Use of magenta filters
Posted 23 September 2010 - 04:36 AM
Would like to know if anyone has an opinion about the use of magenta filters to correct the green cast of the log uncompressed signal, I've heard that it not only color corrects the image, but it also increases the dynamic range and reduces noise.
Would like to know if anyone has used this filters and what correction factor to use in different light situations (day, night, interiors)
Posted 23 September 2010 - 07:41 AM
Workflow depends to an extent on exactly what data-security requirements are required by your insurance and completion guarantee contracts. The usual Viper workflow is to:
- record the unadulterated camera output onto some sort of recorder
- make redundant backups of that recording, which probably comes as part of the recorder rental in your case
- make offline versions of the recording, almost certainly applying a LUT to approximate your intended final grade. This can get complex depending on what you're doing the offline on and whether there are additional requirements for DVDs and so forth.
- conform the picture, then final grade.
It is possible to do all of this on anything from a desktop computer (with a US$500 RAID in it) right up to a full bore post facility ending up on a Baselight colour grading system. Where and when you do these stages, whose responsibility they are, and what technical approaches are used for each, are currently fairly poorly standardised. Storage and HD edit gear is now sufficiently cheap, if you're canny, that you might even choose to cut the uncompressed material, but it depends on what quantity of rushes you anticipate and that's manly done for shorts. It does obviate a big lump of complexity at the offline stage, though.
The chap who shot Benjamin Button (possibly the most natural-looking digital movie ever) has been known to post here; he'll have more practical advice.
Posted 02 October 2010 - 06:14 AM
Posted 02 October 2010 - 08:14 AM
What it doesn't take into account is the fact that the different channels have markedly different sensitivity curves, noise floors, and other variances. People have done tests on this and arrived at various conclusions; I suggest you do the same.