Bit Depth

Bit Depth – What is a Bit?

Good question – from a layman’s point of view it’s the smallest USEFUL unit of computer/digital information; useful in the fact that it can have two values – 0 or 1.

Think of it as a light switch; it has two positions – ON and OFF, 1 or 0.

bit, Andy Astbury, bit depth

A bit is like a light switch.

We have 1 switch (bit) with 2 potential positions (bit value 0 or 1) so we have a bit depth of 1. We can arrive at this by simple maths – number of switch positions to the power of the number of switches; in other words 2 to the 1st power.

How Does Bit Depth Impact Our Images:

So what would this bit depth of 1 mean in image terms:

Andy Astbury,bit depth,

An Image with a Bit Depth of 1 bit.

Well, it’s not going to win Wildlife Photographer of the Year is it!

Because each pixel in the image can only be black or white, on or off, 0 or 1 then we only have two tones we can use to describe the entire image.

Now if we were to add another bit to the overall bit depth of the image we would have 2 switches (bits) each with 2 potential values so the total number of potential values, so 2 to the 2nd, or 4 potential output values/tones.

Andy Astbury,bits,bit depth

An image with a bit depth of 2 bits.

Not brilliant – but it’s getting there!

If we now double the bit depth again, this time to 4 bit, then we have 2 to the 4th, or 16 potential tones or output values per image pixel:

Andy Astbury,bits,bit depth

A bit depth of 4 bits gives us 16 tonal values.

And if we double the bit depth again, up to 8 bit we will end up with 2 to the 8th power, or 256 tonal values for each image pixel:

Andy Astbury,bits,bit depth

A bit depth of 8 bits yields what the eye perceives to be continuous unbroken tone.

This range of 256 tones (0 to 255) is the smallest number of tonal values that the human eye can perceive as being continuous in nature; therefore we see an unbroken range of greys from black to white.

More Bits is GOOD

Why do we need to use bit depths HIGHER than 8 bit?

Our modern digital cameras capture and store RAW images to a bit depth of 12 bit, and now in most cases 14 bit – 4096 & 16,384 tonal values respectively.

Just as we use the ProPhotoRGB colour space to preserve as many CAPTURED COLOURS as we can, we need to apply a bit depth to our pixel-based images that is higher than the capture depth in order to preserve the CAPTURED TONAL RANGE.

It’s the “bigger bucket” or “more stairs on the staircase” scenario all over again – more information about a pixels brightness and colour is GOOD.

Andy Astbury,bits,bit depth,tonal range,tonality,tonal graduation

How Tonal Graduation Increases with Bit Depth.

Black is black, and white is white, but increased bit depth gives us a higher number of steps/tones; tonal graduations, to get from black to white and vice versa.

So, if our camera captures at 14 bit we need a 15 bit or 16 bit “bucket” to keep it in.  And for those who want to know why a 14 bit bucket ISN’T a good idea then try carrying 2 gallons of water in a 2 gallon bucket without spillage!

The 8 bit Image Killer

Below we have two identical grey scale images open in Photoshop – simple graduations from black to white; one is a 16 bit image, the other 8 bit:

Andy Astbury,bits,bit depth,tone,tonal graduation

16 bit greyscale at the top. 8 bit greyscale below – CLICK Image to view full size.

Now everything looks OK at this “fit to screen” magnification; and it doesn’t look so bad at 1:1 either, but let’s increase the magnification to 1600% so we can see every pixel:

 

Andy Astbury,bits,bit depth,tone,tonal range,tonal graduation

CLICK Image to view full size. At 1600% magnification we can see that the 8 bit file is degraded.

At this degree of magnification we can see a huge amount of image degradation in the lower, 8 bit image whereas the upper, 16 bit image looks tonally smooth in its graduation.

The degradation in the 8 bit image is simply due to the fact that the total number of tones is “capped” at 256. and 256 steps to get from the black to the white values of the image are not sufficient – this leaves gaps in the image that Photoshop has to fill with “invented” tonal information based on its own internal “logic”….mmmmmm….

There was a time when I thought “girlies” were the most illogical things on the planet; but since Photoshop, now I’m not so sure…!

The image is a GREYSCALE – RGB ratios are supposedly equal in every pixel, but as you can see, Photoshop begins to skew the ratios where it has to do its “inventing” so we not only have luminosity artifacts, but we have colour artifacts being generated too.

You might look upon this as “pixel peeping” and “geekey”, but when it comes to image quality, being a pixel-peeping Geek is never a bad thing.

Of course, we all know 8bit as being “jpeg”, and these artifacts won’t show up on a web-based jpeg for your website; but if you are in the business of large scale gallery prints, then printing from an 8 bit image file is never going to be a good idea as these artifacts WILL show on the final print.

Become a patron from as little as $1 per month, and help me produce more free content.

Patrons gain access to a variety of FREE rewards, discounts and bonuses.