Do I need a Resolve LUT?

Home Forums General Discussion Do I need a Resolve LUT?

This topic contains 22 replies, has 4 voices, and was last updated by  Florian Höch (@fhoech) 1 week, 4 days ago.

Viewing 15 posts - 1 through 15 (of 23 total)
  • Author
    Posts
  • #20078

    Blake T (@blake-t)
    Participant
    • Offline

    Sorry if this is a stupid question but someone once told me there’s no such thing as a stupid so here goes… lol… ok so do I actually need to generate a LUT for resolve after calibrating my external monitor? And what is the purpose? I make up that it’s so that the GUI in the resolve timeline matches closer to the calibrated external monitor. Is this what it’s for? And is it essential? Seeing as I won’t be relying on my GUI on my laptop screen for grading. Or does it serve another purpose?

    #20089

    Florian Höch (@fhoech)
    Administrator
    • Offline

    Hi,

    do I actually need to generate a LUT for resolve after calibrating my external monitor?

    How is the external monitor calibrated?

    And what is the purpose?

    A 3D LUT simulates a certain target (e.g. Rec. 709 with BT.1886 tone curve) in the actual color space of your monitor.

    #20114

    Blake T (@blake-t)
    Participant
    • Offline

    I calibrated the monitor (Eizo CS2730) using displaycal with the “video (D65, rec. 1886)” preset in the settings tab and using a Spyder5 calibrator.  With no corrections selected.

    Ok thanks for explaining that! I thouht the LUT was was to make the viewer on the resolve GUI (in my case on my laptop) match the external monitor calibration..

    #20115

    Florian Höch (@fhoech)
    Administrator
    • Offline

    How is the external monitor connected? Are you using a DeckLink card or similar? Or is it your main desktop/only monitor?

    #20116

    Blake T (@blake-t)
    Participant
    • Offline

    It’s connected only with a HDMI cable. I’m not using a decklink card or similar because with the latest Resolve Beta update it appears only a HDMI cable is needed to output a fullscreen viewer to external monitor… But am still waiting for someone from BMD to confirm. I’m a bit confused as I never had an external monitor before the resolve beta update.

    No, this isn’t my only monitor I’m using my MacBook Pro laptop for resolve and the external monitor for resolve full screen viewer. Am I missing something crucial here?  Love to know

    #20122

    Gene R (@gene-r)
    Participant
    • Offline

    Hopping into this convo as well since I’m in the same boat. I have a BenQ PV270 monitor. I used the factory Rec.709 2.4 gamma D65 profile on the monitor and then calibrated it using Palette Master,  creating a new profile. The new profile was aiming for Rec.709, 2.4 gamma, D65, 100cd/m2. The profile was made while the monitor was connected directly to my iMac via display port. In Resolve the connections are iMac display port out to BlackMagic UltraStudio Mini Montior to HDMI out to the monitor. So all that being said, should I still create a Resolve LUT?

    #20127

    Florian Höch (@fhoech)
    Administrator
    • Offline

    Resolve can use display profiles, but its support for accurate profiles is lacking. Use a 3D LUT instead. If your monitor is part of the normal desktop, follow the color viewer guide in the 3D LUT creation workflow for Resolve wiki entry, otherwise the pattern generator guide.

    #20141

    Gene R (@gene-r)
    Participant
    • Offline

    Got it. So I took the advice from a user on LiftGammaGain and made a new icc profile for my moniter setting everything to it’s native setting (color, gamma, white point). From there I made two 3D LUTs with DisplayCal using the recommended settings on the wiki and aiming for Rec 709. BT 1886. D65. 100 cd/m2. For the most part both results seem to pass the verification charts but there seems to be a blue bias in the lower grey values. I feel like I’m doing something wrong. Looking at a grey ramp with either LUT loaded to my monitor setting in Resolve, the biases are pretty apparent. So I don’t know what I’m doing wrong. You can see the measurements below.

    https://drive.google.com/open?id=1DvA7Qc3U4EUIxx5qrvwKhYivnyAE7vo8

    #20142

    With LCD displays, blue blacks are the norm.  To create a neutral black, one needs to raise the black level significantly so that R=G=B.  And this will look like a washed out display.

