LG GL850 B and 10 Bit colour mode in Windows

Home Forums Help and Support LG GL850 B and 10 Bit colour mode in Windows

Viewing 15 posts - 1 through 15 (of 36 total)
  • Author
    Posts
  • #36236

    mat a
    Participant
    • Offline

    Hi

    I own an LG 27GL850-B

    Is the 10 bit colour mode implemented well enough to leave it on and not worry about it?

    I use Photoshop/ gaming/ 3D software.

    I read that it creates noises in dark areas due to it not being a native 10 bits (8bits + FRC) mode.

    What do you recommend? Is the image noticeably poorer in practice?

    Thank you.

    #36241

    Raj S
    Participant
    • Offline

    It’s generally recommended to use the native bit depth of your panel (8-bit in your case). But it’s personal preference.

    You can always try 10-bit and see if you notice the dithering (pixels changing colours quickly to blend shades). Unfortunately, most displays use temporal dithering – the worst method that makes things look noisy.

    If banding is a problem, set your GPU to 8-bit and use the Novideo_SRGB tool to enable dithering for your GPU (Nvidia only, AMD is already enabled by default). It lets you choose different spatial dithering modes which perform much better than temporal. The good ones are: Static, Static 2x, and Dynamic. I recommend trying Static 2x or Dynamic with your panel.

    • This reply was modified 1 week, 6 days ago by Raj S.
    • This reply was modified 1 week, 6 days ago by Raj S.
    #36246

    Vincent
    Participant
    • Offline

    Hi

    I own an LG 27GL850-B

    Is the 10 bit colour mode implemented well enough to leave it on and not worry about it?

    I use Photoshop/ gaming/ 3D software.

    I read that it creates noises in dark areas due to it not being a native 10 bits (8bits + FRC) mode.

    What do you recommend? Is the image noticeably poorer in practice?

    Thank you.

    If you can use 10bit in some apps, use it, not because it makes a difference on an SDR display (SDR contrast I meant), but because lots of applications truncate their output to GPU driver bitdepth in an ugly way.

    Photoshop is an example:
    If you use LR or Capture One or even ACR in PS, these apps do their stuff and calculate high bitdepth RGB data on display colorspace but when dealing with GPU driver set at X bitdepth (usually 8) they dither, so truncation is done in a nice way and no banding is showing (unless VCGT calibration induced banding in nvidias and iGPUs and all that stuff).
    Same applies to madVR or DWMLUT LUT3D correction, done at high bitdepth,  then dither to X bit ouput. Nice way.
    Photoshop don’t dither, if your GPU driver & photoshop config is set to 8bit, it truncates in a simple & ugly way. If your GPU OpenGL driver allows to use 10bit hookup (even if actual GPU ouput is set to 8), Photoshop truncate to 10bit, hence not loosing as much info as in 8bit mode.
    This is important: Photoshop won’t care about your GPU output or display capabilities. It only wants a 10bit hook on Open GPU driver. If it has that and 10bit/30bit is enabled, no ugly truncation. GPU openGL driver can use that 10bit data and output 10bit or dither to 8 when link to display does not allow 10bit… it does not matter in SDR contrast range.

    TL;DR Photoshop 10bit oputput is not about more info to screen but actually to avoid ugly and simple data truncation caused by Photoshop itself due its lakc of tmeporal dithered output.
    Unfortunatelly Indesign and Illustrator which are more prone to use synthetic gradients do no not have 10bit OpenGL ouput (AFAIK unless they added it in latest versions). They dont dither..so ugly truncation is always present.
    Hence if display is good enough to be described by a Simple curve + matrix profile… USE IT, don’t use XYZLUT profiles since they will store 3 different TRC that even with a slight difference in them (that you do not see in not color managed gradient) will cause truncation issues and colorarion of grey scale gradients.
    Of course if you want a LUT3D , profile in XZYLUT to make LUT3D or if your display has poor response and cannot be modeled by a matrix profile.
    But if your display behaves in a good way, use Simple curve + matrix profiles for Photoshop, Indesign , Illustrator, GIMP, Krita…

    #36251

    mat a
    Participant
    • Offline

    Thank you for your reply.

    #36252

    mat a
    Participant
    • Offline

    Thank you.

    I need to try and understand your post!! 😉 It is a bit above my skill level.

    This is my monitor: https://www.rtings.com/monitor/reviews/lg/27gl850-b

    And I use a X-Rite i1 Display Pro

    I have calibrated my monitor in the past with Displaycal but now just uses the iprofiler software. I also use no_srgb to force sRGB (it is a wide gamut monitor).

    If I was to calibrate using Displaycal, do you know what correction I should use for this monitor? Best settings? I am not using the wide gamut, I prfer sRGB

    You wrote:

    ‘Hence if display is good enough to be described by a Simple curve + matrix profile… USE IT, don’t use XYZLUT profiles since they will store 3 different TRC that even with a slight difference in them’

    What is the difference between Simple curve + matrix and XYZLUT?!

    Thank you.

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

    #36253

    Vincent
    Participant
    • Offline

    Xrite profile type <-> DisplayCAL profile type:
    matrix <-> matrix + single curve : display is described by its primaries and overall grey gamma
    table <-> XYZLUT: display is described by a 3D mesh of X nodes (amount of profilling patches) and 3 separate gamma, one per channel
    DisplayCAL has more combionations or profie tyoe, but regarding i1Profiler these are the equivalences of the two types yiu can choose in Xrite software.

