Too much red with the ColorHug2

[ageb...@gmail.com]

Hi,

Today I tried calibrating my monitor with my freshly received ColorHug2. Unfortunately, I'm experiencing a red glow after the process, for both my monitors connected to my system. This in addition to some software problems, but those I'd need to file elsewhere I suppose.

1. Everything on the monitor has a red glow. In the FAQ we are pointed to an older article [1] but, due to it's age, it is unclear to me if this still applies, now that we have the ColorHug2.
2. The live USB only detects one monitor, while everything is being mirrored. So I didn't use the live USB but the native installed Ubuntu software as my default OS instead.
3. My right-hand monitor is set to primary, my left-hand monitor as secondary. Both times the calibration is displayed on the left-hand monitor though. The actual colors of the right-hand monitor, however, do change when I try to calibrate that one. Only disabling the left-hand monitor allows me to calibrate the right-hand one.

Attached you'll find three profiles:
- Eizo S1931 (high) 2017-01-01 13-52-56 colorhug2.icc
  A high quality ColorHug2 profile
- Eizo S1931 (low) 2017-01-01 17-12-15 colorhug2.icc
  A low quality ColorHug2 profile after some monitor settings tweaking (forcing 65K)
- GCM - Eizo - S1931 - 84068057 (2014-07-24) [16-52-28].icc
  An old good profile done by a Spyder3 a while back

All three profiles are for the same monitor. My second monitor of the same brand and type gives me the same results.

I also noticed that the temperature in the profiles don't actually reflect the selected 65K in the calibration program. It's set to 64K and 66K instead for some reason.

Device: ColorHug2 (firmware 2.0.7)
Monitor: Eizo S1931
OS: Ubuntu GNOME 16.10 (with GNOME 3.22).
Video card: AMD Radeon RX480

Why does the red shift still occur, even though a CCMX isn't supposed to really be needed any longer with the ColorHug2?
Do the steps in the article still apply nowadays?
How do I get a fully functional calibration as intended with the ColorHug2?
Why is there a difference in temperature in the profiles?

Let me know if you need any more information.

Best regards,

Age Bosma


[1] https://encrypted.pcode.nl/blog/2012/06/15/colorhug-red-shift-workaround/

[Bruno Pagani]

Hi,

Le 01/01/2017 à 18:49, ageb...@gmail.com a écrit :
> Hi,
>
> Today I tried calibrating my monitor with my freshly received
> ColorHug2. Unfortunately, I'm experiencing a red glow after the
> process, for both my monitors connected to my system. This in addition
> to some software problems, but those I'd need to file elsewhere I suppose.
>
> 1. Everything on the monitor has a red glow. In the FAQ we are pointed
> to an older article [1] but, due to it's age, it is unclear to me if
> this still applies, now that we have the ColorHug2.
>

I’ll only answer this one, since I dunno about the others two.

> Why does the red shift still occur, even though a CCMX isn't supposed
> to really be needed any longer with the ColorHug2?
> Do the steps in the article still apply nowadays?
> How do I get a fully functional calibration as intended with the
> ColorHug2?

There is an issue with the ColorHug2 default calibration matrix. You can
see that by using the ColorHug Display Analysis tool for instance, which
will report a way higher temperature than expected when pointed toward
known sources. In order to work around this you can set the matrix to
the EDID provided one (what I did, you can get it using “Auto” mode in
DisplayCAL) or use the identity one (which is called “Raw” in
DisplayCAL). That’s far from ideal, but much better than the ColorHug2
provided matrix. See discussion in this thread[0] to know all about this
and what to do.

This has been further reported[1] several[2] times[3] (and probably
more). Richard partly acknowledged it at some point[4], but sadly we
never heard of him again on this topic…

Happy new year,
Bruno

[0] https://groups.google.com/forum/#!topic/colorhug-users/oHc3aZmJXi8
[1] https://groups.google.com/forum/#!topic/colorhug-users/TZEMdxjCvv8
[2] https://groups.google.com/forum/#!topic/colorhug-users/7nbDjG0ZoVg
[3] https://groups.google.com/forum/#!topic/colorhug-users/xx2Say3y4Qw
[4] https://groups.google.com/d/msg/colorhug-users/oHc3aZmJXi8/wkumj8NIEQAJ

[ageb...@gmail.com]

Hi Bruno,

Thank you for your reply.

