Setting brightness

[Steve Smith]

Hello.

I was wondering, my tv’s backlight will only achieve 130 or 115 brightness, and nothing in between…But I want 120.

So, is it better to input 120 and leave the tv at 130, or better to input 120 with tv at 115?

Is there any quality loss when Display CAL lowers the brightness to reach target?

Thanks.

[Florian Höch]

So, is it better to input 120 and leave the tv at 130, or better to input 120 with tv at 115?

Not sure what you’re asking here. If the TV is set up to produce 115cd/m2, then the only way to increase brightness past that is to adjust the TV controls.

Is there any quality loss when Display CAL lowers the brightness to reach target?

Depends. Moderately modern graphics cards usually have >= 10 bits videoLUTs, but if they actually can make use of them depends on many things (graphics cards make and model, connection, operating system, graphics driver version).

[Steve Smith]

Thanks!

Do you know if the Intel i5 processor’s on-board video is capable of 10 bit LUT?  (Displaycal reports 8 bit)

So about calibration: Always wanted to ask this… If there are 255 color possibilities per Chanel, and there are three channels, wouldn’t that mean that all we need to do is sample 765 possibilities of the display’s response? … If all 765 individual responses are corrected then wouldn’t in follow that any mix of these corrected points would also be correct? … Why so many patches?

Can’t wait to get this answer! 🙂

[Florian Höch]

Do you know if the Intel i5 processor’s on-board video is capable of 10 bit LUT? (Displaycal reports 8 bit)

Sorry, no idea. If 8 bits are reported, that’s likely what you’ll get.

So about calibration: Always wanted to ask this… If there are 255 color possibilities per Chanel, and there are three channels, wouldn’t that mean that all we need to do is sample 765 possibilities of the display’s response?

Your math is slightly off there by about 16.7 million 🙂 (256 possible values per channel in 8 bits, to the power of 3)

Why so many patches?

Considering the above a mere few hundred to a few thousand patches is a very sparse sampling, and still usually enough to get very accurate results.

[Steve Smith]

So lets see here… Lets say we take the red channel. There are 256 possibilities of color determined by a specific hue luminescence and saturation for each… ok, so we correct all 256 points.

Then we do the same for blue and green… Doesn’t that mean that no matter how many millions of possible combinations are possible of RGB, that any mix of these already corrected base points will also insure a correct outcome?

Or does mixing change things? … Maybe voltage shifts across pixels can vary when mixing causing a non linear response?

• This reply was modified 7 years, 2 months ago by Steve Smith.

[Florian Höch]

Profiling does not correct anything. All that is does is measure the light emitted by some combinations of R, G and B to determine the actual color (CIE XYZ), creates a device model from these measurements and stores that information so it can be interchanged and used in a defined way.

[Author]

Posts