    These 95% P3 displays usually are “WLED PFS phosphor” in its cheap variant (no AdobeRGB, not full P3). Correction for this bLED backlight and i1d3 are covered by Panasonic VVX17P051J00.ccss/Panasonic VVX17P051J00.edr and its available in DisplayCAL and i1Profiler/ccProfiler.
    The review you link lack of information (no SPD plot) hence saying its a PFS phoshor LED it’s a guess

    Xrite software make profiles with white point defined as PCS white (D50) with actual whitepoint transformation encoded in a separate tag inside ICCprofile (the actual whitepoint of display like D65). If you see weird blue color cast in white when making a LUT3D, choose “relative colorimetric” instead of “absolute” when making a LUT3D for madVR.

    • This reply was modified 1 week, 5 days ago by Vincent.
    • This reply was modified 1 week, 5 days ago by Vincent.
    #36260

    mat a
    Participant
    • Offline

    Thank you for your help Vincent.

    #36273

    Raj S
    Participant
    • Offline

    Do you edit in sRGB? Or do you edit in AdobeRGB colorspace? Or maybe both? It would come down to using either Novideo_sRGB for systemwide correction (but clamped to sRGB) or to use an ICC profile for application-specific correction (any colorspace, but some apps will not be corrected/accurate). Photoshop with the ICC option would probably be a bit more accurate.

    Also for correction use LCD PFS Phosphor WLED IPS. Your display has 98% P3 so I recommend using “98% Adobe RGB/96% P3 (HP DreamColor Z24x G2)”. Or if you don’t get good results, you can try “94% P3 (Panasonic V… in Lenovo P70)” which is usually right for most consumer displays with P3 coverage.

    And thanks Vincent for sharing your knowledge! Very interesting as always.

    #36275

    mat a
    Participant
    • Offline

    I have created a profile with Xrite iprofiler (to sRGB standard, Gamma 2.2).

    I have then loaded the profile into Novideo_srgb with the Windows OS colour profile set to sRGB.

    Is it better to use EDID primaries?

    I am not sure that this is the right way to proceed. I use a mix of 3DSMax, a bit of photoshop for post and texture and a lot of gaming so I thought that the best way was to clamp to sRGB and be done with it!

    I must say that I used to use DisplayCAL but I thought maybe I was using ‘wrong’ so I went back to the more simple iProfiler. Happy to go back to Display CAL but it is daunting…

    I have followed your directions and now use 8bits with dithering (SpatialStatic 2×2).

    Thank you.

    #36277

    Vincent
    Participant
    • Offline

    I have created a profile with Xrite iprofiler (to sRGB standard, Gamma 2.2).

    I have then loaded the profile into Novideo_srgb with the Windows OS colour profile set to sRGB.

    Is it better to use EDID primaries?

    No, that should be fine… but Xrite profiles usually store WP & primaries as PCS coordinates instead on relative to actual whitepoint. IDNK how LeDoge app deals with that. Ask him.

    I am not sure that this is the right way to proceed. I use a mix of 3DSMax, a bit of photoshop for post and texture and a lot of gaming so I thought that the best way was to clamp to sRGB and be done with it!

    AFAIK CAD/CAM is usually not color managed so using a single setup emulated to sRGB seems fine for YOUR work.

    I must say that I used to use DisplayCAL but I thought maybe I was using ‘wrong’ so I went back to the more simple iProfiler. Happy to go back to Display CAL but it is daunting…

    I have followed your directions and now use 8bits with dithering (SpatialStatic 2×2).

    Thank you.

    Calibration speed to fast or medium, D65, gamma 2.2 or sRGB and profile type “matrix + single curve” should be equivalent to Xrite matrix profile setting using bundled colorimeter corrections. The only difference is that Xrite stores profile data in PCS (profile connection space, D50) instead of illuminnat relative. If you make a LUT3D maybe you want to use relative colorimetric with Xrite profiles, IDNK if DisplayCAL LUT3D maker loads CHAD data from Xrite profiles.

    #36278

    Vincent
    Participant
    • Offline

    Also for correction use LCD PFS Phosphor WLED IPS. Your display has 98% P3 so I recommend using “98% Adobe RGB/96% P3 (HP DreamColor Z24x G2)”. Or if you don’t get good results, you can try “94% P3 (Panasonic V… in Lenovo P70)” which is usually right for most consumer displays with P3 coverage.

    No, that is a mistake. Those corrections differ on main green wavelength. 98% P3 aims for an AdobeRGB natuve grene primary, Macbook Retuna P3 aims for a exactly green P3 primary and 95% aims for a crippled green primary on intersection between 2D plot of P3 & AdoberGB, thus being the smallest colorspace of teh three.

    If your i1d3 colorimeter does not track std observer on those shorter wavelengths in HP Z24x G2 (and firmware stores that too) but your display has main green wavelegth in longer wavelegths (like those consumer 95% P3 displays) it MAY give you an over correction in some coordinates causing a color cast in resulting WP.

    If that consumer display actually had such high P3 coverage (97-99%) but no AdobveRGB green primary is better to use the one for Apple.

    A fast green primary measurement with or without PFS correction should give you a hint about if display has AdoberGB green or P3 green at native gamut, hence giving you a hint about correction.

    #36284

    mat a
    Participant
    • Offline

    Will try that. thank you for your detailed support.

    #36285

    mat a
    Participant
    • Offline

    Many thanks as always Vincent.

    #36290

    mat a
    Participant
    • Offline

    Hi Vincent

    I am sorry to ask but how do I do a green primary measurement?

    #36291

    mat a
    Participant
    • Offline

    Also. In novideo_sRGB, should I enable ‘calibrate gamma to sRGB’?

    Thank you.

Viewing 15 posts - 1 through 15 (of 36 total)

You must be logged in to reply to this topic.

Log in or Register

Display Calibration and Characterization powered by ArgyllCMS