Re: Verifying calibrated gamma

Hi Gernot,

I just want to make sure that you understand that I said 1% at dac 31 ( IRE
of approximately 12) This will typically yield a contrast ratio in excess of
300:1 and on a very good display 500:1 measured as Sequential contrast, Full
Black measurement / Full white measurement . None of this contridicts Alan
Roberts' fine article. In actual practice, a well formed profile should
bring this level in the rendered image properly. It is important that the
actual screen contrast is much higher than the value required for the
proofing operation....

As I said, the effect of a mal-formed profile is probably the biggest
problem in screen to print. I stand by that argument and it is easily shown
in practice.

--
Tom Lianza
Director of Display and Capture Technologies
GretagMacbeth LLC
3 Industrial Drive
Unit 7&8
Windham, NH 03087
603.681.0315 x232 Tel
603.681.0316 Fax


"Gernot Hoffmann" <hoffmann@xxxxxxxxxxxx> wrote in message
news:1148101383.802277.263340@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Tom,

thanks for your explanations:
---
"Setting the black point to the blackest point possible is not a
good general practice.
As a rule of thumb, in a gamma 2.2 system, the level at Dac 31
should be nominally 1% of the full scale luminance :
.00969 = (31/255)^2.2
This 1% set point has a number of advantages over just letting
the black fall visually to a dark value. This procedure will yield
a proper response over two decades of response and it happens
at a luminance value that falls
within the range of most low-cost display calibrators can
measure with reasonable accuracy. Adjustment of the bright-
ness control on the CRT has a large impact on the display
gamma, it is not just an additive adjustment that sets an offset
value. This is because the offset value is essentially summed
inside the parentheses and then raised to a power.
What I have found is that a gamma law relationship holds
extremely well down to this point and below this point there
will be substantial departure from a generic model.
Another major factor in this region is the graphic card performance.
The offset voltage from the graphic card is not well specified
and can sometimes be negative. It is not unusual to see display
cards with missing values or large non-linearities in this sub
1% region."
---

IMO, my doc

http://www.fho-emden.de/~hoffmann/measgamma10022004.pdf

is not in contradiction to your statements, concerning the TRC
models and practical results. One exception:
My recommendation: black level luminance at 0.35...0.50%
of white level luminance, instead of 1.0%.
A contrast ratio of 200 (for 0.50% black level) is IMO realistic
for quality CRTs.
I'm using several EIZO CRTs and one LaCieBlue19IV CRT.

These issues should be taken into account as well:
1. Uncertainties for Blue because of low absolute luminance.
2. Varying luminance, depending on the position on the screen.
3. Varying luminance, depending on the size of the measured
patch and the value for background gray.

Your comments about the accuracy of 8-bit-DACs are surprising.
I would have expected offset and differential nonlinearity less
than 0.5 LSB.

I'm not sure whether there is something like a 'negative output
of a DAC'.
The relevant content in a video signal is anyway shifted by the
reference black level, according to this explanation:
http://tinyurl.com/l5mrr

Best regards --Gernot Hoffmann



.

Relevant Pages

  • Re: Compaq display problem.
    ... appears "washed out" just like a CRT display with the contrast turned away down. ... You may even be able to pick up a complete screen/lid assembly quite cheaply. ... Remove them and gently pry just between the keyboard at one end and the cover should pop up. ...
    (comp.sys.laptops)
  • Re: All Office Programs Open in Low Resolution
    ... Contrast display with the keyboard shortcut ... Toggle the Windows Theme and color schemes in Windows XP ... The display settings in Control Panel are set to high resolution and all ...
    (microsoft.public.office.misc)
  • Re: TFT contrast problem
    ... I'm trying to interface Casio 3,5" TFT display to ... When bar width is increased contrast is bad ... The configuration utility was not able to ...
    (comp.arch.embedded)
  • Re: Why pics are so dark on my computer?
    ... Some video card makers both provide control ... you to make changes to your gamma, brightness, and contrast on a software ... What kind of display are you using (traditional CRT, a flat screen LCD, ...
    (microsoft.public.windowsxp.general)
  • Re: TFT contrast problem
    ... I'm trying to interface Casio 3,5" TFT display to ... When bar width is increased contrast is bad ... The configuration utility was not able to ...
    (comp.arch.embedded)