Calibration Trouble – Nvidia – Too bright

Home Forums Help and Support Calibration Trouble – Nvidia – Too bright

Viewing 14 posts - 1 through 14 (of 14 total)
  • Author
    Posts
  • #36170

    Ultra
    Participant
    • Offline

    Hey all, new to DisplayCal and having some trouble getting things settled. I previously ran an AMD card (R290x) and using the driver color/display controls and my monitor controls achieved a fantastic calibration using some other hardware for reference (prints turned out accurate and great!). The card is power hungry though and loud, so I had the chance to swap to an Nvidia RTX 2060 Ti – a huge upgrade in performance. However, I was never able to get the screen outputs to match (running PIP on my Samsung SD850 32″ – had to use HDMI for one and Display Port for another). I eventually figured out that not only were the HDMI/DP outputs different on account of the cables but one card to the other card also had changes in DP performance. In short, I could only get close to my settings on the Nvidia card and the Nvidia gamma always seemed wonky.

    Enter the Calibrite color checker. Software didn’t even try to work – Enter DisplayCAL.

    So far display cal has worked…OK. I can get the calibration running and the metering to read pretty accurate – my black and white reproduction levels and color resolution is better than ever. Here is my problem. What I see IS NOT what I get in printing. If I override Cal and go to Nvidia settings, lowering the brightness from %50 to about 25-30% gets me print-accurate brightness levels…

    I’m also pretty frustrated that DisplayCal doesn’t seem to have any post-calibration tools DESPITE requiring me to disable all of Nvidia’s controls. Running a calibrated screen and then checking my photos by ducking the brightness software-side to simulate print is something I’m familiar with (Lightroom Soft Proofing – even with WHCC profiles – doesn’t quite get there).

    Somehow the *internal settings* for rendering brightness/gamma and contrast are just a bit too high and I can’t figure out any way to lower them. I tried calibrating at a higher CMD level (100 instead of 50 previously) and this helped slightly (I typically edit in a dim space). I just cant find any way to get user control over my black and white levels/brightness on the software side.

    My screen looks great, but its not telling me the truth!!! Adjusting the screen itself is not possible as its the actual black and white levels in rendering that need changing – if I go to 1 brightness I still can’t get anywhere near how dark the prints turned out. I really don’t want to have to go back to the AMD card or back to trying to run both (constant driver problems). At this point, I don’t even know what to try to adjust to help align my display with my prints. I print from WHCC and NPL mostly.

    Does display cal have post-calibration adjustments? And if not, WHY NOT?!?!?! Help please!!!

    #36176

    Vincent
    Participant
    • Offline

    So far display cal has worked…OK. I can get the calibration running and the metering to read pretty accurate – my black and white reproduction levels and color resolution is better than ever. Here is my problem. What I see IS NOT what I get in printing. If I override Cal and go to Nvidia settings, lowering the brightness from %50 to about 25-30% gets me print-accurate brightness levels…

    Nobody that knows what he is doing will do that. Don’t do that. Brightness must be controled in OSD. DisplayCAL has no DDC/CI, it0’s your responsibility to lower brightness to whatever target you want when tweaking whitepoint.

    I’m also pretty frustrated that DisplayCal doesn’t seem to have any post-calibration tools DESPITE requiring me to disable all of Nvidia’s controls. Running a calibrated screen and then checking my photos by ducking the brightness software-side to simulate print is something I’m familiar with (Lightroom Soft Proofing – even with WHCC profiles – doesn’t quite get there).

    Don’t do that. It’s useless to use software brightness. Nobody does that, unless he know nothing.
    If you want to get a match aim directly for a match
    -get a good light source, measure lux at paper level
    -aim to same white as light source, lux/PI as 1st guess for cd/m2

    Somehow the *internal settings* for rendering brightness/gamma and contrast are just a bit too high and I can’t figure out any way to lower them. I tried calibrating at a higher CMD level (100 instead of 50 previously) and this helped slightly (I typically edit in a dim space). I just cant find any way to get user control over my black and white levels/brightness on the software side.

    My screen looks great, but its not telling me the truth!!! Adjusting the screen itself is not possible as its the actual black and white levels in rendering that need changing – if I go to 1 brightness I still can’t get anywhere near how dark the prints turned out. I really don’t want to have to go back to the AMD card or back to trying to run both (constant driver problems). At this point, I don’t even know what to try to adjust to help align my display with my prints. I print from WHCC and NPL mostly.

    It’ looks that you ar doing all this stuff in the wrong way without knowing the basics.

    Does display cal have post-calibration adjustments? And if not, WHY NOT?!?!?! Help please!!!

