progressed into the craft while the position was being developed. Before you were allowed into a suite with a paying client -- and touch the controls -- for the first couple of years you were responsible for loading the film, fetching craft services, and listening carefully. And keeping it zipped. After that, depending on the facility that was essentially fostering/subsidizing the training, a promising candidate who seemed responsible enough to make a decision, not pirate the footage, or destroy it, was allowed to join the "dailies" fraternity. Churn through thousands of feet (every night for a few more years), sync sound, log camera reports, spot issues, report sensitometry... using RGB only, no secondaries. One-light, best-light.. timed dailies... on-time, under-budget. Figure it out... now. And have an answer ready for when the phone call comes in from the DP later that day, while you were out shopping for groceries after 8-10 hours of midnight shift and a 3 or 4-hour nap. Kind of like ER intern, except with fewer fishhooks-in-the-eyes and gunshot wounds. Then you are assigned to transfer some lab-timed IP's... should be one-lights, but because we are now looking for perfection, usually some scene-to-scene trims. And if you can get through the feature in a couple of days and no melt-downs, great, candidate is approved to progress to some soft-ball no/lo budget student/indie shorts... and that's where a candidate starts gaining some traction... the palette starts to develop, a flair for a bit of innovation may emerge, some word-of-mouth starts getting around... something that happened at 3 AM, 4 years ago happens again and you just deal with it on the spot... a client starts thinking you are some kind of wizard. There you go, because you are. Current grade/colour applications are so highly evolved now, that it is very difficult to isolate fundamental principles of the craft. In the end, it is still one of the most fascinating areas; a curious blend of artistic and technical manipulation. It is far and away not just a matter of being able to achieve a white balance / grey-scale knob/ring solution with a video game controller. A sensitive colorist eventually develops a hue/contrast understanding of imagery the way a skilled musician can blend music theory -- mixing major and minor scales, improvising rhythms, adjusting tonalities -- that will affect an audience on a much deeper level than what they see and consciously process. jPo, CSI
  3. This is not how it is supposed to work. You can reassign the IDT per source clip, so if a composite comes back to you, then assign a 709 IDT to it, and no further conditioning should apply. It might go crazy (will go crazy) if the "Project" IDT (which might be Alexa) is assigned by default. jPo, CSI
  4. This is the second iteration of my master suite, its built into my home, based on what would have been a cozy man-cave (fireplace included -- electric insert). The large monitor is an LG 55" that is starting to get some accuracy now that I have been refining its Expert Settings to not be as far from the Flanders as it used to be (this photo was made over two years ago)... Lights aren't usually up as much as they are in this pic. The redwood panelling is over-emphasized as a result. If you scrutinize the waveform monitor YRGB parade, you can see the immense amount of noise in the blue channel -- I shot the image at the F1 GP Indianapolis 2002 with some re-canned 250D Fuji @ 60 fps, which need to be rated at least another stop slower. And it was a rainy day first day of practice as it turned out. This position is at the so-called "Mickey Mouse" pair of reverses at the end of the infield section leading back out onto the oval, and consequently the slowest section of the circuit. Great for getting some detail in the cars! The film was scanned on a Cintel Ditto, which is now the "Black Magic Cintel Scanner."
  5. If this was hanging driver-side doors on new Camaros coming down the Chevrolet assembly line, you could say 20 seconds per clip, do the math. And wouldn't that be great? We used to be able to plow through a semi-timed IP, feature length, marking events (Auto-detect and fix) 4 or 5 assembled lab rolls in about 20 hours or so. Edit density was low compared to today. DaVinci could store up to 999 events in a list. I am looking at a feature open here now, 4.5K resolution, mostly R3D source, 1,781 events. I lose about 2 minutes every half hour to auto-save. Don't know why its dragging -- oh, wait a minute, ACES IDT/ODT transforms, NR, 4.5K dpx reference stills, tracked, animated secondary qualifications, smart-cache... Because I am re-creating a mismatched lighting setup in many scenes (editor went for story and not necessarily lighting continuity in master scenes) there is a large amount of synthetic foreground/background colorist-roto going on, so I have no way of estimating how long this project is going to take. Although the DP took a great deal of care on-set, that work was undone by shuffling the chronology of the shot order. But all at national commercial quality standard, of course. So when Saatchi & Saatchi stop by and use up a week deciding on 8 shots for a 20-30-second spot... with a 4-figure/hr star facility -- that's for re-grading every pore on your 6-figure/hr supermodel. Expectations really have to be managed because you cannot multiply that by 200 and hope that it even fits in the schedule. jPo, CSI
  6. There were a number of book ideas in play... this was a direction I had that would separate the activity of creating a grade away from the standard approach of "how to work the software" to achieve it -- no knobs, balls, effects... how to think your way into the scene, establish the stage, then maintain it. Not "I am a Camera"... "I am a Colorist" :-)
  7. There is a companion to Alexis' COLOR CORRECTION HANDBOOK, which is the LOOKBOOK. I have seen Itten's treatise here and there, but although the principles don't change, its kind of dated, like Bauhaus philosophy... and then you might as well read Goethe. "If It's Purple, Someone's Gonna to Die" by Patti Bellantoni is a fun read with lots of input from Deakins BSC among others. Its a no-button-pushing insight into color design. As much about Art Direction as anything. There is a curious little book called "The Colour Eye" produced by BBC Books authored by Robert Cumming and Tom Porter as a print companion to a series produced by Suzanne Davies and Julian Stenhouse with BBC2 (1990). I'd spend some time with https://colormax.org/color-blind-test/ Ishihara... <winky> There are always more publications in development... stay tuned. Would anyone be interested...? I set up an outline about 6 years ago and never followed up on it to do a sort of thought approach -- a board-less mental chess game working through the process of grade, using a Sontag-like no-image essay that subtracted all the knobs, rings, styluses, mice... basically no software involved whatsoever...
  8. Gordon Willis wouldn't have been pleased. As far as vision improving, my eyesight "improved" by almost a diopter after moving to a 23" HD monitor. jPo, CSI
  9. Painterly advice is worthwhile. Some other names that are more closely associated with an orthodox "Fine Arts" approach are Cezanne, Seurat and delaCroix. They got very analytic with the deployment of pigment and deliberate theories with respect to chromatic contrast and composition. jpo, CSI