    Fortunately, near blacks read so dark, that we don’t usually perceive the bluish blacks when color correcting.  Just watch your scopes to make sure your blacks are near neutral when grading.

    Also, as noted in the responses in LGG forum, try avoiding the REC 1886 tone curve, as this raises the level of near black tones and will make the bluishness more obvious.  And… lead you to crush your blacks in the grade to over come this 1886 tone curve.

    #20146

    Blake T (@blake-t)
    Participant
    • Offline

    Resolve can use display profiles, but its support for accurate profiles is lacking. Use a 3D LUT instead. If your monitor is part of the normal desktop, follow the color viewer guide in the 3D LUT creation workflow for Resolve wiki entry, otherwise the pattern generator guide.

    Thanks, Florian.

    Turns out I was missing the ‘Ultrastudio Mini Monitor’ hardware which is why I was so confused about needing the 3D LUT or not. I just didn’t understand the overall workflow… So its making more sense now. But I do now have some new questions about correctly creating and installing  the Displaycal 3D LUT vs my external monitor settings vs Resolve settings if you don’t mind having a look at my settings below?


    Software: DisplayCal, DaVinci Resolve Studio
    Hardware: Apple MacBook Pro 2017 > UltraStudio Mini Monitor > HDMI cable > Eizo ColorEdge CS2730
    Colorimeter: X-rite i1 Display Pro

    DisplayCal Settings:

    Video 3D LUT for resolve (Rec.709 / Rec.1886).
    Everything left as default except for
    Correction preset: Spectral: LCD GB-r-LED IPS (Dell U2413)
    Interactive display adjustment: checked
    Whitepoint: Chromaticity coordinates
    White level: custom 120 cd/m2
    Tone curve: Rec. 1886

    EIZO CS2730 Settings:

    Signal Settings:
    Input Color Format: AUTO (other options available are: ‘YUV 4:2:2’, ‘YUV 4:4:4’ & ‘RGB’)
    Input range: FULL (other options available are: ‘auto’, ‘limited 109% white’ & ‘limited’)
    Custom Settings:
    Temperature: N/A
    Gamma: 2.4
    Color gamut: Native
    Gain: RGB sliders adjusted during calibration
    Brightness: adjusted during calibration

    DaVinci Resolve settings:

    GPU Configuration:
    GPU Processing Mode: auto (other options available are: OpenCL, CUDA, Metal)
    Use Display GPU for compute: unchecked
    Use GPU scopes: unchecked
    Looks like it has chosen “GPUs – Video Card 1 AMD Radeon Pro 560 – Metal”

    Video and Audio I/O:
    For capture and playback use: UltraStudio Mini Monitor
    For Resolve Live use: none

    Video Monitoring:
    Video format: HD 1080PsF 25
    Use 4:4:4 SDI: checked
    Use Level A for 3Gb SDI: unchecked
    Use dual outputs on SDI: unchecked
    SDI Configuration: Single Link
    Data Levels: Full
    Video bit depth: 10 bit
    Monitor Scaling: Bilinear
    Use Rec601 Matrix for 4:2:2 SDI output: unchecked
    Enable HDR metadata over HDMI: unchecked

    Color Management:
    Davinci YRGB – Unmanaged
    Timeline Color Space – REDWideGamutRGB/Log3G10
    (fyi …timeline workflow last node transforms to Rec.709)

    —————————

    My questions are:
    1. is the correct correction for my EIZO CS2730 monitor “Spectral: LCD GB-r-LED IPS (Dell U2413)” ?
    (EIZO CS2730 specs are:
    Backlight: WLED
    Panel Type: IPS
    SRGB: 100%
    Adobe RGB: 99%)

    2. the warning that appears before saving the 3D LUT: “The display device’s video card gamma tables 1D LUT calibration is non-linear… Make sure to manually reset the video gamma card tables to linear before using the 3D LUT, or create a 3D LUT without calibration applied…” Does this mean resetting the Eizo video card?  Or the computer running resolve (my MacBook Pro)? Or should I re-create the 3D LUT without the calibration applied? And by doing so does that bypass the adjustments made by DisplayCal during calibration? Not sure what to do here…

    3. Am I correct in having the EIZO monitor input range set to “full” before calibration? other options available are: ‘auto’, ‘limited 109% white’ & ‘limited’…

    4. Am I correct in having the Resolve settings data range as “full” before calibration?  The other option is  “video”…

    5. Are there any other settings above that don’t look right?

    Also apologies for all of the questions, I’m still learning about all of this… and many thanks

    Best,
    Blake

    Attachments:
    You must be logged in to view attached files.

    i1Display Pro on Amazon  
    Disclosure: As an Amazon Associate I earn from qualifying purchases.

