The Art of Monitor Profiling

[BruceGordon]

Hello,

I'm new to this forum and have only recently gotten back into Photography after being away from it for a few years (Sold My AE-1 to help fund College   )

I'm pretty passionate about making images and have an obsessive personality so I'm finding Colour Management frustrating as for something that seems, at first glance, to be a science appears to a neophyte like myself to have certain areas where it appears to be a black art.

First, I would like to say that I've been reading the LL for a couple of years and continue to be very impressed with the sites content and focus on image making and not just the technologies.  I recently started lurking in the forums and find the caliber of the content impressive.

I recently upgraded to an Epson R2880.  I've downloaded the profiles for the papers I use and installed them and configured my proofing.  I'm mucking about with my Rendering Intent and Black Point Compression to find combination's that work best with the images I want to print.  From what I understand these are the subjective controls in the final output process.

I'm running CS4 on Windoze XP Pro with a Samsung 2493HM LCD monitor.  I use a Spyder 2 colourimiter which I acquired with the Express version of the software.

My work environment is very dark (basement room with a grungy window and two luxo lights pointed on the wall with 5000K 13W compact fluorescents)

My first print out of the 2880 was very dark and some research led me to discover I had the White Point Luminance way too bright.  Not being able to set the Luminance with the basic software that came with my puck led me to download a trial version of Color Eyes Display Pro.  I'm very impressed with the software.  Even though my Samsung monitor & ATI HD5670 Video Card don't appear to work with the DDC contols it still magically manages to auto calibrate so I'm not stuck trying to adjust the RGB controls.  Very slick.

My issue is that even though my monitor is now a lot dimmer than it used to be and I have set the luminance target to 100 cd/m^2 my prints are better but still are about a stop darker comparing my prints to the monitor.

My brightness on the display is down to 10% at this point.

Should I try to go even lower?

Also, when adjusting my images I'm using the midpoint levels slider to adjust the gamma.  Images I printed previously on my old HP printer (where it controlled the CM) that looked fine on the monitor & printer now require a Gamma adjustment of 1.25 - 1.4 in order to print with a reasonably nice tonal range on the 2880.  Is there a better method than using the midpoint slider in Levels?

Regards,

Bruce

[Mark D Segal]

Hello,

I'm new to this forum and have only recently gotten back into Photography after being away from it for a few years (Sold My AE-1 to help fund College   )

I'm pretty passionate about making images and have an obsessive personality so I'm finding Colour Management frustrating as for something that seems, at first glance, to be a science appears to a neophyte like myself to have certain areas where it appears to be a black art.

First, I would like to say that I've been reading the LL for a couple of years and continue to be very impressed with the sites content and focus on image making and not just the technologies.  I recently started lurking in the forums and find the caliber of the content impressive.

I recently upgraded to an Epson R2880.  I've downloaded the profiles for the papers I use and installed them and configured my proofing.  I'm mucking about with my Rendering Intent and Black Point Compression to find combination's that work best with the images I want to print.  From what I understand these are the subjective controls in the final output process.

I'm running CS4 on Windoze XP Pro with a Samsung 2493HM LCD monitor.  I use a Spyder 2 colourimiter which I acquired with the Express version of the software.

My work environment is very dark (basement room with a grungy window and two luxo lights pointed on the wall with 5000K 13W compact fluorescents)

My first print out of the 2880 was very dark and some research led me to discover I had the White Point Luminance way too bright.  Not being able to set the Luminance with the basic software that came with my puck led me to download a trial version of Color Eyes Display Pro.  I'm very impressed with the software.  Even though my Samsung monitor & ATI HD5670 Video Card don't appear to work with the DDC contols it still magically manages to auto calibrate so I'm not stuck trying to adjust the RGB controls.  Very slick.

My issue is that even though my monitor is now a lot dimmer than it used to be and I have set the luminance target to 100 cd/m^2 my prints are better but still are about a stop darker comparing my prints to the monitor.

My brightness on the display is down to 10% at this point.

Should I try to go even lower?

