-
color space issues 4444 to DCP
Hi
I’ve just graded a film shot CinemaDNG and rendered it out 4444 XQ, video levels, gamma 2.4, rec 709 for a DCP. Viewing the DCP in a theater, it was washed out.
The house making the DCP didn’t set the video level or gamma, their software was auto detecting it at 2.2. We tested it again with a gamma of 2.4, it was better but now it was a bit dim.
It looks like some color space conversion issue. 4444 XQ carries data levels. I forced resolve to render video levels when I made the initial Quicktime from Resolve. On conversion to DCP it seems data levels may be being detected?I went back and ran a Pro Res 422 HQ from the 4444 file, and we will test that. Pro res HQ carries only video levels so there can be no confusion.
Any ideas what is going on here? The DCP software has nowhere to set levels as video or data, i looked at their screen. It just does an auto detect. Is the XQ file generating data levels even though i forced video on render?
I would prefer a 4444 go to conversion, but if only an HQ will convert correctly……. -
6 Replies
-
Or you could output the jpeg2000 directly from your timeline and supply that to the DCP house. I’ve never had issues with correct color space conversion that way. In version 14 software you can set the data rate for the jpeg2000.
I tend to avoid Quicktime based 4444 codecs as there seems to be a gamma identification issue with Quicktime ProRes 4444.
-
When you export to ProRes 4444 from Resolve you need to set the Video Output to Full to get the correct Video Levels. Not sure why this has never been addressed by BM but maybe they will in the future.
\”MY MEDIA/PROJECT MOTTO: If you think three copies of your Media or Projects are enough. Take a moment to place a value on them and then maybe add two more. Hard Drives are now stupidly cheap. A RE-SHOOT AND YOUR TIME AREN\’T.\”
-
just for info: i always master in 4444 and never had a problem! how did your proreshq test look in the cinema?
-
Well it turned out the house creating the DCP had set the gamma incorrectly, and didn’t set the video levels at all, despite the fact that gamma and levels were stated in the slate.
-
They also had their projector doing a conversion from XYZ to 709 at 2.2, which made the image dim. So the projector wasn’t in p3 space, so I was just really seeing a blowup of rec 709. Not ideal.
After re running it, it seemed to match, but I don’t love the fact that I’m taking raw CinemaDNG down to 709 422.I intend to QC it in a theater with P3 space in future, as well as re creating the DCP from DPX. Live and learn.
-
[Sean Sartori] ” I don’t love the fact that I’m taking raw CinemaDNG down to 709 422.”
This is a bit like exporting a low-quantized codec into a higher one in the hopes that the quality will improve, except going the other direction. For the “down to 709” to be valid, you would be assuming that your scene gamut contains values that would be outside the 709 colorspace to begin with. Otherwise, they just remain what they are.
The ability to keep things in RGB, rather than Y’CbCr is not a visual impairment, since human eyesight is not as critical in chromaticity as it is in luminance. It’s how we’ve been getting by for the past several decades, before tankerloads of storage and bitrates came available. But now we think we need the Valdez to deliver a litre of anything… besides even bigger disasters.
jPo, CSI
“I always pass on free advice — its never of any use to me” Oscar Wilde.
Reply to this Discussion! Login or Sign Up