Display bit-depth?

[AstroBenny]

1 hour ago, Glenwing said:

Well ultimately the goal of calibration is to get multiple devices to look the same. Mostly it's only important for print media, so that images on the screen look exactly the way they will look when printed out on a printer which is also calibrated.

 

The corners of the color space (full red, full green, and full blue) are the reference points for a color space, since all other colors are combinations of those primary colors, the monitor's entire capability is dictated by what the monitor's most intense red green and blue are. Color space isn't just the boundary colors though, since you could have two color spaces with the same boundaries (same definitions of what color red 255, green 255, and blue 255 are) but different distribution of shades, so that red 160 for example is not the same on both color spaces.

 

There are 256 shades of each color (including off/black) on an 8-bit per channel monitor, numbered 0 through 255, with 255 being the most intense shade of that color. 254 is a slightly duller shade, 253 slightly duller, and so forth. But the 256 shades are not equally spaced, it's a non-linear distribution (because the sensitivity of the human eye to brightness changes is also non-linear), instead the darker shades are closer together, and as you get brighter, the shades get further apart. The exact distribution of shades is determined by a function known as the gamma function. Each color space specifies a gamma number which is used in the formula to set the steepness of the curve, and all 256 shades are set accordingly. A monitor needs to use the same boundary colors as a certain color space and follow the gamma curve defined by that color space to ensure that every possible color is accurate, not just the boundaries, but all the inbetween colors as well.

So is the reference of a colour simply the brightest it will go? You describe the levels of 0-255 as the intensity of the colour.  But does that mean brighter? For example is 255 the brightest colour being produced in a 8 bit channel?

[Glenwing]

4 minutes ago, AstroBenny said:

So is the reference of a colour simply the brightest it will go? You describe the levels of 0-255 as the intensity of the colour.  But does that mean brighter? For example is 255 the brightest colour being produced in a 8 bit channel?

Yes, brighter is more intense. And the brightness setting of a display does affect color accuracy if you're being strict about it  But for most people it's not really much of a concern.

[AstroBenny]

23 hours ago, Glenwing said:

Yes, brighter is more intense. And the brightness setting of a display does affect color accuracy if you're being strict about it  But for most people it's not really much of a concern.

Well, thanks for that

Guess what's next on my list of the understanding-ish mission.

Scaling.

As I understand it, when content is a lower resolution than the native resolution of the monitor, the monitor can apply up-Scaling so that the content can be viewed full-screen. I did watch Linus' fast as possible on upscaling but I still don't get why it works the way it does. Supposedly, monitors use pixel interpolation to guess what the missing  pixels should look since the raw data isn't there.

 

However, what I don't understand is why (my method :P) wouldn't work. Here's how I see it:

Using the example of 1080P content on a 4K display, since 4K contains 4x the number of pixels as 1080P, for every 1080P pixel, couldn't you just spread the colour of that pixel across 4 of the pixels on the 4K display? Why is the inaccurate interpolation approach adopted instead of my 'genius' idea?

[Glenwing]

1 hour ago, AstroBenny said:

Well, thanks for that

Guess what's next on my list of the understanding-ish mission.

Scaling.

As I understand it, when content is a lower resolution than the native resolution of the monitor, the monitor can apply up-Scaling so that the content can be viewed full-screen. I did watch Linus' fast as possible on upscaling but I still don't get why it works the way it does. Supposedly, monitors use pixel interpolation to guess what the missing  pixels should look since the raw data isn't there.

 

However, what I don't understand is why (my method :P) wouldn't work. Here's how I see it:

Using the example of 1080P content on a 4K display, since 4K contains 4x the number of pixels as 1080P, for every 1080P pixel, couldn't you just spread the colour of that pixel across 4 of the pixels on the 4K display? Why is the inaccurate interpolation approach adopted instead of my 'genius' idea?

You could, but that wouldn't work for any other set of resolutions, you'd have to put in a special exception just for running 1080p on a 4K display. Interpolation algorithms work with any two resolutions.

[AstroBenny]

On 08/04/2016 at 11:07 PM, Glenwing said:

You could, but that wouldn't work for any other set of resolutions, you'd have to put in a special exception just for running 1080p on a 4K display. Interpolation algorithms work with any two resolutions.

So how do applications like Photoshop work? Since they have the ability to zoom in and out of an image. Does it rely on scaling from the GPU/Display or does it use its own algorithm of some sort?

[Glenwing]

31 minutes ago, AstroBenny said:

So how do applications like Photoshop work? Since they have the ability to zoom in and out of an image. Does it rely on scaling from the GPU/Display or does it use its own algorithm of some sort?

They use some kind of scaling algorithm of their own choosing. I know GIMP uses linear interpolation, so at 200% zoom each image pixel takes exactly 4 pixels on the screen so the image is still displayed natively, I assume photoshop probably uses the same. For scaling images to different permanently using the resize function, most advanced image editors will let you choose what type of interpolation to use (linear, cubic, etc). I'm not really competent to talk about the differences/advantages between those though.