Colour management in Photoshop on Dell u2711

[Mr_S]

Hi all,

I'm hoping to get some definitive answers about colour management from some of you clever types, as I can only seem to get mixed opinions from trawling the net.

I'm using a Dell u2711 monitor (in it's factory calibrated sRGB mode) under Win 7 and have mainly been using it for 3d work where colour management wasn't really an issue.  I'm now doing some design work in Photoshop with an aim to get the images printed and would obviously like the prints to look as close as possible to the onscreen image.  When I received the monitor it came with an .icm (which I didn't use, but downloaded the newest one from the Dell site) and installed it. 

So in windows colour management/devices tab, I have added the dell profile which now displays Dell U2711 Color Profile, D500 (default) - U2711.icm.  Everything else I've left alone in the other color management/advanced tab.  Is this correct?

Now, I've just open up Photoshop Elements 8 and have enabled colour management on an open document.  What I get is very muted saturation in "optimise for computer screens" and with the "optimise for print" setting, I also get muted saturation but not to the extent of the "optimise for computer screens".  I would assume this to be correct?

So I put this down to even though the monitor is hardware calibrated when using either sRGB or Adobe RGB modes, the standard icm profile is a generic one that is expected to give a ball park result.  Investing in a colorimeter and profiling to my exact display would give me much better results?  I'm looking into getting an eye one display 2 colorimeter, though I believe Xrite has discontinued this and have new models out..

Also, I've noticed that Photoshop will let me embed the Dell profile into the file - I'm guessing this is the correct thing to do if the printers have a profiled setup, other wise it will default to a generic sRGB profile?

Thanks in advance for any answers - I would greatly appreciate any help on all this!

[NikoJorj]

Investing in a colorimeter and profiling to my exact display would give me much better results? 

Yes, much better than assigning the wide-gamut profile to the screen in sRGB mode.
You can read this on the subject.

[Steve Weldon]

Hi all,
 
the standard icm profile is a generic one that is expected to give a ball park result.  Investing in a colorimeter and profiling to my exact display would give me much better results?  I'm looking into getting an eye one display 2 colorimeter, though I believe Xrite has discontinued this and have new models out..

Perhaps.  Probably.  It is possible your monitor is perfectly matched to the Dell profile provided.  But not likely.  It is possible for manufacturers to do this as NEC has demonstrated with their new PA series, and remain very accurate.  But from what I've seen with the Dell monitors, they're not in the same quality/accuracy realm as the NEC's.  Still, it's possible.  Have you tried making a print to see if you're satisfied with the results?  If an occasional use print is your only goal, perhaps you'll be satisfied. 

If not satisfied buy a good colorimeter/software and go through the learning curve of color management. 

[DaveCurtis]

I'm using a Dell2711 monitor. I have it profiled using a Spyder3. Im running the monitor in "Custom Color Mode" to take advantage of the wide gamut display.

The Spyder3 with Spyder3Elite does reasonable job and is suppose to handle wide gamut monitors. After calibration (6500K 2.2 120Lum) I have noticed that the 2711 is slightly too magenta when compared a to my prints. Still working on that though. I can adjust the RGB sliders slightly to fix this but I shouldnt really have to.

Dave

[Mr_S]

Thanks for the replies - sorry for the late reply.  Appreciate the help...

[digitaldog]

Perhaps.  Probably.  It is possible your monitor is perfectly matched to the Dell profile provided.  But not likely. 

Considering the graphic system has an effect on what comes out of that display, and everyone’s is possibly different, I’d say like you, not likely.

[hjulenissen]

Considering the graphic system has an effect on what comes out of that display, and everyone’s is possibly different, I’d say like you, not likely.

Are you saying that if Lightroom ship a 8-bit integer equal to e.g. [32 42 66] or a floating point equal to [0.42 0.666 1.0] to the graphics subsystem, a significantly different value may appear at the DVI link?

-h

[Coloreason]

Are you saying that if Lightroom ship a 8-bit integer equal to e.g. [32 42 66] or a floating point equal to [0.42 0.666 1.0] to the graphics subsystem, a significantly different value may appear at the DVI link?

