DisplaCAL is currently calibrating my monitor. I set the calibration speed to high, which gave me an estimation of 14 minutes IIRC. A look at the the screenshot below determined that was a lie.
Is this normal? I’m using a Spyder4 pro. Calibration using the spyder software under windows was much much faster.
It’s normal. Total time does not depend only on the speed but also on the patch count.
Ok, is that something that can be influenced?
Indeed, in the Profiling section, e.g. using the “auto-optimized” test chart lets you configure the number of patches. A higher number of patches is needed for a good LUT profile than for a matrix one, IIRC >=1000 is great.
This is one of those “start it before going to bed and come back to it in the morning” kind of things for me. In my case I’ll start just before leaving the office and just let it run.
I’m no stranger to overnight processes and will be treating calibration as such in the future. It just surprised me as the calibration on windows with the spyder software was a matter of only a few minutes. The results were “good enough”. Granted i’m no expert on that matter.
There is a Spyder Pro or something like that will do it quicker, but the price is steep.
are you sure the settings were the same? eg if you want a matrix-only profile (not the default with displaycal) you can do it much faster. and if you have a good screen the result might very well be in the ‘good enough’ set
As for screens, mine for sure falls in the ‘not so good’ category, and I have found that the ‘not as quick as a lightening’ displayCAL does a really good job profiling the beast to the best possible outcome. Great software indeed!
but “good enough” is not good enough
I have the spyder4 pro and the spyder-software produces a blueish colour fault on my two Eizo Screens, over and over again. With displaycal everything is fine, no colour faults. I think this is absolutely worth the longer time that displaycal needs. To me it is.
I try to let it run at least an hours worth of patches which is about 2000. The profile is very detailed. displaycal and the $10 eBay Huey do somersaults over Pant*ne’s utility.
There is this common misconception that DisplayCAL inherently requires more time. It does not! it’s entirely up to the settings you use. You can make it zoom through profiling your monitor in a few minutes, but why would you - profiling is something you do once a year, so do it well.
Could you elaborate? I was under the impression that you should calibrate once a month?
If your profiling hardware is really good, the monitor is really consistent (the same lightness day in and day out) and your needs are really high you might want to profile monthly. But in my experience there is so little variation that it’s hardly worth it. YMMV.
@matejmarti don’t take my word for it. Calibrate it once a month using the same settings, then compare the graphs and measurements (DisplayCal can generate those), or just switch between the different ICCs, at the end of the year. They will probably all look identical.
We’ll see in late december of 2018 :D.
For now i’m quite happy, it took some time to calibrate both my monitors (both nothing fancy) but displaycal made them look almost identical.
I’ve noticed how limited my knowledge on calibration and profiling is, i’ll be looking more into that in the future.
Regarding how slowly or quickly a monitor’s display characteristics might change over time, I’ve been using the same monitor - an NEC LCD2190UXI - since somewhere around 2008. Over the years I’ve kept more or less complete (though very disorganized) profiling notes, including information on the uncalibrated display characteristics (ArgyllCMS command “dispcal -R”).
So for what it’s worth, over the last 9+ years, and always using the monitor’s native white point and gamma, and with the measured “White level” kept in the range 60-75 cd/m^2:
The measured “Black level” has stayed reasonably close to 0.24 cd/m^2, with no obvious trend up or down over the years.
The measured “Approx. gamma” has stayed reasonably close to 2.2 - which is what the manual that came with the monitor says it’s supposed to be - with no obvious trend up or down.
The measured “Contrast ratio” has stayed fairly close to 256:1, with no obvious trend up or down.
The measured “Color Temperature” started out around 5800K with a DE of around 5 or less, and over the years the temperature seems to be steadily trending downward and the DE drifting upward, currently measuring around 5200K, with a DE of around 10.
If anyone else has records of how your uncalibrated monitor display characteristics have changed over time, I’d be very interested. I’m guessing the reason the Color Temperature for my monitor is steadily falling and the DE rising might be because the CCFL backlight is getting dimmer over time.
Anyway, I suspect my monitor might be on the downward slope of useability - I’ve noticed that using my usual “no calibration file, use native white point and gamma” approach to profiling has resulted in neutral gray midtones looking a bit green. So for my latest monitor calibrating/profiling I’ve resorted to two changes in how I calibrate and profile my monitor:
I used the monitor RGB Temperature controls to set a custom white point, setting R=255, G=238, B=250, using DisplayCal to guide changes to the monitor RGB values.
I changed the colprof parameters from “-as” to “-aS”, which eliminates abrupt hue changes in the deep shadows of what are supposed to be neutral gray gradients. I’m still experimenting, and might also end up using a calibration file loaded into the video card LUTs.
Regarding the slowness of monitor profiling, it’s the calibration file that takes such a long time to create. On my reasonably fast computer a low quality calibration takes a couple of hours, and a high quality calibration+profiling is best left to run over night. The really frustrating part of making a calibration file is that there are so many parameters to experiment with, and every time you want to try a different parameter value, a new calibration file is required.
Does anyone know what part of the calibrating process makes it so slow? It doesn’t seem to be affected by the amount of RAM or the speed of the processor. Is it related to the color measuring instrument?
On another topic, I noticed that the uncalibrated Black level and Contrast ratio are both fairly much affected by how bright the ambient light is. I think it might be best to calibrate and profile one’s monitors with the room as dark as possible, preferably in total darkness, and at the very least with the lighting held constant thoughout the profiling process. But this isn’t the lighting conditions under which we actually use our monitors, so I was wondering what sort of ambient light control other people use while calibrating/profiling their monitors.