    Why should have it?
    Profile is made after VCGT calibration is applies. It is made by measuring calibrated display response.

    If you modify display after calibration & profile, profile is no longer valid. It is common that just varying OSD brightness makes whitepoint to drift. Profile usually remains valid as all software rendering from image to screen is relative to whiteppoint.

    #36178

    Ultra
    Participant
    • Offline

    I don’t think you get what I mean, but its very clear you think I know nothing. What is being rendered to the screen is the problem. If I don’t know the basics then dumb it down for me, I’m not a professional and have limited access to tools – lowering brightness software side is a super easy way to get my screen to emulate print output and I’ve used it successfully for years. Obviously its not the right way to do things. Adjusting the OSD does not get me where I want to be. Its ridiculous that I’d have to re-calibrate over and over with different settings at a massive cost of time to maybe get where I need to be. Are there really not default ballpark values people use to get started? My screen is rendering images brighter than print output by far, and soft proofing only gets about halfway there even using the print shop profiles.

    If I need content going to my screen to be calibrated darker what adjustment can I start with in DisplayCal? Its bad enough to have to constantly order prints and reconfigure but if calibrating with different settings is the only option then I need some guidelines to try. I have no way to measure lux. If I’m in a dark, lamp-lit room what can I assume? I tried calibrating at 50cd/m2 and it got me my super bright prints. Going to 100cd/m2 helped a tiny bit and I don’t have time to calibrate 10x a day to play with all the options when its almost 30 minutes per calibration.

    Without the ability to compensate after calibration the only option is to calibrate over and over. I’m going from AMD software color settings to DisplayCal, obviously there is a learning curve and I’m not used to it, but that’s why I’m asking for some ballpark values to try in different variables to get started.

    #36183

    Vincent
    Participant
    • Offline

    I don’t think you get what I mean, but its very clear you think I know nothing.

    I’m sure about that now.

    What is being rendered to the screen is the problem. If I don’t know the basics then dumb it down for me, I’m not a professional and have limited access to tools – lowering brightness software side is a super easy way to get my screen to emulate print output and I’ve used it successfully for years.

    And it is the wrong way since you have a tool for measuring & calibrating display with DisplayCAL. Set brigthness on device.

    Obviously its not the right way to do things. Adjusting the OSD does not get me where I want to be.

    That does not mean that your target is valid. From what I read on the message is very clear that it is not.

    Its ridiculous that I’d have to re-calibrate over and over with different settings at a massive cost of time to maybe get where I need to be. Are there really not default ballpark values people use to get started? My screen is rendering images brighter than print output by far, and soft proofing only gets about halfway there even using the print shop profiles.

    If I need content going to my screen to be calibrated darker what adjustment can I start with in DisplayCal? Its bad enough to have to constantly order prints and reconfigure but if calibrating with different settings is the only option then I need some guidelines to try. I have no way to measure lux. If I’m in a dark, lamp-lit room what can I assume? I tried calibrating at 50cd/m2 and it got me my super bright prints. Going to 100cd/m2 helped a tiny bit and I don’t have time to calibrate 10x a day to play with all the options when its almost 30 minutes per calibration.

    It’s just that you do not know what you are doing.

    If your light source for hard copy cmparisons is so dim that 50cd/m2 n display is bright you are doing all this wrong.
    Also it’s very very very likely that such light source has a wrong spectral power distribution unsuitabe to be used for graphic arts. Read about lightsources for evaluationg prints.
    Also if yiu are using DisplayCAL, you need a measurement device and all or almost all of them are able to take ambient light readings. Them apply that 1st guess based on Lamberts law and get a hint of suggested brightness for screen, or suggested light source intensity for a givenb display brightness.

    Without the ability to compensate after calibration the only option is to calibrate over and over. I’m going from AMD software color settings to DisplayCal, obviously there is a learning curve and I’m not used to it, but that’s why I’m asking for some ballpark values to try in different variables to get started.

    You cannot compensate anything after calibration & profile because if yiou do that profile may be no longer valid.

    It’s all explained in previous message:

    -get a suitable light source for evaluating prints.
    -measure it at paper level, color & lux
    -use it as a hint for display target whitepoint & brightness
    -calibrate & profile

    These are REQUIREMENTS for getting closer to match +- additional WP & brightness tweakings ON DISPLAY, not in useless GPU driver controls.

    Use softproof on display to simulate paper gamut limitations and optionally simulating paper white (which is relative to lightsource, hence display must be a match in color white to light source) & loss of contrast due to “minimum amount of reflected light by ink” (aka “low contrast due to being a print”)

    #36186

    Ultra
    Participant
    • Offline

    I GET IT, you have a PhD in calibration and being a prick. Please don’t comment here anymore, I don’t even care if you lay it all out step by step. Insufferable people like you make giving up a lot more attractive, please just go away. I literally don’t want to hear anything else you have to say.