-h

I think it is similar to the hardware controls of the monitor. If you change the settings of your video card that affect color, brightness, gamma, and etc the displayed colors will change dramatically on screen but their color values like  [32 42 66] remain the same as reported by a screen capture image. On the other side color managed software does change the actual values like [32 42 66] to something else to translate colors from simulated color spaces to the monitor's color space. The difference between video card and hardware controls is that using the video card for correction will leave less colors to be used for mapping of  8-bit image display which will cause banding and reducing the size of the monitor's color space (narrower gamut) thus making it harder to simulate other gamuts. That's why calibration is achieved best when achieved entirely by the hardware controls of the monitor and the video card is not contributing for calibration at all.

[hjulenissen]

I think it is similar to the hardware controls of the monitor. If you change the settings of your video card that affect color, brightness, gamma, and etc the displayed colors will change dramatically on screen but their color values like  [32 42 66] remain the same as reported by a screen capture image. On the other side color managed software does change the actual values like [32 42 66] to something else to translate colors from simulated color spaces to the monitor's color space. The difference between video card and hardware controls is that using the video card for correction will leave less colors to be used for mapping of  8-bit image display which will cause banding and reducing the size of the monitor's color space (narrower gamut) thus making it harder to simulate other gamuts. That's why calibration is achieved best when achieved entirely by the hardware controls of the monitor and the video card is not contributing for calibration at all.

What you are saying is that the graphics hardware/driver adds a layer beneath the application in which image data can be changed. My questions are:
1. Will typically AMD/Nvidia/Intel/... hw/drivers be shipped in such a default state that this will happen, or will it typically happen only for those that actively mock around with driver settings (e.g. "vivid" colors). I would assume that any sane driver developer would have transparancy (i.e. 1:1 mapping) as the ideal default behaviour.
2. Will applications like Photoshop use graphics APIs where it is possible for drivers/user to mess this up, or will they write directly into buffers that are "as close as possible" to the hardware framebuffer?

[Coloreason]

What you are saying is that the graphics hardware/driver adds a layer beneath the application in which image data can be changed. My questions are:
1. Will typically AMD/Nvidia/Intel/... hw/drivers be shipped in such a default state that this will happen, or will it typically happen only for those that actively mock around with driver settings (e.g. "vivid" colors). I would assume that any sane driver developer would have transparancy (i.e. 1:1 mapping) as the ideal default behaviour.
2. Will applications like Photoshop use graphics APIs where it is possible for drivers/user to mess this up, or will they write directly into buffers that are "as close as possible" to the hardware framebuffer?

1. My understanding is that the default state of the video cards settings doesn't interfere with the processing range of the card, any change of the default settings only takes away from the display possibilities - it can't add anything. For example if a monitor has much more intense Red  and Green signals in comparison to Blue this will result in tinted color when neutral numbers (R=G=B) are used, calibrating such monitor to remove tint is done by reducing the Red and Green signal capabilities because Blue is physically limited and can't be increased. While both hardware and video card controls use the same approach, video card does this at the expense of its 8-bit graphics processing. It actually reduces from the bits available for producing shades in the colors and the result on the display is banding.

2. I don't think color management in programs like Photoshop uses the video cards at that level, it simply takes a value like [32 42 66] as recorded in the image file, reports it in the color measurement tools like the Eyedropper as such but sends another different value for display with the video card when the color space of the image is different from the color space of the monitor and this simulates another color space. The video card is probably used from programs on that level for other graphics tasks like Open GL, and from software that works on a system level like calibration software that uses the vcgt tag in the assigned monitor color profile to control the video card contribution of the calibration when the computer starts up.

[Geraldo Garcia]

Are you saying that if Lightroom ship a 8-bit integer equal to e.g. [32 42 66] or a floating point equal to [0.42 0.666 1.0] to the graphics subsystem, a significantly different value may appear at the DVI link?

I will try to answer with a real world example:

Recently the graphic board of one of the computers in my studio died. I replaced it with a simmilar model, same manufacturer and specs with a bit more memory. The rest remained the same, I didn´t even change the driver and even so I noticed a difference on the colors and could prove that with a profile validation tool.

The unit to unit variation is one of the most important reasons to always calibrate and profile your devices.

Best regards.

[DavidB]

Geraldo in that scenario where you replaced the graphics card, is the link to the monitor analog or digital?
Certainly with analog there're more places where the values can change (e.g. resistance of cables/connectors, performance of DACs, etc) but although it's harder to imagine the opportunities for change with digital, I'm sure they're still there.

For projectors with digital connections we've found it quite effective to copy display profiles between machines (e.g. to a visiting speaker's laptop) but obviously we don't hold projectors to quite the same standard as our desktop monitors...