Error message while calibrating

[ajohn]

I had this error message pop up when doing an uncalibrated check and again when I calibrated the display.

2017-08-10 13:15:50,466 Checking video card gamma table access for display 1...
2017-08-10 13:15:50,476 Dispwin: Warning - new_dispwin: Expected VideoLUT depth 11 doesn't match actual 10
2017-08-10 13:15:50,530 VideoLUT has 1024 entries, interpolating to 256
2017-08-10 13:15:50,692 Dispwin: Warning - new_dispwin: Expected VideoLUT depth 11 doesn't match actual 10
2017-08-10 13:15:50,700 Dispwin: Warning - new_dispwin: Expected VideoLUT depth 11 doesn't match actual 10
2017-08-10 13:15:50,700 Verify: 'test.cal' IS loaded (discrepancy 0.0%)
2017-08-10 13:15:50,707 Dispwin: Warning - new_dispwin: Expected VideoLUT depth 11 doesn't match actual 10

Not sure what this means or if it matters. It’s an MVA panel that I believe is a true 10 bit one covering something like 104% sRGB.  Also not sure what excess coverage means in practice. All of my other monitors have fallen a tiny bit short.

The graphics card is an nvidia GTX 1050 Ti silent with the nvidia driver for linux,

John

–

[Florian Höch]

Hi,

that’s not an error, it’s a warning, and it’s harmless. Generally, everything that doesn’t generate an error dialog can be safely ignored.

[ajohn]

I know you are a busy man Florian but feel that I need to bother your further. New monitor and departure from what I usually use and I’m checking if I should send it back.

The monitor is a 40″ Iiyama and clearly covers more than the sRGB gamut. I understand it is a 10 bit panel so the warnings about LUT depth concern me.  I assume  I am using an 8 bit one and that may have some implications. On the other hand I may be using 10bit. If so no worry.

I understand that the calibration will ensure that only  sRGB colours are displayed  but to do that if it’s 8bit it can’t be using all of the range that this implies. Taking a 10bit aRGB monitor for instance that still has adequate resolution when profiled to sRGB.  As mentions though this is all black magic to me so can only try to figure out how this aspect works.

I have run the large verification chart on it and that has picked out some larger error peaks. They are acceptable. I do read the info you provide, mostly on the old forum, and remember mention that xyz LUT’s may not produce a smooth colour profile so would expect some figures that are worse. If I’m using 8 bit though there may be some much worse ones between those that can be tested. If so me thinks I need to send it back.

I’ve attached the gamut.  All the other monitors I have used just about cover it with a slight lack of blue.

John

–

Attachments:

You must be logged in to view attached files.

[Florian Höch]

The monitor is a 40″ Iiyama and clearly covers more than the sRGB gamut. I understand it is a 10 bit panel so the warnings about LUT depth concern me.

If your OS is actually capable of sending 10 bit output to the display depends a lot on the graphics card and driver you use. Also, applications need to be specifically written to support true 10 bit all the way through (e.g. Photoshop can only do it in fullscreen mode with AMD and nVidia Quadro graphics cards). 10 bit support may still provide benefits even with an otherwise 8 bit framebuffer in case the graphics driver can apply the videoLUT with high precision on top of the 8 bit framebuffer (AMD cards seem to have an advantage here, with nVidia it’s more hit and miss, e.g. on my own Windows 10 system, my GTX 1070 only sends a 10-bit signal over HDMI after booting up, if the system is put to sleep and then back online, 10 bit output is lost and the only way to bring it back is a reboot, which is very annoying).

I understand that the calibration will ensure that only sRGB colours are displayed

No, calibration does not (cannot) correct gamut (it can reduce effective gamut slightly in case the max per channel output needs to be adjusted to hit a given target whitepoint via the videoLUT).

I do read the info you provide, mostly on the old forum, and remember mention that xyz LUT’s may not produce a smooth colour profile

That is outdated information. XYZ LUT profiles are comparable in smoothness to matrix-based profiles since version 2.0.