    #36187

    MW
    Participant
    • Offline

    Literally no one can help you if you refuse to learn basic concepts and take some direction. That’s the definition of megalomania so let’s not put that on Vincent.

    #36188

    Vincent
    Participant
    • Offline

    I GET IT, you have a PhD in calibration and being a prick. Please don’t comment here anymore, I don’t even care if you lay it all out step by step. Insufferable people like you make giving up a lot more attractive, please just go away. I literally don’t want to hear anything else you have to say.

    Again, start with basics:

    -get a suitable light source for evaluating prints.
    -measure it at paper level, color & lux
    -use it as a hint for display target whitepoint & brightness (cd/m2=lux/pi)
    -calibrate & profile

    For example ISO P2 appraisal level aims to something like:
    -monitor D50 , 160cd/m2 which is a little bright
    -light source close to D50 spectral power distribution (or 5000k thermal/blackbody equivalent), 500lux (500/3.1416 = …)
    From that starting point you can rely on paper white & ink contrast simulation in your softproof tool, or disable these features in software and fine tune white point & contrast (and then calibrate & profile)

    It’s fairly simple to extrapolate this setup for a more laid back enviroment for example 100cd/m2 on monitor and that rule gives you a hint on lux for your light source. Same for 90 or 80 cd/m2, although the less iluminated your print you are less likely to be able to spot print issues in darker colors, like color cast in a dye printer doing B&W (even if profiled if you do not optimize such dye printer profile fro better B&W).

    #36191

    Алексей Коробов
    Participant
    • Offline

    Vincent answers all things right here. But here’s also another side that has been skipped: what is your print? Is it wet print or inkjet one (…laser/Indigo)? What kind of paper you print on? What light sources are in your room?

    1. Papers brightness under consumer lights is usually much less than consumer-familiar screen brightness. I prefer 90 cdm on 27″ screen for prepress in my room that has windows and is lighted by consumer 3000K LEDs and sky both, 70…300 lux.
    2. Technical paper contrast is not perceped one (we have two eyes, we tilt print in hands, so our brain easily compensates for paper/inks glare). This problem is still not resolved in modern applications proof mode. Moreover, they don’t consider screen matt covering that is well seen by eye. This may be considered in the last CIE/ICC standards but has not yet implemented. Use contrast (and WP) emulation with caution.
    3. Human perception of contrast depends on illumination intensity. Yes, you need contrast correction to ambient lux level integrated in calibration part of profile. DisplayCAL can do it. But this tool is quirky for the next reasons: a) screen is a light source itself, big screen changes ambient light with different images displayed on; b) ambient lux correction together with proofing mode in Ps/etc. may show wrong contrast. So, better to make recommended lighting conditions in your room: quite dim (and not cool, not intensively tinted with wallpapers) light, than you may follow typical 80-120 cdm for printed photo editing. But you have to check prints under more bright lights. If you still prefer to work under bright light, the only way is to find out good enough lux correction level. You can do it fast enough building low quality profiles.
    4. Print WP emulation in proofing mode is poorly implemented in Ps, especially for viewing under LEDs. Print profiles are made for D50 light source (5000K of daylight) and either 200 or 2000 lux condition in most cases. This is old standard, this may be really good for “natural white” papers or offset prints, but most of photo papers use OBAs (colors and brightness depend on light source very high) and we have warm or cool, but not 5000K LEDs with ugly spectrum in most rooms. However, I use WP+contrast emulation to check color and contrast relations. I note that ColorGATE RIP has better implementation of this mode than Photoshop has.

    We’re not magicians, so, please, provide sufficient information for support and make effort to dig up color measurement, good printer/prepress specialist must know it at some level.

    #36192

    Алексей Коробов
    Participant
    • Offline

    (color management, sorry for mishmash)

