[Michael Stirling]

I think that everyone is misunderstanding the problem I’m having. There is nothing wrong with easyDCP’s colour transformation. Putting a rec709 file through the easyDCP chain looks great in the cinema in XYZ space there is no issue there – no video to data level LUT is required. This is now my workflow.

It’s Da Vinci’s rec709-XYZ LUT that looks bad. It does look like it needs a video to data LUT then the rec709-XYZ LUT but I’d have thought that it should also presume the start point of rec709 was video levels and transform accordingly.

This all might have come about because the source material was graded ProRes HQ at video levels and Resolve could be expecting the source to be full range when applying the LUT?

M

[Michael Stirling]

There was a thread about this about 2 months ago which cover it quite comprehensively

net based – search for these…
Warren Eagles’ fxphd fast forward course.
Patrtick Inhofer
Alexis Van Hurkman.

For a course you attend ICA (international colorists academy) do classes.

M.

[Michael Stirling]

The easyDCP manual says it does this:

Is that doing more than the rec709toXYZ LUT in Resolve?

[Michael Stirling]

It does…

That’s what I did for the trailer. I’m going to be remaking the feature DCP using easyDCP for the colour transformation. My question was more does anyone have an opinion why the feature where I used to LUT from within Resolve instead looked so bad.

Thanks,
Michael

[Michael Stirling]

Ha!

Thanks… I did search through the manual for quite a while. Still works the old fashioned way too btw 🙂

[Michael Stirling]

Well, it’s exactly the same in v9 other than that this workflow description has totally disappeared from the manual! Thanks for posting it Greg – I’m not sure I’d have thought to read the v8 manual for this.

MIchael

[Michael Stirling]

I’d definitely be treating the skin and the white separately…

node 1 overall luminance/ contrast balance.
Node 2 balance skin tone (either in a HSL key or not)
Node 3 key white and balance.

I’d do a couple of master grades I really liked – say on a wide and a close up then balance each shot to these as well as the shot either side of it in the timeline so the grade didn’t drift over time. Also, you will probably find that all of the shots aren’t different from each other but rather you have 3 or 4 groups of shots that need grading differently from each other. Once you’ve recognised these and got a shot balanced to the master you’ll get a long way by just adding that grade to the next clip of that sort rather than reinventing the wheel each time.

M

[Michael Stirling]

By fully manual you mean frame by frame right? It was last time I used it (in v7 I think) which was painful.

[Michael Stirling]

It’s also worth pointing out that even when the grade is there things like tracked windows will not track in the handles as you only did the tracking on the section of clip in the timeline. This will make the windowed grade jump so the handles will be unusable unless you go into the master timeline and lengthen the track at each end of the used media.

M

[Michael Stirling]

Hi Peter

Thanks for responding to my forlorn plea and offering to look into it. I just had dive into some pretty fast problem solving here.

I finally tracked it down to the clients drive with a dodgy connection (1 day job, drives arrived at 8:30am – no time to xfer files to my RAID) as I had 2 drives daisy chained in FW800.

It was a hard spot because whatever was wrong with the connection wasn’t enough to kill the drive in Finder so I could still browse the HDD there but it did crash Resolve’s Quicktime decoder so I couldn’t see any Quicktime on my system either which was what led me to thinking Resolve was at fault. On one of the many reboots and fails Resolve gave an error message about the decoder. That helped me figure it out in the end and just about have the project conformed by the time the client arrived (mercifully late)

Cheers,

Michael