Also, when adjusting my images I'm using the midpoint levels slider to adjust the gamma.  Images I printed previously on my old HP printer (where it controlled the CM) that looked fine on the monitor & printer now require a Gamma adjustment of 1.25 - 1.4 in order to print with a reasonably nice tonal range on the 2880.  Is there a better method than using the midpoint slider in Levels?

Regards,

Bruce

What specific brightness value does ColorEyes Display tell you your monitor is operating at? What gamma setting for the monitor are you using - 1.8 L* or 2.2? What paper are you printing on? Do you adjust the prints with soft-proofing activated? When you examine the prints after making them, what illumiation are you using to see them (brightness and temerature)?

[BruceGordon]

What specific brightness value does ColorEyes Display tell you your monitor is operating at? What gamma setting for the monitor are you using - 1.8 L* or 2.2? What paper are you printing on? Do you adjust the prints with soft-proofing activated? When you examine the prints after making them, what illumiation are you using to see them (brightness and temerature)?

I'm using D65 2.2 I thought but the software reports "L*"
The software reports 33cd/m^2 currently.  I'm using Epson Ultra Premium Photo Paper Luster and soft proofing.  Prints are examined using the Luxo Lamp about 1m above and the print at a 45 deg angle to the source which compares similarly to viewing in another room with a North facing window.

[Mark D Segal]

I'm using D65 2.2 I thought but the software reports "L*"
The software reports 33cd/m^2 currently.  I'm using Epson Ultra Premium Photo Paper Luster and soft proofing.  Prints are examined using the Luxo Lamp about 1m above and the print at a 45 deg angle to the source which compares similarly to viewing in another room with a North facing window.

Something about that monitor brightness setting is awefully suspect. I've never heard of a reading as low as 33cd/m^2. Is this a typo or what the software actually reports as the monitor brightness? Normally you would be working in the range of 110-120 for the kind of dimly lit environment you describe, and if it really were 33, it would be tempting you to vastly over-brighten the prints, which is opposite to your problem, so the evidence and the symptoms are contradictory.

[BruceGordon]

Something about that monitor brightness setting is awefully suspect. I've never heard of a reading as low as 33cd/m^2. Is this a typo or what the software actually reports as the monitor brightness? Normally you would be working in the range of 110-120 for the kind of dimly lit environment you describe, and if it really were 33, it would be tempting you to vastly over-brighten the prints, which is opposite to your problem, so the evidence and the symptoms are contradictory.

Agreed about the luminance.  Here are the screen shots from a Validation I just ran:

[Mark D Segal]

The display profiling exercise looks defective. The average Delta-E should be less than 1, and likewise for the maximum Delta-E for any one of those 15 values. You average is three times more than it should be and one of the values is above 18. This indicates that either something you did in the profiling process was not correct, or Color Eyes and Spyder 2 aren't playing well with your display. I recommend that you redo the monitor profile from scratch paying very close attention to implementing all the steps through the process, making sure the Spyder is recalibrated before you start, don't press it into the display - just let it sit on the display with no gap (tilt the display back). Set monitor Luminance to 110 as a start. Set gamma to L*. D65 or D50 depending on whether you want a cooler or warmer white point respectively. Use Color Eyes Monitor White function and the on-monitor controls to get the RGB values in balance before you start the profiling. See if you can produce a better monitor profile and then do a test print to see whether the matching is improved.

[BruceGordon]

The display profiling exercise looks defective. The average Delta-E should be less than 1, and likewise for the maximum Delta-E for any one of those 15 values. You average is three times more than it should be and one of the values is above 18. This indicates that either something you did in the profiling process was not correct, or Color Eyes and Spyder 2 aren't playing well with your display. I recommend that you redo the monitor profile from scratch paying very close attention to implementing all the steps through the process, making sure the Spyder is recalibrated before you start, don't press it into the display - just let it sit on the display with no gap (tilt the display back). Set monitor Luminance to 110 as a start. Set gamma to L*. D65 or D50 depending on whether you want a cooler or warmer white point respectively. Use Color Eyes Monitor White function and the on-monitor controls to get the RGB values in balance before you start the profiling. See if you can produce a better monitor profile and then do a test print to see whether the matching is improved.

