C300: C300 II & 1DX II Color Matching HDMI & Internal

[jcs]

After carefully matching the 1DX II and C300 II via HDMI into a Win10 PC and OBS (Magewell 4 channel HDMI card), I later shot tests using the internal CFast cards at another location, where the WB and onscreen LCDs looked like a good match. Upon reviewing the recorded footage, the colors didn't match at all.

Today I repeated the test using my main 4K editing box and a Magewell and Epiphan AV.io 4K USB 3.0 devices. I used the custom WB systems of both cameras and a grey card to set WB (where WB is based on a live capture/image, not manually dialed in). Light was provided by a Fiilex P360EX (set to about 4100K). It looked pretty good however these custom WB systems don't seem to adjust for WB offset, they just appear to find the nearest Kelvin and call it a day. So I had to tweak the WB Offset on the 1DX II by adding a lot of M (magenta) and a little B (blue) to get them to match via the live HDMI output. Upon reviewing the recorded footage (both 4K, both using my custom Filmic Skin settings (currently dialing the C300 II to match Filmic Skin on the 1DX II)), once again totally different color and gamma. Left side 1DX II, Filmic Skin picture style, right C300 II, close-matching Custom Picture settings:


I was able to relatively quickly/closely match the internal shots to HDMI adjusting WB, Tint, and contrast/curves, however the color shifts required were pretty large.


The lesson is the built-in LCDs and HDMI output do not match the color and gamma of the internal recording systems.

Is this a common issue: LCD & HDMI output is not calibrated to the internal recording system? A slight shift would seem reasonable, however these results are pretty far off.

 

[AndreeOnline]

I personally never use any exotic settings in my recordings. I much prefer to use the 'Daylight' WB preset whenever I get away with it (as I like the natural color shifts to be reflected in the footage anyway). If that doesn't work, I go to a fixed Kelvin number that I sync between cameras. And finally, if I'm in a mixed lighting scenario that I think might exaggerate differences between camera models, I shoot a few frames of ColorChecker, or similar.

I find that with this basic workflow, even if different cameras have different gamma curves and contrast characteristics, I can match them in the timeline. The benefit that I only need to focus my matching in one place and if they match in the timeline it's all good.

And no, I would never use on camera displays or external monitors hanging off HDMI cables as reference when dialling something in by eye.

 

[jcs]

I did the test again, re-performed the custom WB settings (gray card), ignored the LCD and HDMI, and was able to match close enough with minor WB and tint adjustments:


Another issue is I used the 24-70 F2.8 II (with peripheral illumination on) on the C300 II and the 24-105 F4L on the 1DX II (with peripheral illumination turned off: I had turned it off on the 1DX II when I noticed brightness shifts while focusing)). That's the main reason the background looks vignetted on the 1DX II shot.

I'm performing 'exotic' tests as I want to make sure the C300 II's 'exotic' color changes I made work as expected for skin tones and nothing unexpected happens elsewhere. Not sure if this is surprising or not, I'm finding the 1DX II color science being more accurate than the C300 II (after I optimized them both for skin tones).

As for LCD/HDMI color accuracy, I recall someone noticing similar issues with the C200. It would be useful for a future firmware upgrade to allow us to adjust WB+WB-Offset on the LCD, HDMI, etc. These are professional camera systems, I'm surprised WB/calibration is not possible for the monitoring systems, especially when they aren't accurate from the factory. Do higher-end cameras have calibrated and/or WB-adjustable monitoring systems?

 

[jcs]

Interesting- these embedded jpgs went really green on MacBook Pro and iPhone X. 1DX II held up better. Need to check monitor calibrations (edited on Win10).