I'm sorry to hear that the ColorHug2 does not work out of the box. In order to find a reliable solution I found the following page: https://colorimetercorrections.displaycal.net/?get&type=*&manufacturer_id=ENC&instrument=*&html=1
This has my monitor (S1931) with three Spyder4 CCMX's. Is this something I can use and if yes, which one and how? How reliable can these files be considered?

Best regards,

Age

[Richard Hughes]

On 1 January 2017 at 18:55, Bruno Pagani <bruno.n...@gmail.com> wrote:
> Richard partly acknowledged it at some point[4], but sadly we
> never heard of him again on this topic…

So, for ColorHug2 the calibration matrix is used for pretty much one
reason: It makes the results a lot more accurate on the hardware I
own. If the raw matrix (i.e. no correction) works better for people
(and, according to the sensor vendor, no correction matrix should be
required) then I can certainly switch to using no compensation by
default. There is a unity.ccmx factory cal matrix somewhere, but the
best way to do it is just to load the unity matrix into a spare spot
and then map the default LCD and LED matrices to that. If anyone wants
more detailed instructions just let me know.

Richard.

[Bruno Pagani]

Le 03/01/2017 à 21:10, Richard Hughes a écrit :

> On 1 January 2017 at 18:55, Bruno Pagani <bruno.n...@gmail.com> wrote:
>> Richard partly acknowledged it at some point[4], but sadly we
>> never heard of him again on this topic…
> So, for ColorHug2 the calibration matrix is used for pretty much one
> reason: It makes the results a lot more accurate on the hardware I
> own.

I indeed remember you stating this in the other thread. ;)

> If the raw matrix (i.e. no correction) works better for people
> (and, according to the sensor vendor, no correction matrix should be
> required) then I can certainly switch to using no compensation by
> default. There is a unity.ccmx factory cal matrix somewhere, but the
> best way to do it is just to load the unity matrix into a spare spot
> and then map the default LCD and LED matrices to that. If anyone wants
> more detailed instructions just let me know.

That would be nice (even if I can probably find how to do it reading the
man). :) Maybe add it in the FAQ?

Bruno

[ageb...@gmail.com]

Thanks you for your response.

Sorry for the delay. I wanted to check the CH2 on different types of monitors first.

In addition to the two Eizo S1931 I tested before, I now gave the meter a go on a HP EliteDisplay E232 monitor as well. In both cases I'm getting way too much red if the factory calibration is used. Only using "raw" mode in DisplayCal results in much more desired colours. Using "auto" mode gives me a too red-ish result as well, but less extreme than the LCD preset.

So yes, in my case I'd say I prefer no compensation by default as well. I don't, however, understand why unity.ccmx is needed if we are talking about no compensation. But if it is, I'd love to receive the instructions as well.

Best regards,

Age

[christia...@boku.ac.at]

Hi,

Am Dienstag, 3. Januar 2017 21:10:03 UTC+1 schrieb Richard Hughes:

[...]There is a unity.ccmx factory cal matrix somewhere, but the

best way to do it is just to load the unity matrix into a spare spot
and then map the default LCD and LED matrices to that. If anyone wants
more detailed instructions just let me know.

Richard.

Yes, please!
 
Yesterday I tested my colorhug2 (3292) the first time with the live-system and got also a reddish color cast (Eizo 2436). It looks similar to the results I got with version 1.

For the colorhug1 I created sometime ago the attached ccmx with a borrowed ColorMunki. Right after the creation it was working fine. However, I tried it some days later and I recive no readings from the colorhug device with this ccmx.
Life took over and I had no time to have a closer look on this. Anyway, I guess the ccmx will work correctly with the colorhug2?

Christian