Thanks Mark - I was just figuring out that the dE smelled rotten.  The display panel is recessed 1/4" from the case so I was using a piece of tape to hold the puck down (with the ends attached to the case not the screen of course   ).  I've reset the display to defaults and turned of the "Magic" functions and, as you suggest, will start out by attempting to get the white point set on the display using the OSC.

[Mark D Segal]

Bruce - don't do the tape-down - not even to the edges. If the puck got pressed into the display because of too much tension, that could be a good part of the profiling problem. As I mentioned above, best to lean the display back a bit and just let the puck sit on it.

[terrywyse]

I agree with Mark that both the luminance level and your validation dE seems a bit whacked. dE tolerance should generally be under .50 average and <1.0 peak. Luminance level can be whatever best matches up visually with your print viewing conditions. Typically this would fall in the 100-150 cd/m2 range (ISO standards say this should not exceed 160 cd/m2 for graphic arts viewing conditions).

On the white point, try using "native" or at least measuring what the white point is before pinning it to a specific value. Anything in the range of about 5500-6500K should be fine. It's generally better to use the native white point....as long as it's in a reasonable range...rather than forcing it.

Some other more basic stuff would be to CONFIRM that the system is using the display profile and also confirming that Photoshop is using it as well. In Photoshop, open Color Settings and then open the pop-up menu for RGB default working space. At the very top, verify that "Monitor RGB" is the correct display profile (obviously, don't set your RGB workfing space to be Monitor RGB, just use the pop-up menu to confirm that Photoshop is using/recognizing your display profile).

Terry

[terrywyse]

Forgot to add.....use Lab (checkbox) for your validation results. Might help us determine why the dE is so bad on steps 1-11 and step 12 in particular. I would guess if you were to display a gray ramp that you'd be seeing color inconsistencies up and down the gray scale.

Terry

[ChasP505]

I've used CEDP for quite some time and I also have experience with TN paneled Samsung LCDs like the 2493HM.  My recommendations for CEDP settings are:  

Brightness/Gains
ICC v.2
D65
2.2 gamma
120cd/m2
Black Point  - Minimum/Relative

On your monitor, start by resetting all settings to factory defaults, then set Brightness to 20%.  
Important: When you adjust the "Monitor White" using the RGB sliders and Brightness control, don't try for perfection.  With CEDP, it's OK if you end up with as much as +/-2 on each color channel.  AND... the Brightness level should NOT be at 0... it should be at about +5.  Adjust your Brightness control until it gets to about +5.  

[attachment=22262:adjust_m...e_ranges.jpg]

Regarding the Spyder2 Puck...  Make sure you have removed the suction cups, then tilt the screen back to about 25 degrees so the Spyder can rest on it flush without any pressure.  Allow it about 10 minutes to warm up to the screen's temperature.  Then profile.  With CEDP, an average dE of below 2.0 is fine for a consumer level TN monitor.

Adjust your ambient light higher to offset the higher luminance level.

[Ethan_Hansen]

The validation screenshot you posted has the appearance of light leaking under the puck edges. The points made by Mark and Chas about sensor positioning, removing the suction cups, etc. are spot on. I do, however, have qualms about the luminance target. High contrast, high brightness (10000:1 and 400 cd/m2 in the case of your Samsung) TN panels do not perform at their best at lower luminance values. Anything under 140-150 cd/m2 starts chopping the color gamut and introducing banding and posterization artifacts. Cranking up theillumination in your viewing area and, most importantly, soft proofing with the simulation options enabled in Photoshop is the key to accurate print matching.

[Mark D Segal]

I've used CEDP for quite some time and I also have experience with TN paneled Samsung LCDs like the 2493HM.  My recommendations for CEDP settings are:  

Brightness/Gains
ICC v.2
D65
2.2 gamma
120cd/m2
Black Point  - Minimum/Relative

Chas, any particular reason why you would use v2 rather than v4 profiles?

