RGB and YUV
If I were to finish on HDCAM SR and had two masters, one RGB and one YUV, I would use the RGB master for the film-out and the YUV master to create various video deliverables such as Digibeta, BetaSP, and DVD copies, correct? Thanks, Ryan
The only reason to master as RGB is if your entire post process (or at least the color correct and beyond) was done as RGB. Converting it at the end gains you nothing. A 35mm negative or a Beta tape can be easily generated from either.
And oddly enough with SR, if you go with RGB at the standard 440Mb data rate, you actually have to apply more compression as RGB contains a third more data that a 4:2:2 sampling scheme. Because of this, converting to RGB at the end could technically yield a lower quality master.
However, one reason to have two separate masters is if one was to be a 'broadcast master' that is limited for safe video levels. This may be an unwanted constraint for a film master. Further, anything intended for film out that is over 20 minutes should be broken up into reels. You can't deliver a 90 minute tape for film out and you can't deliver a program for air broken up across five tapes.
Thanks for the info-very helpful. So, would it be better to color grade a film shot on the Red One Camera using 2K DPX files and then create all my video deliverables from that (by converting to a YUV color space I assume) or should I color grade using 2K Prores HQ files which would simplify the creation of my video deliverables. If I work from the DPX I'll be able to use those same files when I go to do a film-out (though there will have to be another color grading session). If I work with the Prores footage the creation of the video deliverables will be easier (and I'll save money on storage) but I'd have to go back later anyway and create DPX files when the movie is color graded for the film-out, correct? Thanks again, Ryan
Okay, that just got a whole lot more complicated. I cannot speak in anyway to the truth or accuracy possible through a Final Cut finish. Also, DPX and RGB (as used here) are not the same thing. If you can really do a DPX in to DPX out conform and color correct, then I would deliver for film out as DPX files. Everything with Red is a multistep process with a lot of back and forth. The flexibility is great, the process is not. Just dealing with the issues of Log and Linear and creating the correct type of DPX file seems like a job in itself.
However you end up working, I would, as you stated, do color for film first and then work from that to do an adjustment pass for video (if even needed). You can master for video in whatever way best suits your needs.
Can't really offer any more detailed advice, the Red thing started almost completely unworkable and continues to change all the time as the kinks get worked out. I haven't tried too hard to keep up as it rarely impacts me. What made sense 2 months ago may seem silly 2 month from now.
Best of luck.
Just so I can wrap my head around this... Why would a post house provide you with both an HDCAM SR 422 YUV and an HDCAM SR 444 RGB master? How does the difference in color space affect the image and what purpose would each master serve? I really appreciate your help! Ryan
They might suggest both just to make sure you are covered. Not all SR decks can play the 444 tapes. Or as I mentioned previously, it might have something to do with the idea of 'broadcast safe' levels where the 444 master is unconstrained and the 422 master is NTSC safe? You should ask them.
Working at RGB 444 is technically higher resolution as it allows for a third more samples per line. YUV (lets say component instead as YUV is a very specific subset that doesn't apply to HD) is designed for data efficiency while taking advantage of the way our eyes work by preserving the luminance detail but limiting the chroma detail. The limitations were much more significant back in the analog composite days. Modern component (specifically designated YCrCb) digital systems preserve much more information than would be possible in NTSC which not only limits detail but also the range of possible colors. Except of course, many HD formats apply further limits in the interest of cost efficiency over technical accuracy. This is seen in the subsampling of formats like HDCam and DVCProHD and even more so in the 4:2:0 scheme of HDV or XDCam HD. In fact, only HD-D5 and HDCam SR are true full raster 422 formats. Oy.
RGB is fully unconstrained and is most useful for situations such as heavy color correction or complex effects and compositing where a lot of heavy math could lead to unwanted distortions and artifacts. And then there is DPX which stores information as logarithmic rather that linear. This was derived from the Kodak Cineon format. I'll provide a link rather than try to explain it:
And then there is still bit depth and how that effects both media and processing. No easy answers.
I'm trying to determine how deeply (and where) to look into all this information as I'm the editor and post supervisor. So, if I ask the post house to do an online using DPXs for an HDCAM SR master as well as a downconvert to HDCAM and DVDs for festival screenings I assume the HDCAM SR master could be used for the film-out (but not for screenings). I know it's probably better to use the DPX files off the hard drive but it seems, from what I've read, that many people and post houses just do the film-out from the HDCAM SR master. Any thoughts? Now for the big question... how deep should I journey into this bottomless pit of information? I assume as much as possible but I just want to make sure that I'm at a level where I can communicate the needs of the film and deliver the product. I truly appreciate your assistance Dino. Ryan
I think the biggest determinant is cost. Staying DPX throughout the process should be more expensive as it requires more of the equipment and the people using it and demands a level of expertise greater than that for video. There is certainly no reason that a great looking piece can't be delivered on SR. I mentioned delivering files for the film out only because if you are sitting there with a finished piece that can easily be output as files, it saves the time and effort of generating and checking tapes and removes the same issue at the film end. Probably more of a personal preference developed from dealing with finicky machines. The last time I delivered for film out, I made uncompressed QuickTimes of each reel. Not because it was the absolute best thing to do, just good for me in that situation.
Even skipping the DPX route and doing a more video style finish can still produce beautiful pictures. At the end of the day, mastering to SR is definitely the smart thing to do (whether RGB or component) as it represents a great way to archive all your hard work. Depending on what sort of equipment you have access to regularly, an HDCam or even DVDProHD dub (or even just as a file) could be good to have for generating different deliverables.
What leaves me unable to be more specific is all the uncertainties surrounding Red and what is the best way to handle it and what capabilities the people you are working with have. I really think you should be expecting answers from your post house. You are hiring them to be the experts. Maybe concerns over money have stalled the conversation? If you are not satisfied with their answers, it may be time to find a new partner.