Color shift when viewing images full screen in DR 17.2

I'm getting an odd color mismatch between the timeline viewer window when viewed "normally" as part of the edit mode User Interface and when viewed full screen. The image when viewed full screen is much less saturated and a bit lower contrast than the image viewed in the viewer window as part of the UI. The same thing happens in Color mode.

The difference is significant and disturbing, since I don't know which is accurate (and, since it's the same image viewed on the same monitor, there shouldn't be any difference!) Have I screwed up a setting somewhere, or is this a bug? I'm pretty sure this started happening after I updated to Resolve 17.2 a few weeks ago; I certainly didn't notice it in earlier versions.

I'm working with ProRes 422 HQ files on a 1080 29.97 timeline. Resolve 17.2 on an M1 Mac Mini running Big Sur 11.3.1 Monitor is an Eizo CS240 connected via DisplayPort.

Thanks for any advice!

Update - As mentioned above, in color mode if I enter full screen (command-F) the image is desaturated and contrast is slightly lowered. However, I just realized that if I use shift-F to enter full screen mode with the jog controls still visible at the bottom, the image is not desaturated and matches what I see in the timeline viewer window in the UI.

Shift-F is not available in edit mode so this work-around is of limited use, and I still don't know which image is accurate.

This may be related to an existing bug I logged a year or so ago. Are you using an external monitor with a LUT on it?

I'm using an external monitor, but there are no LUTs on it (and no LUTS applied in Resolve).

Also, I'm now realizing the problem appears to be intermittent. Sometimes when I quit and then restart Resolve, the full-screen images created by both shift-F and command-F match the timeline viewer window. Other times, I can quit and restart Resolve and the full-screen image is desaturated. It's weird, and I can't figure out any pattern to this behavior.

since I don't know which is accurate

Neither. For image accuracy, you can't use the GUI monitor. You need an I/O card feeding a properly calibrated external display.

https://www.blackmagicdesign.com/products/decklink

Jim, thanks for your reply.

I'm using an Eizo CS240 monitor calibrated to Rec. 709 using a probe and Eizo's Color Navigator software. I just checked with Eizo, and they said that Color Navigator bypasses the color management functions built into the Mac OS and a separate I/O card should not be necessary to get a clean 1080p signal out of my Mac. Is there some reason specific to Resolve that I would need a separate card?

Incidentally I have no illusions that my system is adequate for professional use; I'm just using it to do a rough grade on a micro-budget documentary. My main concern is that it perform consistently and give me a reasonably accurate image.

Update -
This problem continues, and I've been following up with Eizo tech support. Their best guess had been that there was a glitch in the way their Color Navigator software was interacting with the Mac OS (Big Sur, running on an M1 Silicon Mac MIni). I was hoping that the 11.4 Mac OS upgrade would resolve the problem, but unfortunately it hasn't - it continues to intermittently recur. The problem was present using Resolve 17.2, and remains with 17.2.1

When the problem popped up again a few days ago, I had the bright idea of seeing whether I could record the difference in color by taking screen shots of the UI window and the full screen window. I assumed that the problem was at the display end, and therefore probably would not be visible in the screen shots. To my surprise, however, the color difference DOES show up in the screen shots. Eizo tech support thinks this indicates that the problem is upstream of Color Navigator, and more likely resides in Resolve, the Mac OS, or possibly an interaction between the two. I've attached two screen grabs below. No changes whatsoever were made to the timeline or the image, other than switching to full-screen. A few hours later, the problem came back, only this time the situation was reversed and the UI timeline window was less saturated and the Full Screen image was more saturated. Does anyone have any suggestions?

Because of possible interaction between OS and DR, I second Jim’s advice.
Be careful, though to get an I/O device that is compatible with M1.

Caching and proxies can also throw a wrinkle into the mix, especially with prores on macs. See the big long thread on rec.709-A. This could also explain the intermittent-ness of your issue. I feel like recent versions of Resolve are bad about displaying proxies even if I specifically disable them. Haven’t upgraded to 17.2.1 yet, though. Maybe they fixed it.


Sent from my iPad using Tapatalk

Thanks for the replies!
Uli - I was seriously considering getting a unit such as the BM UltraStudio Monitor 3G, but Eizo tech support recommended I NOT do so - they were concerned that there could be issues between the Color Navigator software and an external video box.

Jason - I wasn't able to find the thread you are referring to, but FWIW I'm not using any proxies. The ProRes 422HQ files are my source material. I do have a few cache files, but they are not where the problem is appearing.

I'm having a similar problem - when I enter full screen for the Resolve viewer the blacks are brighter but it also matches output QT files when viewed outside of DR. It actually looks better. All of my tagging seems to be correct. My external monitor and resolve viewer both have different viewer luts as calibrated.

Any updates on this issue?

but Eizo tech support recommended I NOT do so - they were concerned that there could be issues between the Color Navigator software and an external video box.

That's strange advice, did they give any more specific info on why? I'm thinking of getting a CG319X as an extra client monitor and as I understood it there were no problems of this kind. You can calibrate the internal probe to an accurate external one using probe match, or you simply use other calibrating software with Resolve as a patch generator.

but Eizo tech support recommended I NOT do so - they were concerned that there could be issues between the Color Navigator software and an external video box.

That's strange advice, did they give any more specific info on why? I'm thinking of getting a CG319X as an extra client monitor and as I understood it there were no problems of this kind. You can calibrate the internal probe to an accurate external one using probe match, or you simply use other calibrating software with Resolve as a patch generator.

They did give some more specific information, but I'm not that knowledgeable about monitors and color science and am not sure I can pass it on accurately. My sense, and I may be wrong, is that Eizo is still working out the bugs with how Color Navigator interacts with Mac OS Big Sur and the new M1 Silicon Macs, and their concerns about using external video in/out boxes may be specific to their use with M1 machines.

I'd strongly suggest that you contact Eizo directly with any questions. Their tech support is very responsive and easy to reach - they answer the phone almost immediately. Contact info is here:
https://www.eizo.com/support/

UPDATE MARCH 2022
Just wanted to update and correct some information I wrote in this and previous posts in this thread in case anyone with a similar problem stumbles upon them.

1. The information I was given by the first Eizo tech support person about not using Color Navigator with an external video box such as the BMD UltraStudio 3G was incorrect, or perhaps I misunderstood. I later spoke with a second Eizo tech support person, who clarified that it is generally fine to have Color Navigator running while using an external video box. The only time that the external video box should NOT be used is when using Color Navigator's calibration routines to calibrate an Eizo monitor.

2. Both BMD and Eizo tech support escalated this issue, and I spent several months working with them to try and determine where the problem lay and find a solution. Unfortunately, we never did. Neither BMD nor Eizo had seen a similar problem in the past, which leads me to think the issue is very specific to my set up - some issue between Mac OS running on an M1 chip and Eizo Color Navigator specifically controlling an Eizo CS240 monitor. If the problem also applied to other Eizo monitors, I'm sure that Eizo would have received at least a few other complaints!

3. I did in fact purchase an UltraStudio 3G Monitor box, it works fine with Color Navigator, and it does in fact avoid the problem - color and contrast remain consistent, even when they vary on the UI monitor not connected via the UltraStudio box. So this is in fact a good solution for those with space for a second monitor. Still, sometimes it's more convenient to use a single monitor when accurate color isn't critical, and the arbitrary changes in saturation I was experiencing are annoying.

Right I see, thanks Zack, that makes sense.

I've just updated to the 17.4 Version - seems to be working fine now