Monitor brightness: NEC SpectraView Profiler 4 vs Multiprofiler

[Malcolm Payne]

Can anyone please explain or otherwise comment on the following discrepancy in reported monitor brightness levels between NEC SpectraView Profiler 4 and MultiProfiler?

The monitor is an NEC Spectraview Reference 271 (UK version) with SpectraView Profiler 4.1.24, profiled with an i1Pro (Rev. D) spectrophotometer, all running on Win 7 Pro 64-bit.

Target settings are White Point 5800, Black Point Min. Neutral, Luminance 120 cd/m2, Gamma 2.2

Running a profile with SVP produces the expected result, with achieved values very close to the target values and confirmed by a profile validation which shows similar values with good Delta-E results.

However, with a white background the screen feels perhaps uncomfortably bright, if not slightly dazzling at times, and MultiProfiler reports a Brightness level of 182 cd/m2 compared to the SVP value of 120 cd/m2. If I lower the Brightness setting in MP to 120 cd/m2 and re-run the SVP validation, SVP then reports the actual luminance value achieved as 78 cd/m2.

Quite apart from the need to achieve the correct brightness for accurate print matching, I am concerned about the large discrepancy in reported values between the two softwares and, particularly, the likelihood of premature backlight ageing if the monitor is indeed running at >180 cd/m2. Both programs are showing the correct profile loaded and I have made no changes directly in the MP settings.

If anyone can shed any further light on this (sorry, pun entirely unintentional!) I shall be very grateful.

[howardm]

I dont have an answer but only that I've seen the same thing w/ my USA software.  Often, Multiprofiler will seemingly have the wrong lum. level.

[Czornyj]

Try to set:
Settings>Luminance/contrast ratio>Black luminance: Min. Native

When "Min. Neutral" is selected SpectraView Profiler increases the backlight luminance to avoid sensor errors in the shadows.

[gromit]

Can anyone please explain or otherwise comment on the following discrepancy in reported monitor brightness levels between NEC SpectraView Profiler 4 and MultiProfiler?

Make it easy for yourself and just calibrate with MultiProfiler. A luminance value of 100cd/m2 is a good starting point if there's no light hitting the screen, but you can adjust manually up or down until you feel there's a good match. The absolute value doesn't matter (the same for white point). If you really must use an instrument, after calibration with MultiProfiler use SpectraView Profiler and do a "No calibration (profile only)". If you compare the results in Photoshop with the profile built by MultiProfiler you'll likely see little difference. I have the SpectraView 271 model, an i1 Pro and a number of colorimeters and prefer the results I get with MultiProfiler alone (doing both the calibration and profiling).

Given the literally thousands of posts here from people who stuff up what is in essence a pretty simple task, it makes you wonder when there already exists such an effective (and free!) solution.

[digitaldog]

If you really must use an instrument, after calibration with MultiProfiler use SpectraView Profiler and do a "No calibration (profile only)"

Interesting. So the trick is to turn off the “Calibration On” check box, then click the Calibrate button?

I have the SpectraView 271 model, an i1 Pro and a number of colorimeters and prefer the results I get with MultiProfiler alone (doing both the calibration and profiling).

So you just use MultiProfile without an instrument and adjust the calibration aim points until you get a match? Interesting, I’ll have to give that a try.

[gromit]

So the trick is to turn off the “Calibration On” check box, then click the Calibrate button?

This is with SpectraView Profiler (or basICColor display). I don't know what the comparable settings are for SpectraView II. It's good visual check of the correlation between MultiProfiler and whatever instrument you're using.

[Malcolm Payne]

Thanks to all for your comments and suggestions - I'll investigate these further, though Gromit's recommendation just to use MultiProfiler's internal calibration would seem to negate the entire monitor profiling industry as far as these high-end monitors are concerned.

I've just received the following response from basICColor tech. support:

"the SV Profiler reads out the real luminance of the monitor with a colorimeter/spectralphotometer and displays the measured values.
The MultiProfiler sets the monitor to an internal setting where it assumes that the monitor is set to that luminance. It doesn't measures the real luminance of the monitor."

"This is why you get different readings from the two applications."

Which explains the reason for the difference and confirms that the SV Profiler luminance value is the correct one, though I still find it odd that there's such a large discrepancy. I'm not concerned about the precise numerical value, except insofar as excessively high backlighting levels are liable to reduce the useful life of the monitor.

I do wish manufacturers would provide more technical details about the operation of their software though, I'm sure it would avoid a lot of confusion.

[howardm]

For me, Multiprofiler is almost always <10% off from the SV set value.

[gromit]

Which explains the reason for the difference and confirms that the SV Profiler luminance value is the correct one, though I still find it odd that there's such a large discrepancy.

No, only that the values are derived differently. I measured the luminance with my monitor set at 100 cd/m2 with MultiProfiler. One instrument I had reported 90 cd/m2, another 110 cd/m2. People are deluding themselves if they think the luminance and white point readings from their consumer grade instruments are "accurate".

The absolute value isn't important anyway as it's something you need to set by eye.

(Actually this was a while ago and maybe the measurements where 95 and 105 respectively. I do remember MultiProfiler being in the middle.)

[shewhorn]

No, only that the values are derived differently. I measured the luminance with my monitor set at 100 cd/m2 with MultiProfiler. One instrument I had reported 90 cd/m2, another 110 cd/m2. People are deluding themselves if they think the luminance and white point readings from their consumer grade instruments are "accurate".

This is kind of the thought I had. Ethan Hansen has already done quite a bit of testing on various colorimeters and spectrophotometers. Hopefully if he continues to update the calibration section at http://www.drycreekphoto.com he'll eventually get a PA241W on his bench and can compare the target luminance displayed on the OSD with the actual value read in by their Photo Research spectroradiometer. That would be a pretty definitive result that I'd trust right there (although it would only apply to one unit... ideally you'd want to test multiple units but if he set the luminance to 120 cd/m^2 and the reading came back as 121.27 cd/m^2 that would at least be somewhat encouraging).

My big question with the PA series monitors and the direct luminance settings is how does NEC account for the CCFL losing its brightness over the life of the monitor?

Cheers, Joe

[howardm]

My hunch would be that they (the factory) has a very good handle on the average deterioration  lumens vs. time curve and they given they have an ambient light monitor, they may even have a 2nd (low precision) photocell to monitor the backlight.

[shewhorn]

My hunch would be that they (the factory) has a very good handle on the average deterioration  lumens vs. time curve and they given they have an ambient light monitor, they may even have a 2nd (low precision) photocell to monitor the backlight.

I would think they would have to have something to measure the backlight anyway (how else would the luminance smoothing work?).

[gromit]

That would be a pretty definitive result that I'd trust right there (although it would only apply to one unit... ideally you'd want to test multiple units but if he set the luminance to 120 cd/m^2 and the reading came back as 121.27 cd/m^2 that would at least be somewhat encouraging).

A quick and informal test I did with my PA271W calibrated to 100 cd/m2 and 5500K with MultiProfiler:

Optix XR  .....  98.3  4732
Spyder 3  .....  94.7  5263
i1Pro (rev B)  .....  96.2  5180

I also tried a reading with my Display 2 but the values were way out. I had another that also died. Good riddance I say.

[Czornyj]

IMHO it's rather a matter of fact, that profilers tend to increase the backlight intensivity to reduce the problems with dark readings inaccuracy. The PA displays are suprisingly well calibrated "out of the box":