[BruceGordon]

The validation screenshot you posted has the appearance of light leaking under the puck edges. The points made by Mark and Chas about sensor positioning, removing the suction cups, etc. are spot on. I do, however, have qualms about the luminance target. High contrast, high brightness (10000:1 and 400 cd/m2 in the case of your Samsung) TN panels do not perform at their best at lower luminance values. Anything under 140-150 cd/m2 starts chopping the color gamut and introducing banding and posterization artifacts. Cranking up theillumination in your viewing area and, most importantly, soft proofing with the simulation options enabled in Photoshop is the key to accurate print matching.

Wow, thanks for all the help.

I'm a little confused by the suggestions to remove the suction cups.  The Spyder 2 has a frame for LCD monitors that fits over the base of the unit and has a honeycombed opening and a coloured disk to filtering the illumination from the target.  The instructions with the unit[/u] suggest that this should be used for LCD displays.  Is the suggestion then I remove the bottom plate and then remove the suction cups?

Cranking up the illumination in my working area could be problematic as it is a basement and would mean additional artificial light sources which would introduce glare.  The thought seems to be that this type of monitor is optimized for gaming and/or office productivity applications and that a monitor more suitable for proofing would be natively much less bright.

I've used CEDP for quite some time and I also have experience with TN paneled Samsung LCDs like the 2493HM. My recommendations for CEDP settings are:

Brightness/Gains
ICC v.2
D65
2.2 gamma
120cd/m2
Black Point - Minimum/Relative

On your monitor, start by resetting all settings to factory defaults, then set Brightness to 20%.
Important: When you adjust the "Monitor White" using the RGB sliders and Brightness control, don't try for perfection. With CEDP, it's OK if you end up with as much as +/-2 on each color channel. AND... the Brightness level should NOT be at 0... it should be at about +5. Adjust your Brightness control until it gets to about +5.

Regarding the Spyder2 Puck... Make sure you have removed the suction cups, then tilt the screen back to about 25 degrees so the Spyder can rest on it flush without any pressure. Allow it about 10 minutes to warm up to the screen's temperature. Then profile. With CEDP, an average dE of below 2.0 is fine for a consumer level TN monitor.

Adjust your ambient light higher to offset the higher luminance level.

Yes, tilting the monitor back has helped and as Mark and you suggested I adjusted the Monitor White first.  Acheiving perfection with the sliders was impossible but after many iterations I managed to get it to 0,0,-1 with a brightness at 0.  I then adjusted for D65, 2.2 & 110 cd/m2 & min. luminance for the Black Point.  At this point the calibration validation results show much less variation (i.e. the max dE is much less) but the over all average dE is much worse (1.8 vs, 0.58).  Should I be using the profile validation?

[ChasP505]

Chas, any particular reason why you would use v2 rather than v4 profiles?

Many browsers and image viewing apps don't support v.2 profiles.

http://www.color.org/version4html.xalter

[ChasP505]

...in the case of your Samsung) TN panels do not perform at their best at lower luminance values. Anything under 140-150 cd/m2 starts chopping the color gamut and introducing banding and posterization artifacts. Cranking up theillumination in your viewing area and, most importantly, soft proofing with the simulation options enabled in Photoshop is the key to accurate print matching.

Ethan, I thought about this after posting my recommendations and I have to agree with you.  I run my office Samsungs at 160-170cdm2.

[ChasP505]

Wow, thanks for all the help.

I'm a little confused by the suggestions to remove the suction cups.  The Spyder 2 has a frame for LCD monitors that fits over the base of the unit and has a honeycombed opening and a coloured disk to filtering the illumination from the target.  The instructions with the unit[/u] suggest that this should be used for LCD displays.  Is the suggestion then I remove the bottom plate and then remove the suction cups?

Sorry for the confusion. It's been years since I used a Spyder2 device.  What I meant is don't use the suction cups.

