Michael Most

Oops, sorry, slight typo. That should, of course, be 5400 feet/hour. The total is actually $324,000. Much less than $330,000.

Okay. 2.5 hours x 5500 feet/hour = 13750 feet x 16 frames/foot = 220,000 frames x 1.50/frame = $330,000. You really have a lot of strange ideas as to where technology is today in the real world of film production/post production. Unless, of course, you also are independently wealthy and actually have a spare $330,000.00 sitting around for a personal project.

Yes, but out of necessity. Features are not television, and digital intermediate is not television post production. There is little difference in cost and no difference in time between transferring on a telecine to SD or HD. Tape stock costs are about the same, as well. So for television, it is economical and sensible to do an HD transfer of all dailies (print takes, in the case of most scripted productions), assemble from those sources, and tape to tape color correct for the final product. With digital intermediates, it is not practical (at the present time) to scan all print takes at the daily stage, so video is used as a substitute for editorial purposes. For the final finish, the scanning starts once the picture is locked (or close to locked), which keeps the cost and time factors for this step sensible. The final timing is still done from assembled material, which allows for continuity color correction as I previously mentioned.

Personally, I think a "retransfer selects" approach is insane for anything but a commercial, and even then, I question its value. You can't color correct properly without continuity, and you only have continuity when the piece is assembled. The only tricky part of finishing via a tape to tape correction is the need to have your dailies transferred within acceptable bounds so that you have enough latitude to do what you want to do. Other than that, it's pretty straightforward and very efficient. Essentially every network television drama is done that way, and I think for the most part, network dramas look pretty darned good.

Yes, you were somewhat wrong. But you're missing my point. At your level, you should be asking questions here rather than answering them. I could only dream of having access to some of the people who post here when I was starting out, and if I had that access, I would have used it to expand my knowledge base rather than try to prove how much smarter I am than they are. Those of us who actually work in the film industry for a living easily see through the facades of those who don't, and a little humility often goes a very long way. Good writing and spelling skills don't hurt, either.

John, you don't really believe that Panasonic is going to cede the 4:4:4 market to Sony, do you? Have you forgotten that D5 started out as a standard def only format? The HD version was originally an add-on, outboard chassis that connected to the original transport. Although I don't have any hard proof, my guess would be that you will see a 4:4:4 enabled D5 format within a reasonable time frame that will be backward compatible with current recordings, the same way the 3700 (I think that's the model number of the current version, with additional audio tracks) is backwards compatible with the 2700 (the previous format with fewer audio tracks). Unfortunately for those of us who have to deal with it, I also believe that we will never see a single, universal format standard in the video world in my lifetime.

Then why are you answering his question, particularly when there are a multitude of others here who have the experience to properly answer?

The only step in a 35mm film production that will cost you .20 per foot is a daily print or a telecine transfer. Television programs, which can sometimes shoot 30,000 feet in only one day (multiple cameras in a courtroom, for instance) and who often shoot over 100,000 feet per episode (i.e., over 2 million feet in a season!) pay at least double that for raw stock, and that would be a very, very good price. Besides, the real cost of shooting 35mm on a per foot basis is more like .60 per foot (.45 for stock, .15 for development - you develop everything you shoot, even if it's not printed), and that's without creating dailies, which will cost at least .20 per foot regardless of whether they're film or video dailies.

I certainly hope not. They are two different positions with two completely different skill sets and mind sets. With the possible exception of Robert Rodriguez, I don't know of any cameramen of skill and reputation who don't appreciate the skills of a good lab timer or electronic colorist, and I also don't know of any who think they can do the job as well or better - or would even want to try. A cameraman's skill set is conceptualizing lighting design and camera framing and movement in an artistic manner that's complimentary to the story being told, then translating that into technical terms in order to create proper exposures on either film or video, thus capturing the desired look. The colorist's skill is accurate interpretation of the images that have been recorded on either film or video, and on occasion, artistic enhancement of those images through various tools available in the color correction suite - all in collaboration with the cameraman, the director, or both. And, in the case of commercials, about 14 other producers, agency types, and hangers-on. One job is creation, from scratch, of an image. The other is enhancement of something that has already been created. Like I said, two completely different skill sets and mindsets.

Compared to what? Discreet Lustre? That's about the same price. DaVinci? A 2K costs over $350,000. Nucoda? About $170,000 last time I checked. Someone mentioned Baselight, whose cost I'm not completely sure of but I can assure you it's nowhere near $20,000. None of these systems sell in the hundreds, let alone thousands, of units. They require tremendous software development expertise (and hardware as well, in the case of DaVinci) and a tremendous amount of support. They must be accurate and reliable, not to mention versatile enough to be useful in a constantly changing segment of the industry. Silicon Color is a startup company that has some very smart guys working for them, and they sell a turnkey system that includes a very good control panel (the Tangent Devices panel) and all necessary monitoring, as well as storage that has been tested and proven. Someone mentioned that it was "available" for the Onyx 4, but that is not the case. They put the SGI version (they were attempting to build a 4K capable system) on the shelf before any units were delivered, and even now, I only know of a few sites that have the Macintosh product. It is a nice product, but without name recognition, it's difficult to justify placing it in a high end DI facility over, say, Lustre - which has the Discreet name, a proven track record (both as a Discreet product and as the original 5D Colossus), and more features, such as full conforming capabilities (Silicon Color's product does not have conforming tools of any sophistication) - for almost the same price.

The picture was NOT post produced in 4K. While it's true that the work at Efilm for the DI was done using 4K scans and was recorded at 2K, the fact is that nearly 60% of the film is visual effects shots - which were all done at 2K, NOT 4k. The pipeline was scanning at 4K, down-rezzing to 2K, all visual effects work (CG and compositing) at 2K, and upconverted to 4K again at Efilm. This is hardly "4K postproduction," and I really wish that those who claim it is would be more forthcoming.

Lookup tables convert one set of values to another based on calibration and conversion parameters. In the case of digital intermediate work, there are two places you can use LUTs, input and output. The output LUT is normally used for the display, to account for the differences between your display device (i.e., either a CRT or a digital projector) and the final output device (i.e., print film of a particular type). The output LUT is always the same because it's there to make what you see look as close as possible to the intended film print, adjusting for gamma differences, color differences, and approximating colors that are out of gamut (i.e., colors that print film can display but a CRT can't). In other words, it's a system adjustment based on the particular viewing device and the environment it's in, and what type of print stock you're going to be using. The input LUT, if used, converts the scanned log format film frame (almost all scans are done in 10 bit log format) to linear colorspace to allow "video style" color correction. Systems such as the Discreet Lustre or the Baselight system can color correct in either log or linear space, so in those systems an input LUT is not always used. Frankly, I don't really see where either LUT would need to be adjusted in any way for a 16mm original, because the color corrector doesn't know or care what format the material was shot on. It is working from digital files, and any calibration adjustments are presumed to have been made when the scanning was done.

Yes. approx. $220,000.00 Still think it's interesting? Just because something runs on a desktop platform doesn't mean it's intended for home use. As an example, Discreet Lustre runs on Windows XP and Avid DS Nitris runs on Windows as well. Not to mention the latest version of Discreet Smoke, which is available under Linux running on - yes, a PC.