    #36195

    Ultra
    Participant
    • Offline

    I think my issue is I need really, really basic input apparently. Jumping straight to the information given here when I don’t even have the ability to situate my current setup or measure anything doesn’t give me much room to move. I’ve switched back to the Calibrite software which now actually works fine, previously would only glitch and give up. My room is literally 4 or 5 cd/m2 under lamplight (Calibrite software let me measure) and that’s not something I can change in my current situation. Again, I’m not a pro and often don’t have the luxury of daytime editing. I never had prints as off with my old non-calibrated setup as my calibrated one, that’s what really gets me. The contrast is fine, only the brightness is off, equivalent to a few stops of underexposure. My monitor is giving me more color resolution/separation and better viewing of bright and dark values than ever before, its just not allowing me to get the whole thing dark enough to get close to prints in a dark room. With Calibrite, I can cheat over to Nvidia for a minute with the brightness down to at least approximate my dark and light levels matching them to physical prints in the room. However, this still isn’t ideal since even using printer profiles from my print shop, I can’t emulate a print output on my screen during editing. No adjustments I’ve made during calibration seem to make any large difference either and I’m just scratching the surface on what D50, D65 etc. are. I’m used to editing at 6600k temp on my screen so my recent calibration was at D65. My colors always measure good and my black and white reproduction seems fine as I can see all values clearly, I just can’t get a good on-screen print output without (I’m assuming) waiting to confirm prints in bright daylight. The question is how to get a good print proof (brightness) on screen in a dark room without being able to achieve it through calibration. I don’t need anything crazy and I feel like most of the advice here is for accuracy concerns that are far outside the realm of what I’m trying to do. I’m not getting prints in magazines I’m trying to not have to re-edit and re-print family pictures over and over due to brightness inconsistency. The shadow content of my screen is far brighter than prints.

    #36196

    Ultra
    Participant
    • Offline

    I’m perfectly familiar with megalomania. A pearl served on a plate of dung is still offensive, regardless of its value.  

    #36197

    Vincent
    Participant
    • Offline

    Buy a solux halogen lamp, or at least a neutral high cri Yuji LED bulb for a desktop lamp (standard E26/E27 screw-in base).

    Your current setup is totally useless for your print evalution purpose. Then proceed to calibrate as instructed, otherwise you are wasting time.

    • This reply was modified 2 weeks, 5 days ago by Vincent.
    #36199

    Алексей Коробов
    Participant
    • Offline

    Buy a solux halogen lamp, or at least a neutral high cri Yuji LED bulb for a desktop lamp

    High-CRI LEDs still have huge pit in deep reds, they don’t actually emulate normative D50 for prints (may be good for shooting though). Probably High-CRI D55…D65 LED + halogen (less powerful, but with deep reds) mix in a single matte lamp shade (need to avoid left/right look difference) will give good result. However, principal problem is print machine profile made for D50. I use synthetic light condition when I build profiles for typical consumer photo papers, this provides compromise between consumer LED and out of window lighting. I think it is better to get sample print from the lab to compare screen with it under user lights. All in all, you should buy some desktop/compact lamp or viewing box to check prints in your dark room.

    Also, you must set correct proofing intent (the one your lab uses): either relative colorimetric with black point compensation (relatively precise, but may to clip saturated colors) or perceptual (compresses saturation to machine capacity, make some color shift, make little DR compression).

    Check for output signal in NVidia control panel. If your display black is too bright, this may be 15-235 signal range for TVs, NVidia loves to set it for HDMI. Set RGB 0-255 output (install NVidia Studio-type driver, not gaming one).

    Reset display settings before calibration/profiling, set user color with RGB control, but don’t move the bars. Set desirable brightness only (actually add 10-15%), don’t calibrate with 100% of brightness, you’ll get limited contrast and too bright black.

    If you calibrate with colorimeter, you must select proper display panel type (SD850 seems to be near-sRGB WLED-lighted, close to BenQ PD2700Q which correction is downloaded to DisplayCAL database). You don’t need it for spectrophotometer, however, better to use DisplayCAL with its Hi-Res spectral mode. Always set black point compensation in profile settings for consumer needs. Note, that white point is a slippy thing, its perception depends on ambient light and consumer devices are not precise enough to get it 100% right. You may move RGB bars in your display menu to 1-2 points to tune WP buy eye after calibration.

    #36200

    Vincent
    Participant
    • Offline

    Buy a solux halogen lamp, or at least a neutral high cri Yuji LED bulb for a desktop lamp

    High-CRI LEDs still have huge pit in deep reds, they don’t actually emulate normative D50 for prints (may be good for shooting though).

    I said that as lowest cost solution that can give hime some “quality” light, check out YUJI high cri SPD… since it’s very very very likely that a YUJI D50 tube/module (violet led + multiphosphor) or a normlicht D50 CCFL is out of his scope at this point in time (=it seems clear to me that at this point of time he won’t buy them, maybe in near future if he will be more serious at tis stuff, but not now).
    Hence a “quality” bulb “E27” that can be plugged in a flex desktop lamp is his best chance to improve light for print evaluation “under lowest cost contraints”.

Viewing 14 posts - 1 through 14 (of 14 total)

You must be logged in to reply to this topic.

Log in or Register

Display Calibration and Characterization powered by ArgyllCMS