Rick Harding & Chris Bone have written a paper on behalf of Sony, proposing a method for full 4k digital intermediate production using an HDCAM SR data pipeline. I will outline the pertinent features in this article.
The basic idea is that you shoot on film and then telecine the film material to HDCAM SR, maximising the dynamic range, cropped to the aperture required (unlike film, HDCAM SR has a 16:9 aspect ratio, like most HD formats, meaning the film must be letterboxed or cropped). The video tapes then serve as the basis for viewing dailies, and offline editing. Now, here’s where the complicated bit comes in: the HDCAM tapes are then used for a pseudo-online edit. Basically, the HDCAM material is conformed and colour graded, as if it were the 4k material (t is possible that the feature is edited offline at uncompressed HD resolution, in which case it does not need to be conformed, just graded). This means that the graded HDCAM material can be used for high-quality screening purposes (and by extension, as video masters, though the paper does not suggest this explicitly). At the final stage, the original film is dug out, scanned at 4k resolution (probably using selected takes from the final EDL), conformed using an EDL, and colour grading applied via a CDL. This then forms the basis of the digital master.
Effectively, the HDCAM SR material serves as a proxy for the 4k film scans until the end of the production, rather than the current trend, which is to use 2k data proxies for the 4k scans.
The benefits should be clear– it is much more convenient (and faster and less expensive) to manage HDCAM SR data than it is to manage 4k scans throughout the pipeline, but there are a few potential problems with this method:
#1 – Dustbusting– both the HDCAM telecined footage and the 4k scans will have to be dustbusted separately, assuming that the HDCAM SR output is used for the video master.
#2 – Risk of damage — The film is effectively scanned twice. Right now, that’s about the same as what most DI processes involve- once to do dailies, once for digitisation. However, one of the advantages of DI is that it potentially allows for film to be scanned only once, particularly if digital dailies are used. With time, I reckon that most filmmakers will become as obssessed with limiting film handling as they are with secondary colour grading.
#3 – Limited options for pan & scan — because the film image is cropped during the telecine process, it is not possible to pan & scan the entire film image from the HDCAM material (unless the footage is zommed in). This means that selected shots must be panned & scanned from the 4k material.
#4 – There is limited support for CDLs in most grading applications, and CDL workflows, although good on paper, have not been reliably used across different systems.
#5 – Cost savings — it remains to be seen whether the money saved is significant in practice, especially considering that the material is either conformed twice or edited directly at HDCAM SR resolution.
#6 – Visual effects — visual effects would probably be done outside of this pipeline, and supplied as 4k data. This means that they must also be converted to produce HDCAM SR copies for inclusion during the conform.
UPDATE: I’m including a copy of the original slide, which should explain the process a little more clearly.
Click the thumbnail to view the full-size image.
Posted: June 23rd, 2005
Categories:
Articles
Tags:
Comments:
9 comments
