- This topic has 3 replies, 2 voices, and was last updated 10 months, 2 weeks ago by
.
-
Posts
-
Hi everyone.
Here is my setup
An EIZO CS2400S for a budget reference monitor.
An I1D3 prope with Stuart’s EDR for the monitor.
I’m using this with my GPU, not with the Decklink because the monitor has already burned through my budget.I calibrated this monitor using CN7, and to my understanding, it now holds an internal 1D LUT and an icc profile to describe the behavior to color managed softwares like Photoshop and Lightroom.
Since Davinci Resolve ignores icc profile. I presume my calibration now relies solely on the internal 1D LUT. Would this mean I get some degree of color accuracy when using this with the video clean feed from DR or do I still need and external 3D LUT box for 3D LUT calibration workflow?Thank you.
Calibrite Display Pro HL on Amazon
Disclosure: As an Amazon Associate I earn from qualifying purchases.CS monitors and CG2420/CG2730 do not have 1dlut but a 1d lut- matrix – 1d lut.
An 1dlut would be like GPU, to calibrate greyscale only, for example like NEC EA models.
A lut-matrix-lut kike yours allows to mix primaries and simulate ideal colorspaces that fall inside your monitor gamut, like sRGB/Rec709, but this simulation assumes that if you mix widegamut primaries to get sRGB primaries, all colors from primaries to black will behave in an ideal additive way.
A LUT3D like in CG2700-X does not assume that and allowes you to aoptionally correct not only primaries but all other colors with a 17x17x17 mesh.Now, if you use CN7 to calibrate to Rec709 D65, 100nit and 2.4g (it’s just an example) it will simulate that colospace based only on primaries mixing… and usually this simulation / HW calibration is very good. Calibrate to that target in CAL1 for example, then validate ressults with DisplayCAL. If this is good enugh to you (expect ALL colors <1.5dE), use Resolve as is Windows, not LUT3D. To validate in DisplayCAL such HW calibration you can set as simulation profile Rec709, 2.4 gamma relative black output 100%, use simulation profile as display profile, then run “measurement report”
If you are in macOS you can calibrate to native gamut 2.2g and let Resolve use macOS ICC profiles, instead of simulating Rec709 in HW.
If your results simulating Rec709 with CN are not good enough for whatever your requirements are, then do the LUT3D. But CN7 rec709 calibration should be very good.
CS monitors and CG2420/CG2730 do not have 1dlut but a 1d lut- matrix – 1d lut.
An 1dlut would be like GPU, to calibrate greyscale only, for example like NEC EA models.
A lut-matrix-lut kike yours allows to mix primaries and simulate ideal colorspaces that fall inside your monitor gamut, like sRGB/Rec709, but this simulation assumes that if you mix widegamut primaries to get sRGB primaries, all colors from primaries to black will behave in an ideal additive way.
A LUT3D like in CG2700-X does not assume that and allowes you to aoptionally correct not only primaries but all other colors with a 17x17x17 mesh.Now, if you use CN7 to calibrate to Rec709 D65, 100nit and 2.4g (it’s just an example) it will simulate that colospace based only on primaries mixing… and usually this simulation / HW calibration is very good. Calibrate to that target in CAL1 for example, then validate ressults with DisplayCAL. If this is good enugh to you (expect ALL colors <1.5dE), use Resolve as is Windows, not LUT3D. To validate in DisplayCAL such HW calibration you can set as simulation profile Rec709, 2.4 gamma relative black output 100%, use simulation profile as display profile, then run “measurement report”
If you are in macOS you can calibrate to native gamut 2.2g and let Resolve use macOS ICC profiles, instead of simulating Rec709 in HW.
If your results simulating Rec709 with CN are not good enough for whatever your requirements are, then do the LUT3D. But CN7 rec709 calibration should be very good.
Thanks Vincent. I forgot to mention that I’m in windows and yes, I’m calibrating to rec709 gamma 2.4.
I just did a measurement and it seems the result are pretty good considering the monitor isn’t exactly high end level. I’ve attached the report.A side question if you don’t mind. I opened some stills from Davinci in Photoshop to compare and notice they lose quite a bit of contrast. The same also when opened in other icc aware softwares like Lightroom or Capture One. However, images posted online seems to have the same level of contrast I see in Davinci. This only happens to stills eported from Davinci, everything else seems to be behaving normally.
The project those stills came from were in rec709 gamma 2.4, color settings in Photoshop was set to sRGB iec61966.The project those stills came from were in rec709 gamma 2.4, color settings in Photoshop was set to sRGB iec61966.
You need to reencode Rec709 g2.4 content in sRGB, otherwise won’t match. You cannot use direct RGB number copy.
Example, assuming 0-255 levels for simplification. If RGB content in Rec709 2.4 is RGB 150,150,150, and that color has some color coordinates, the same color coordinates in sRGB will have different numbers, since gamma is lower something like ~130,130,130 (this values are not accurate, just a qualitative example)
An alternative take is to copy RGB numbers from Resolve to Photoshop, but insted usig sRGB as image colorspace, use a image profile that matches Rec709 g2.4.
IDNK how do you export content from Resolve to sRGB .jpg, but it seems that this reencoding stage is not working as expected of that you had some misconfiguration in Photoshop or in the way you configure jpg export in resolve.
Also remember that you cannot change colorspace in Eizo when Photoshop is running. On photoshop startup it will read default display profile in WIndows color management for that display, and such profile must match the CALx in display, all the time. If you change Cal1 to Cal2 photoshop won’t know you changed it, unless you or CN7 assign the new cal2 profile to display in windows (CN7 does that) and you restart photoshop so it will notice the ICC profile change on Windows.
