timoth

One thing you might want to try is to use Image Cache at the lowest setting of 1. Sure it is

a trade off. Some things run slower, Auto Levels don't work, and you have to close the

histogram view to get PS to have a good responce time, but some things run faster and

you get a screen preview generated from a 16bit image instead of the 8bit image cache.

This will yield much better quality in the darks when editing at zoomed out views such as

25%.

Timothy Farrar - http://farrarfocus.com

Some LCD's suffer more than others from the viewing angle problem. My 17inch

Powerbook G4 has serious viewing angle problems. Even at the same angle the screen

gamma is inconsistant from the top to the screen. The other, and perhaps the most

serious, problem with LCD's is screen uniformity. Basically the intensity and even the

color shifts across the screen. Usually best in the center and gets worse as you get

closer to the edges. Sometimes the problems of viewing angle and uniformity are hard

to seperate. In the case of the 910T the viewing angle does not seem to be a problem,

however it does have uniformity issues.

To put this in perspective,

Looking at uniformity results from various expensive ($1500+) LCD reviews on

www.tomshardware.com,

The 23" Apple Cinema has uniformity that starts at 80% on the edges and goes to 95

-100% for the center of the screen. The Eizo L997 ranges in the 100-95% range for

most of the screen and down to 85-90% in one corner. The LaCie 321 seems to have a

rather poor 80-85% for a good amount of the screen.

From my understanding, it seems as if Delta-E is the difference between two colors in

LAB color space, so in terms of a brightness difference, a Delta-E of 1 is like a 1%

brightness difference between two colors. The tomshardware tests use a photodiode

to check uniformity, so the 0%-100% of the uniformity tests is intensity instead of

brightness. To convert intensity (I) into brightness (L) is, L=116*((I/100)^0.33)-16. So

here would be the translations of uniformity=deltaE: 95%=2.0, 90%=4.0, 80%=8.3.

On my inexpensive 910T screen I went and checked the maximum delta-E as close as

I could get to the 4 corners of the screen: 2.29, 2.73, 3.15, 4.43. The worst uniformity

I could measure was about 89%. It seems to do on par with the Pro LCD's.

I no longer have a good CRT to test uniformity with. However my opinion is that this is

one area where the CRT is many times better than even the best Pro LCD. Even though

you can calibrate a large part of the LCD screen well, the edges are always going to be

visably out of calibration.

In practice, if you have a LCD (such as a laptop) that has viewing angle problems, use

something like a Adobe Gamma style background image so that you can check that

you are always editing at the same viewing angle. If you want I can give you some

really good tile-able backgrounds that show a very visable color shift based on

viewing angle. This is what I use on the road for my laptop.

- Timothy Farrar -- Farrar Focus

For those who have not color calibrated a display, my previous post will not be of

much help. However one thing to remember is that typical LCDs from the factory are

usually way off in color calibration so you will have to do this anyway to get good

results.

Below I have included some links to full reviews of 2 pro LCD monitors (one LaCie and

one of a $1900 Eizo). Keep in mind that color calibration of all displays is much

tougher at low color temps (like 5000K).

From the results I got, it seems as if the 910T at 5000K calibrated to about as good as

the $1900 Eizo with 14bit LUTs at 6500K. Except the 910T might have a slightly

lighter black point. So the combination of good LCD (16.7 not a 16.2) and good color

calibration using a VGA cable (and hence the video card's LUTs) can yield extremely

good results.

Here are the web sites showing the 910N, http://www.behardware.com/articles/534

-8/comparative-test-of-low-priced-8-12-16-and-25-ms-19-lcds.html, a LaCie321,

http://graphics.tomshardware.com/display/20050706/professional_lcd-08.html, and

an Eizo FlexScan L997, http://graphics.tomshardware.com/display/20050830/

pro_lcd-06.html.

Hope that helps for anyone who is looking for a good LCD for proofing.

- Timothy Farrar -- Farrar Focus

From Fry's Electronics, the 19" Samsung SyncMaster 910T LCD is $399. It is a true

16.7 (8bit per channel display). I calibrated this monitor using ColorEyes Display using

the X-Rite puck on a Mac Powerbook G4, targeting D50 (5000K) and using L*

(perceptual similar to gamma 2.5) instead of using a specific gamma (like 2.2 or 1.8).

I used their 16bit LUT ICC profile output choice.

This LCD does NOT have a 10bit LUT, so I didn't use the DVI (digital) cable. Instead I

used the standard VGA cable to the monitor so the video card's high precision LUTs

are adjusted to calibrate the monitor.

Here are the settings I ended up with on the display controls,

Contrast 80, Brightness 80, Red 50, Green 35, Blue 18,

Gamma choice 3 (the highest gamma / darkest output).

And the results after verification with the software.

Maximum Delta-E -> 0.43.

Average Delta-E -> 0.20.

Luminance -> 185 cd/m2.

And looking at the ICC profile using ColorSync.

Contrast Ratio (Y ratio of white to black point) -> 370:1.

Which I believe would put the black point at somewhere around 0.5 cd/m2.

The color gamut looks about the same size as a very good CRT.

So do you need $1700 to get a good LCD? Perhaps not.

I used the following internet resources when looking for an LCD display,

http://graphics.tomshardware.com/display/index.html.

http://www.behardware.com/.

They both had real calibration results and seemed creditable.

One has a review comparing the LCD to CRT.

The 910T was not in any reviews, however it's sister the 910N (no DVI input) was.

- Timothy Farrar -- Farrar Focus