View Single Post
Old 09-12-2006, 06:06 AM   #5
ath
Addict
ath doesn't litterath doesn't litter
 
Posts: 222
Karma: 110
Join Date: Jun 2006
Location: Malmo, Sweden
Device: iLiad, Sony PRS-505, Kindle Paperwhite & Oasis
Quote:
Originally Posted by Malder1
Thanks for new photos!
According first one, iLiad's white is 3 from 22 (1..22). It's a very good result. Like usual white book's page.
I should have explained better: I was not primarily out to get an absolute value of the grey level of the screen of the iLiad, but rather to suggest that very 'white' photos may have been (intentionally or unintentionally) biassed. (I can't decide if the third? photo/thumbnail was taken of both devices at the same time: that white area to the right seems suspicious.)

Difference between two screens can be compared only if both screens are shown in the same photo. Establishing the absolute difference needs a standard target in the same photo. (Even then, it may not give an accurate idea of what the screen looks in *your* light.)

But if you want absolute level ... 3 is far too high. I placed it at around 80% myself, but that was not based on a very careful inspection, as I know I have two level of screen adaptions going (one Adobe gamma adjuster, and one Nvidia adjuster) that can affect my perception of the images.

Some lab work follows. I'm checking histogram of a manually selected part of the original image (i.e. before they were resampled for the web. Resampling was 'intelligent' which probably means that things may have changed). I use the L value (lightness) of the HSL model as an approximation. (There should be no difference in using either greyscale model, or converting the image to greyscale before checking the pixel value -- and indeed repeated testing this way produce the same values. )

A) Iliad + target picture:

White box has L median 227, black box 31. A selection of the white area below the first paragraph gives 161. Box 7 median is 158, box 6 median 169, so it's between levels 7 and 6. The HSL lightness of the best part of my screen in this photo is approx. 71% (i.e. 161/227).

B) Book + target picture:

White box: 193, black box 3 (so either daylight changed, or my camera rebalanced the image: probably a little of both). A selection from the central part of the right margin of the page has a L median of 188, which ends up as HSL lightness 97%.

Based on this, the iLiad screen should not be described as paper white -- at least not the paper I used as comparison. It ends up at a little better than 2/3.

(In each case, the histograms of the selected image area are symmetrical and fairly 'tight', so median should be about the same as average. It's first when I select the entire book or visible iLiad page that the histogram gets clearly lopsided towards lower values, and so would give a obviously lower average than median.)

However, for reading ease, I don't think paper whiteness means as much as contrast between foreground (text) and background (page). I have no idea of how to measure that, though, as colour has to be taken into account (black text on light yellow paper reads better than black on pure white, for instance).

Last edited by ath; 09-12-2006 at 06:11 AM.
ath is offline   Reply With Quote