Does alpha channel really need 8-bits?

[Flashbond]

Hi,

I am looking for the differences between 16-bit, 24-bit and 32-bit colors for a while. I saw that actually 32-bit color is also like 24-bit color but plus 8-bit alpha channel. Since 8-bit information represents 256 possible values for each color channel, what does 8-bit do exactly for alpha channel to store color or transparency info?

I hope I didn't ask a stupid thing. Thanks a lot!

 

[berkeman]

Hi,

I am looking for the differences between 16-bit, 24-bit and 32-bit colors for a while. I saw that actually 32-bit color is also like 24-bit color but plus 8-bit alpha channel. Since 8-bit information represents 256 possible values for each color channel, what does 8-bit do exactly for alpha channel to store color or transparency info?

I hope I didn't ask a stupid thing. Thanks a lot!

For those folks (like me) who had no idea what you are asking about, here is a link: https://en.wikipedia.org/wiki/RGBA_color_space

Are you asking specifically about Photoshop? This thread probably belongs in one of the technical forums -- Physics or Computing?

EDIT -- Thread moved by the Mods to Computing.

 

[Flashbond]

Ah, ok I got my answer. It holds the percentage information of the alpha channel. This makes sense. Otherwise 8-bit channel badwidth wouldl be very unneccesarry if it was only holding transparent or not. Thanks a lot!

 

[berkeman]

Hey, always happy to help...

 

[newjerseyrunner]

It's exactly the same as the other components, why would it be any different? r is a 0-255 value for how intense the red part of the color is, a is how intense the alpha is.

 

[glappkaeft]

The alpha channel could use any depth from not existing, 1 -bit and up depending on you needs.

 

[newjerseyrunner]

The alpha channel could use any depth from not existing, 1 -bit and up depending on you needs.

You can’t use 1 bit for alpha. That’s not alpha, that’s masking. 1 bit would give you the option of having the pixel fully there or not at all, no transluscency at all.

 

[Flashbond]

You can’t use 1 bit for alpha. That’s not alpha, that’s masking. 1 bit would give you the option of having the pixel fully there or not at all, no transluscency at all.

Yes, in this case I must agree with @newjerseyrunner because alpha must able to have at least 101 value in order to represent transparency percentage.

 

[newjerseyrunner]

Yeah, but because engineers tend to think in base 16 rather than base 10, 100 would be completely arbitrary. Thinking of parts per 256 is simply easier when you do base 16 stuff all day.