AVS Forum banner
1 - 6 of 6 Posts

·
Registered
Joined
·
189 Posts
Discussion Starter · #1 ·
I propose a new "universal" way to evaluate contrast ratio & black levels across displays, to make things somewhat fair & objective:


-all displays must be professionally calibrated to get full black detail (no crushing).

-all displays must be tested in completely light controlled room (dark)

-any other lightsources (LEDs from the display or other components) be taped over

-ambient light must be low enough as to not be detectable by the sensor used.


3 measurements then take place:

Measurement 1 - entire screen displaying black image. Measure actual light output level (if there are hot zones use middle of screen). Call this number Black level.

Measurement 2 - pure white image. Measure actual light output level (if there are hot zones use the middle of screen) Call this number White level.

Maximum contrast: difference between Measurement 1 & 2

Measurement 3 - white/black checkered pattern. (1 pixel wide). Run the difference between a black pixel & a white pixel. This is close to ANSI contrast right?


What this method will give is: objective, fair black level, brightness level, contrast ratio & ansi contrast ratio (light bleeding) for all displays under 100% optimal conditions for each & every display. I'd expect the actual recorded numbers to be quite different from the manufacturer's specs in these cases.


Footnote: so-called dynamic irises can be turned on if desired for an additional set of numbers but cannot be directly compared to baseline measurements.
 

·
Registered
Joined
·
2,477 Posts

Quote:
Originally Posted by lexx_kun /forum/post/0


I propose a new "universal" way to evaluate contrast ratio & black levels across displays, to make things somewhat fair & objective:


-all displays must be professionally calibrated to get full black detail (no crushing).

-all displays must be tested in completely light controlled room (dark)

-any other lightsources (LEDs from the display or other components) be taped over

-ambient light must be low enough as to not be detectable by the sensor used.


3 measurements then take place:

Measurement 1 - entire screen displaying black image. Measure actual light output level (if there are hot zones use middle of screen). Call this number Black level.

Measurement 2 - pure white image. Measure actual light output level (if there are hot zones use the middle of screen) Call this number White level.

Maximum contrast: difference between Measurement 1 & 2

Measurement 3 - white/black checkered pattern. (1 pixel wide). Run the difference between a black pixel & a white pixel. This is close to ANSI contrast right?


What this method will give is: objective, fair black level, brightness level, contrast ratio & ansi contrast ratio (light bleeding) for all displays under 100% optimal conditions for each & every display. I'd expect the actual recorded numbers to be quite different from the manufacturer's specs in these cases.


Footnote: so-called dynamic irises can be turned on if desired for an additional set of numbers but cannot be directly compared to baseline measurements.

I disagree with "-all displays must be tested in completely light controlled room (dark)" The 'average' viewer does not spend all their viewing time in a dark room. Perhaps the videophile who can afford a dedicated home theater room would appreciate the dark room but IMHO the average viewer will have some lights on in the evening and will also watch tv during daylight hours.
 

·
Registered
Joined
·
625 Posts
Indeed I too agree. Rather than testing the display in an artificially dark environment, what ought to be done is to test it in a series of ambient brightness steps, maybe about 10-50 lux each, starting from 0 to maybe 200 or more.


A display which appears to have better black level than another in a perfectly black environment, will often have poorer black level in a lit environment.


By plotting a curve which measures black level against ambient illumination, and comparing these curves for different displays, it becomes possible to determine what the best display would be for a given ambient luminous flux average.


If we take two devices, A and B, their ALECs (ambient light ejection curves) when displaying a black image, may be parallel or convergent, with a crossing inside or outside the area of reasonable environmental luminances.


If a crossing exists, this is known as a 'Black Crossing'. The BC is a most Gothic concept, yet is also eminently practicable. A display which has lower black level than its partner under examination below BC luminance, will have higher black level above the BC.


If the two devices under comparison are LCD and plasma, it is pretty certain that a crossing of the curves will exist. What's most important, is at what environmental luminance that crossing occurs. The location will determine device suitability for the given environment.
 

