CHROMiX

How to linearize monitor output

Hi Everybody,

My needs are a little strange as I am not trying to optimize monitor output for viewing. Instead, I need input values, 0-255, to the monitor (say from MatLab) to correlate to luminance values output from the screen in a linear fashion.

I know this sounds a little odd, but I research vision and the stimuli I need to present require a linear relationship between input from the software and output from the monitor.

I have ColorEyes Display Pro (v. 1.42.0) with a DTP-94 colorimeter. I have tried creating a profile using a gamma correction value of 1.0 (I thought this would generate a linear output). However, the resulting profile still outputs luminance values in a nonlinear fashion.

Any suggestions are most appreciated. Thanks!

Though a gamma of 1.0 will provide a measurably linear result in cd/m^2 it is in no way visually linear. The result will, as I imagine you have found out already, be very washed out. What I think you are looking for is to achieve a perceptually linear result, where equal steps in RGB result in equal differences in perceived brightness. To do that you need to calibrated using a color space that is also perceptually linear. For this we have Lab. Lab is a perceptually linear color space and so if the tone response of the monitor is calibrated to match the L curve you will get your visually linear results.

In ColorEyes you are able to set a calibration target of L*. This, I think, is what you are wanting to do. Then, if you are measuring afterwards to validate, you want to look at the L value to verify linearity.

Bob

-----Original Message-----
From: “doc_ck07”
Sent: Friday, October 10, 2008 6:34am
To: MonitorForum@colorforums.com
Subject: How to linearize monitor output

Hi Everybody,

My needs are a little strange as I am not trying to optimize monitor output for viewing. Instead, I need input values, 0-255, to the monitor (say from MatLab) to correlate to luminance values output from the screen in a linear fashion.

I know this sounds a little odd, but I research vision and the stimuli I need to present require a linear relationship between input from the software and output from the monitor.

I have ColorEyes Display Pro (v. 1.42.0) with a DTP-94 colorimeter. I have tried creating a profile using a gamma correction value of 1.0 (I thought this would generate a linear output). However, the resulting profile still outputs luminance values in a nonlinear fashion.

Any suggestions are most appreciated. Thanks!

Post generated from email list

Thanks for the info Bob. I am actually looking for a measurably linear result in cd/m^2. The gamma of 1.0 definitely caused a washed out look, but when I measured output with the colorimeter, it still wasn’t providing a linear output. I also tried using L*. Although it looked much better, it still wasn’t what I was going for as the cd/m^2 result was not linear. Do you know of any reason using a gamma value of 1.0 wouldn’t return a measurably linear output?

Thanks again for the help!
CK