    #20158

    Florian Höch (@fhoech)
    Administrator
    • Offline
    1. Yes, this Eizo has a GB-r-LED backlight.
    2. This warning can be ignored for a video monitor that is not part of the desktop.
    3. & 4. Monitor input range only needs to be consistent with Resolve’s “Video Monitoring” levels settings (so “full” – “full” or “limited” -“limited” are both correct).
    4. Looks good to me.
    #20166

    Blake T (@blake-t)
    Participant
    • Offline

    Thanks so much, Florian!

    Almost there. Just a few more questions if you don’t mind…

    1. I forgot to mention that I did not do any prior calibration to the Eizo monitor through “Color Navigator”… I don’t even have it installed. Does this matter? I see in a lot of the forums that people say that they had to calibrate through Color Navigator prior to creating the 3D LUT in DisplayCal. Do I need to do this?
    2. Should I have the “clipping” setting on the Eizo monitor “on” or “off” before creating the 3D LUT?
    3. Should I have the “due priority” setting one the Eizo monitor as “uniformity” or “brightness” before creating the 3D LUT?
    4. I noticed that you mentioned to another Eizo user to change the curve in the 3D LUT tab to 100% output offset instead of BT.1886. Should I do this too? And if so is it just for the 3D LUT tab only? or for the Calibration tab also? Here is a link to that forum: https://hub.displaycal.net/forums/topic/another-confused-davinci-resolve-user-eizo-cg2420/
    5. Should black point compensation be ticked?
    #20175

    Florian Höch (@fhoech)
    Administrator
    • Offline

    I forgot to mention that I did not do any prior calibration to the Eizo monitor through “Color Navigator”… I don’t even have it installed. Does this matter? I see in a lot of the forums that people say that they had to calibrate through Color Navigator prior to creating the 3D LUT in DisplayCal. Do I need to do this?

    There’s no need, although it probably wouldn’t hurt either unless it would restrict the monitor gamut.

    Should I have the “clipping” setting on the Eizo monitor “on” or “off” before creating the 3D LUT?

    I don’t know what that does.

    Should I have the “due priority” setting one the Eizo monitor as “uniformity” or “brightness” before creating the 3D LUT?

    Personally I would rank uniformity higher than peak brightness.

    I noticed that you mentioned to another Eizo user to change the curve in the 3D LUT tab to 100% output offset instead of BT.1886. Should I do this too?

    Personally I’d do for any display below (roughly) 2000:1 native contrast, i.e. most if not all IPS computer monitors.

    And if so is it just for the 3D LUT tab only?

    Only 3D LUT tab. Keep calibration tone curve “As measured”.

    Should black point compensation be ticked?

    No.

    #20218

    Blake T (@blake-t)
    Participant
    • Offline

    Thanks for the help, Florian.

    Ive spent the last week creating LUTs for Resolve and comparing the results to make sure everything is good before I start grading. So sorry in advance if any of these questions have been answered before or are in the manual… I’m just trying to understand it further from my own trial and error since I’m a beginner and don’t know all the scientific terms just yet.

    1. When watching the progress dialogue after the patches are finished, I notice it says “there are 2 rev cache instances with 3382 Mbytes limit” and “there is 1 rev cache instance with 6764 Mbytes limit”. Is this a problem?

    2. When opening the validation reports, by default the “use absolute values” box is unchecked. However when creating the LUTs most of the time the default setting on the tone curve is “absolute”. So does this mean to be reading the report correctly, I should be checking “use absolute values” whenever “absolute” has been applied during the LUT creation? And should it be left unchecked if “absolute” has not been used during LUT creation?

    3. Similar to the previous question, Is it better to use the “XYZ” gamut preview since the 3DLUTs default is an XYZ space?

    4. My EIZO monitor covers 99% of Adobe RGB. So in the advanced profiling settings, should I leave source profile to the default “sRGB IEC61966-2.1” or change it to “Interchangable with Adobe RGB 1998” ?