·
Registered
Joined
·
3,186 Posts
Way too complicated and it would only make things worse IMO. Do you think that most consumers will understand this??


They need to simplify things. Get rid of dynamic contrast, static contrast, ansi contrast etc which are confusing, ambiguous, and most importantly not repeatable.


Settings are so dynamic between different displays there is no way to use a standard image reference or pattern to determince contrast in any environment. It is just ludicrus.


The only way to simply and objectively differentiate between displays is to start with dynamic range. All that I would need to know to determine the capabilities of the display with regards to contrast would be:


Dynamic Range

Peak White

Black Level

Reflectance


All of these measurements can be made objectively and should be repeatable regardless of who takes the measurement. For instance at work here we use a reflectance measurement that can be confirmed by any photonics lab in the world.
 

·
Registered
Joined
·
625 Posts
I'll revise my proposal: three measurements of both contrast and black level should be made: the first at ambient light level of 0, the second at one typical of indoor daytime household, and a third at high ambient light level, which would correspond to some houses I've seen with lots of windows.


The current published values by Home Theater magazine and others are highly misleading, because they were obtained in an environment that is highly atypical of the average. People read those numbers and make decisions based on them, which will be incorrect.


Many plasma displays which show better contrast/black level at zero ambient illumination, will fall far behind their LCD cohorts when tested at average day ambient flux.


So the numbers published are only applicable to a small fraction of users and usage times, namely those that watch in the pitch black at night. Yet these numbers are being used to recommend displays for totally different uses.


Here is the master formula for black level:


B = E + R


Black level = Emission + Reflection


expanding this formula:


B = E + (r * a)


Black level = Emission + (reflection coefficient {aka. reflectivity} * ambient flux)


The key parameter here is r, the reflection coefficient. Plasmas have high values, while LCDs by their nature have small r values. Having a higher r value means having a steeper ambient luminance vs. black level curve:




The graph is a bit deceptive because it implies that the Black Crossing of a typical LCD/Plasma pair is somewhat high. Rest assured, that is just an artifact of my hastily thrown-together graph.


I predict that if a typical pair of such devices are place side-by-side, both on and showing black images, that the B values will converge at a fairly low ambient luminance level.


So this is an open challenge to Home Theater magazine, and any other testers out there who have both LCD and Plasma displays which can be tested side-by-side with three or more different environment flux levels: do the tests, and report the results. I'm hungry for some real-world data to back up the theory!


In addition, I'd like to request that Home Theater magazine re-test as many displays as possible, providing their black level and contrast ratio values at typical interior ambient luminance, and above-average ambient luminance.


It's important to remember that even though the arbitrarily chosen 'mid-range' luminance for testing is indeed arbitrary, it will still be closer to the average user's environmental luminance than the current 'standard' of 0 lux. Total blackness is, apart from crazily brilliant whiteness, by definition as far as possible from the average user's envirolume.


Testing displays as far as possible outside their real-world usage context of course results in values that are applicable only to the unnatural test condition that gave them birth. And that, my friends, is not something most people are even thinking when they read those deceptive numbers. It's called lying by omission.
 

·
Registered
Joined
·
189 Posts
Discussion Starter · #6 ·

Quote:
Originally Posted by xrox /forum/post/0


Way too complicated and it would only make things worse IMO. Do you think that most consumers will understand this??


They need to simplify things. Get rid of dynamic contrast, static contrast, ansi contrast etc which are confusing, ambiguous, and most importantly not repeatable.


Settings are so dynamic between different displays there is no way to use a standard image reference or pattern to determince contrast in any environment. It is just ludicrus.


The only way to simply and objectively differentiate between displays is to start with dynamic range. All that I would need to know to determine the capabilities of the display with regards to contrast would be:


Dynamic Range

Peak White

Black Level

Reflectance


All of these measurements can be made objectively and should be repeatable regardless of who takes the measurement. For instance at work here we use a reflectance measurement that can be confirmed by any photonics lab in the world.

Aye, I like that proposal. Though ANSI contrast is somewhat important - when light bleeds out from pixel to pixel it hurts high-contrast scenes.
 
1 - 6 of 6 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top