[ajohn]

Thanks Florian – I did expect the LUT to reduce output.

Still not sure about the bit depth aspect. Is there any way I can find out if a 10bit has been applied.

I’m using display port and am pretty sure that it can support 10bit to the card even at 4k.  Also the monitor is ok – as far as bandwidth is concerned. I’ve no wish for a 10bit workflow so am only concerned about the LUT etc.

John

–

[Florian Höch]

Choose “Tools” -> “Report on uncalibrated display device”. Note that bitdepth detection past 8 bits may not be very reliable because it challenges the repeatability of the instrument.

[ajohn]

I get this which I assume means it’s using 10?

13:34:40,352 dispcal: Warning – new_dispwin: Expected VideoLUT depth 11 doesn’t match actual
13:34:40,352 10

John

–

[Florian Höch]

You should see the assumed measured bitdepth a bit further down the log.

[ajohn]

LOL It’s not very reassuring Florian – excuse my sense of humour.

It states  “Effective Video LUT entry depth seems to be 10 bits”

I’ve found some data in the info gamut/gama pop up. Shot attached, 16bit depth. That ties in with something else I noticed – more space for LUT entries, interpolating to 256 entries. Then comes a mention of 8 bit entries in it if I am understanding correctly.

I take your point about my colorimeter.  A spectrometer is too expensive for what is really hobby use and no one I am aware of provides any info on how accurate others are. I also suspect that spending more may not achieve much in this respect. I selected on the basis of a UK review – this one ok and suitable for LCD types. At the time some were not.

John

–

Attachments:

You must be logged in to view attached files.

[Florian Höch]

It states “Effective Video LUT entry depth seems to be 10 bits”

So, that would indicate 10 bit.

I’ve found some data in the info gamut/gama pop up. Shot attached, 16bit depth.

Don’t confuse the vcgt data format with what the graphics hardware (including the display) may or may not be capable of.

[ajohn]

Thanks for your time Florian. It’s  good of you to have this forum.

One more to finally clarify things for me.  I’ve never had a monitor that exceed the sRGB gamut before.

I’m assuming that when I run a verification I am checking the values an 8bit sRGB image could provide over it’s entire range.  Is that correct?  In other words an image of this type can’t show colours out of the sRGB gamut even though the monitor can.

Just a comment. I looked at the reviews again. The colormonkey display compares very favourably with the i1display pro. The only difference is software and it’s also slower than the pro.  There doesn’t seem to be a better alternative really. Also a monitor calibration review site mentions use of a colorimeter for certain tasks because it’s black end response is better. I’d guess for my needs a change would be swings and roundabouts.

Subject to the query this monitor looks like a keeper.

John

–

Calibrite Display Pro HL on Amazon  
Disclosure: As an Amazon Associate I earn from qualifying purchases.

[Florian Höch]

I’m assuming that when I run a verification I am checking the values an 8bit sRGB image could provide over it’s entire range. Is that correct? In other words an image of this type can’t show colours out of the sRGB gamut even though the monitor can.

Yes (in a color managed application).

The colormonkey display compares very favourably with the i1display pro. The only difference is software and it’s also slower than the pro.

It’s basically the same device, but the firmware of the colorMunki Display limits speed to a minimum time of one second per read.

[ajohn]

Thanks Florian. The problem for me in this area is that it’s a bit specialised and has it’s own vocabulary.

😉 Anyway it looks like I have a pretty decent monitor. The sRGB colour balance as it comes is pretty good but the grey scale is off mostly in respect to the white point. de of 4

Thanks again.

John

–

[ajohn]

Just a note. Concerned about filters in colorimeters and using an unusual type of panel I just changed to a ColorMunki Photo. Results differ from those I had from the Display but there wasn’t much in it. I ran a verification with the Photo on a calibration done with the Display. I showed one aspect as being slight worse – just one patch really. The worst one the monitor has but still within limits.

Which instrument is the most accurate.? Well they don’t tell us that.

John

–

• This reply was modified 6 years, 8 months ago by ajohn.

[Author]

Posts