    5. In the “self check report” I notice some dialogue that says “Channel 1, 2 & 3 unique values 256 @ 8 bit”. My monitor offers 10 bit so does this mean it read my monitor wrong and not using the 10bit correctly?

    6. Before hitting the start button I always make sure to uncheck the “black point compensation” box, but after completion it is always re-checked by default. Is this supposed to happen? Just thought it was a bit odd as its the only default setting that changes itself back to default…

    7. In a previous answer to one of my questions you told me to leave the calibration tone curve to “As measured” even though my monitor hasn’t been previously calibrated. So am I right in thinking that my monitor still gets a calibration (by applying the LUT) because of the way that the LUT simulates a certain target within my monitor?

    8. Should I use trilinear or tetrahedral interpolation when applying the LUT in Resolve?

    9. By comparing my verification reports created over the past week I feel really confident that I have created a 3D LUT that I can work with in Resolve when it comes to balancing primary colors and working with saturation etc… but not so much in terms of working with light, so I want to be sure that when working between 0-1023 on the waveform, that it is accurately represented on my Eizo monitor with 0 being pure black and 1023 being whitest with all the mid tones scaled and spaced out correctly in between. I have tried both “absolute” and “relative” custom gamma 2.4 with output offset set to 100% (Calibration tone curve left “as measured”).. but the results looks almost the same when comparing the Gamma diagrams on the verification reports. I also tried the same settings just mentioned but with the rendering intents changed to “luminance preserving perceptual appearance” and “luminance matched appearance” but still no improvement. I have read that “absolute” and “relative” is an output at 50% input. I Don’t exactly understand what this means. Does it mean the 3DLUT only contains 50% of the gamma curve? And is the other 50% supposed to be the calibration part to make up 100%? Because this would explain a lot since my calibration tone curve is left “as measured”…

    Sorry if i’m sounding a bit naive, especially with that last question! And thanks again for your time.

    Blake

    #20221

    Florian Höch (@fhoech)
    Administrator
    • Offline
    1. That’s normal progress information. It’s a bit verbose and not that useful to the end user though.
    2. The meaning of “absolute” on the reports is different. Normally, all values are adapted to a common whitepoint (D50), which enables comparing whitepoint delta E independently of the rest of the patches dE. When you check that box, the whitepoint difference becomes part of the overall dE (unless “use display profile whitepoint as reference white” is checked).
    3. L*a*b* provides a representation that is easier to understand, because the neutral axis is centered and hues are spread in a more perceptually uniform matter around the neutral axis. This is independent of profile type.
    4. This is only relevant if you want to create a “perceptual” gamut mapping for applications like e.g. Photoshop, and only makes sense if you monitor gamut is smaller than AdobeRGB and you want to compress a larger source gamut into the monitor gamut (no clipping)..
    5. This is just informational and always uses 8 bit for the calculations. It is completely independent of actual display bitdepth.
    6. If you are under macOS, this depends on the profile type you are using. The default macOS profile type curves + matrix should always use BPC, since some macOS applications (Preview, QuickLook) only support such profiles. This is not relevant for applications that support LUT profiles though (e.g. Photoshop) or use 3D LUTs.
      Edit: I realize it might be a bit unexpected having to re-disable BPC when you already had it disabled. I’ll change this behavior in the next update for consistency sake, so that it only gets re-enabled when choosing another (non 3D LUT) preset.
    7. Yes.
    8. There should be little to no difference with large LUTs (e.g. 65^3). Leave at default.
    9. When using 100% output offset, it is normal that there is no conceivable difference between “absolute” and “relative” gamma. The 50% target means that if you aim for a certain gamma value, the logic will match the value at 50% input and adjust the other values accordingly. If you set output offset to 0% for testing purposes, you will see the difference between “absolute” and “relative” more clearly (i.e. at relative, even with 0% output offset the effective gamma at 50% input will still be the chosen target gamma, while with absolute it will be lower by a certain amount depending on the display’s black level).
    • This reply was modified 2 weeks, 4 days ago by  Florian Höch.
Viewing 15 posts - 1 through 15 (of 23 total)

You must be logged in to reply to this topic.

Log in or Register

Display Calibration and Characterization powered by ArgyllCMS