DaVinci Resolve Forum
exports darker confusion
exports darker confusion
by Enginn Heima on Apr 4, 2019 at 11:57:43 am

I calibrated my monitor with a x-rite i1 display probe and DisplayCal to rec 1886 tone curve and I'm using the generated LUT in "3D monitor LookUp table" under color management and monitoring via HDMI on a Decklink card. Furthermore, my project is set to rec709 2.4 gamma.

Now, when I export (AppleProRes 444) and play it in Quicktime player or VLC player, both players give crush my blacks more than I'd like. I've tried to screen my files on different systems, more or less calibrated and the result is always the same, I'm losing too much detail in my blacks as opposed to my color grading view inside Resolve.

I know that QT and VLC players aren't the exact color science winners, but I'm thinking that maybe I'm doing something wrong here. If anyone has any idea to throw at me, I would be much obliged.

Regards


Re: exports darker confusion
by George Dean on Apr 4, 2019 at 3:22:45 pm

Enginn, just a thought, if you change the output in BCM to Rec709 2.2, render, then what does it look like in VLC?

Best Regards......George


Re: exports darker confusion
by Michael Gissing on Apr 5, 2019 at 12:18:57 am

I would first try other codecs and wrappers. ProRes4444.mov has in my experience been unreliable in replaying correct gamma in various players.


Re: exports darker confusion
by Enginn Heima on Apr 5, 2019 at 7:47:52 am

Michael, do you have any codecs to suggest?


Re: exports darker confusion
by Enginn Heima on Apr 5, 2019 at 7:46:48 am

That's actually one of the first thing I tried and go figure, I don't see any difference when I render out in 2.2. Can that be normal?


Re: exports darker confusion
by Andi Winter on Apr 5, 2019 at 3:00:48 pm

prores4x4 is a totally fine mastering codec. never had any issues.

did you leave data levles at auto during import? during export? i would suppose there is
a problem. i never touch those and leave them at auto.

if you are in a non color-managed-project, it is totally normal that the timeline gamma
doesn't do anything. if you render out dcps the timelinegamma will be taken to linearize
and put the 2.6 dci gamma on top, so, in that case it would make a difference.


Re: exports darker confusion
by Enginn Heima on Apr 5, 2019 at 3:50:56 pm

It is actually in "Davinci YRGB Color Managed" mode with timeline set to rec709 2.4 and output color space set to rec709 2.4


Re: exports darker confusion
by Marc Wielage on Apr 6, 2019 at 4:29:54 am

Be sure to read this, which covers some of the problem:

https://jonnyelwyn.co.uk/film-and-video-editing/colour-management-for-video-editors/

Basically, things will look different in VLC, in QuickTime, in different browsers, in different OS's. In a color-managed playback situation, you can generally predict how it will look if it's going to a calibrated display. Without that, it's the Wild West. I draw comparisons to sound mixes: they sound different on small speakers, big speakers, earbuds, TV sets, laptops, and theaters. Same mix, totally different sound in each.


Re: exports darker confusion
by Joseph Owens on Apr 6, 2019 at 7:18:45 pm

Video Levels vs. RGB. Read all about it. Range issue.

Gamma is not involved.

Look at your signal measurements. If all that is reported is "gee it looks darker", then you haven't looked at your levels.

ProRes 4:4:4 can contain either RGB or Y'CbCr coded media, and it will make a sometimes arbitrary decision (wrong almost 100% of the time) about your video range. If you are converting from a Y'CbCr origination, it is pretty much certain you will get a range error when attempting to render out to a 4:4:4 container.

jPo, CSI

"I always pass on free advice -- its never of any use to me" Oscar Wilde.


Re: exports darker confusion
by Enginn Heima on Apr 7, 2019 at 3:29:00 pm

Thank you Marc for the Article!

And Joseph, any suggestions about codecs? You think a ProRes 422 is more likely to be my friend? I will do a test...


Re: exports darker confusion
by Marc Wielage on Apr 23, 2019 at 9:38:15 am

[Enginn Heima] "You think a ProRes 422 is more likely to be my friend? I will do a test...
"

Joe is correct (as always). One key idea is always to export a short segment of SMPTE Color Bars prior to program and check that on scopes. If the export is identical to what you see in Resolve, then you have a color management problem and the playback engine (like QuickTime) is misinterpreting the image. This is discussed in the article I linked to.

Neither 422 or 444 is really any better or worse -- they both can be misinterpreted under some conditions, as can DNxHD and DNxHR.





© CreativeCOW.net