Cranking up the illumination in my working area could be problematic as it is a basement and would mean additional artificial light sources which would introduce glare.  The thought seems to be that this type of monitor is optimized for gaming and/or office productivity applications and that a monitor more suitable for proofing would be natively much less bright.

Not necessarily less bright, but more capable of running at a low luminance level without introducing color problems and backlight instability.

Yes, tilting the monitor back has helped and as Mark and you suggested I adjusted the Monitor White first.  Acheiving perfection with the sliders was impossible but after many iterations I managed to get it to 0,0,-1 with a brightness at 0.  I then adjusted for D65, 2.2 & 110 cd/m2 & min. luminance for the Black Point.  At this point the calibration validation results show much less variation (i.e. the max dE is much less) but the over all average dE is much worse (1.8 vs, 0.58).  Should I be using the profile validation?

Well if you follow the settings from the top down, you would have set all your profile parameters first, THEN adjusted the Monitor White, then profiled.  As I said, there's no need to strive for perfection in the Monitor White dialog.  CEDP will go directly to your video card controls and make adjustments there.  That's why I said set the Brightness not to ), but to about +5, as this compensates for the software lowering the RGB values at the video card.  Make sure you select Minimum/RELATIVE and don't specify an absolute black value.

And the profile validation has no effect on the profile just created.  Here's a typical validation of a profile on an entry level monitor, using my recommended settings:

[attachment=22268:validation.jpg]

[BruceGordon]

Thanks for all the input.

I adjusted for 140 cd/m^2, D65, Gamma of 2.2 and Relative/Min Black Point and the results while not impressive are tolerable.

I set the brightness to +4 as suggested and now the blacks seems much more in tolerance.

Now I'll have to try some prints though I suspect I'll still have to adjust the gamma to get useable results.

For me, not being a pro, it's a trade-off of $1000-$1500 for a better monitor versus saving the funds towards the 24-70 f/2.8L IS  

Dis a quick print of an image today and the result is still darker but nowhere as near as previously - less that 1/2 stop by my estimate

[Mark D Segal]

This validation result is good. If your print is still coming out a bit too dark, it means that your monitor brightness is a bit too high, inciting you to over-darken the file relative to what you see on paper. Don't trash the profile you just made, but try another with luminance down to about 120 cd/m^2 and try setting gamma at L*. See what the profile validation and a re-adjusted (i.e. brightened) print using the new monitor profile look like.

[BruceGordon]

This validation result is good. If your print is still coming out a bit too dark, it means that your monitor brightness is a bit too high, inciting you to over-darken the file relative to what you see on paper. Don't trash the profile you just made, but try another with luminance down to about 120 cd/m^2 and try setting gamma at L*. See what the profile validation and a re-adjusted (i.e. brightened) print using the new monitor profile look like.

Thanks Mark,
120 cd/m^2 did the trick.  Prints are matching the softproofing so far with the paper I tested (Premium Glossy).  I'll try a print with the Premium Lustre.
The dE seems to have degraded slightly but is still acceptable.  I'm assuming that I'm right on the edge of the monitor falling apart due to being driven so low as some others have pointed out.

I've been running a Demo License of CEDP which expired today.  I made the mistake of installing it before I left to go down to Toronto last weekend for Art Wolfe's seminar (which I really enjoyed) so I haven't had as much time to play with it as I hoped but the Demo had the desired result so I purchased it as it clearly demonstrated itself superior to the bundled software.  Thanks ICC for providing a demo.

I'm going to try a couple of luminance values between 120 and 140 to see if I can get the dE down a bit.

Again, thanks to all the posters here for pointing me in the right direction and sharing their knowledge of this rather black art.

Lessons Learnt:
A Really Nice Bright Monitor that is good for gaming is probably not useful for proofing prints in PhotoShop
There may be differences in colorimeter hardware but there are also differences in the software.  You do get what you pay for.
If your serious about your craft and willing to spend money on lenses & bodies be prepared to also send money on the right tools to post process if you want consistency in your results.