Color Grading & Finishing

[Tommy]

Do you need to do two separate color grading sessions to create an HDCAM SR master for screening and one for the film-out or will the first color grading session allow for both?

[Kaku Ito]

Panasonic Post Production Studio has been working on Varicam workflow to film out and I had some good long discussion with the manager there about film out for R3D. He mentioned that if we shoot in 2K, then we should edit and color grade using REDCINE and FCP or other NLE, then give them DPX or transfer to D5. After that they can use Quantel iQ Pablo to do the final color grading, then to film out. He said it does take some experiences to preserve the feel of the master version.

So making the long story short, I'd think it takes multiple stages if you are picky about it, or if you ask the film out company to keep the color on HDCAM SR master and they can do that, then you probably wouldn't need the second session, would it?

I haven't finished with the project we are working on, but since we'd spoken to many people so I thought I could share my findings.

[Tommy]

Thanks for the reply! Say you go to a Post House and have the color graded using a LUT with the intention of doing a film-out to a particular film stock. Now, that color grading session would only produce something useful for film-out not an HD master that could be used for screenings, correct? Meaning you definitely have to have a second color grading session if you want an HD master for screenings on your first pass.

[Kaku Ito]

Tommy, yes, I think so.

[M Most]

Thanks for the reply! Say you go to a Post House and have the color graded using a LUT with the intention of doing a film-out to a particular film stock. Now, that color grading session would only produce something useful for film-out not an HD master that could be used for screenings, correct? Meaning you definitely have to have a second color grading session if you want an HD master for screenings on your first pass.

No, that is not correct. Digital Intermediate facilities generally do their color correction from 10 bit log files, with a film preview LUT in place. When the work is complete, the files are used directly by the film recorder. But they are also transformed with another LUT to yield the HD video deliverable, and in most cases today, that is accurate enough for almost any indie feature. In some cases, for studio pictures, a "trim" pass is made on this output, but in other cases, it is unnecessary as the lookup table gives a very accurate translation of the look of the film print on video (both the film and video LUTs are built with a film profile as their target). So basically, the only steps necessary for a video deliverable are a render pass (to bake in the video LUT) and a recording step to get it to HD videotape.

FYI, the same files are also put through a different data transform to yield Digital Cinema deliverables (in the form of a DCDM).

[laguun]

mmost sum up explains it perfectly.
i would like things to add two things:
1) one can also use -one- colorspace-target as reference and translate it to other targets later on. typically 709 to film.
2) Once your master is going into distribution, you are in wild west. consumer displays vary extremly, and so do cinemaprojectors and screens, especially mechanical ones.