Chroma Subsampling Numbers Explained

By Sareesh Sudhakaran

Disclosure: Links in this post may be to our affiliates; sales through affiliate links may
benefit this site. Please help support wolfcrow and buy from Amazon . It won’t cost you
anything extra.

Everybody talks of luma sampling and chroma subsampling, but few understand what
these numbers really signify.  When someone uses the notations 4:4:4, 4:2:2 and
4:2:0, etc., they usually mean chroma sub-sampling. But what do these numbers
really mean, and why?

How is 1920 x 1080 sampled?

Before the advent of digital HDTV, there was analog HDTV. The problem engineers
faced was that they had to make this higher resolution standard work within the
systems already established, mainly PAL and NTSC.

The major problem they faced, was the marriage between PAL and NTSC. The new
HDTV standard would need to account for both PAL and NTSC as a common factor.
E.g., if PAL is 2 and NTSC is 3, then HDTV couldn’t be 5 or 7 or 8. It had to be divisible
by both 2 and 3, in this case resulting in 6, the lowest possible value.

Standard definition luma (Y) was sampled at 13.5 Mhz, and the lowest common factor
that pleased both PAL and NTSC was 2.25 MHz (6 x 2.25 = 13.5).

This new system, which saw many variations, is now established as having 1125
vertical lines, of which 1080 is reserved for the image. The highest frame rate at the
time for 1080 was 29.97 fps (NTSC), and for 720 was 59.94 fps (NTSC).

What about the horizontal size? The uncorroborated legend goes that 16:9 (1.78:1)
wastes the least space when you want to fit in every kind of aspect ratio, from 4:3 to
2.39:1.

For 1125 vertical lines, the horizontal size for this aspect ratio (16:9) is 2002.
Multiplying 2002 x 1125 x 30 (frame rate), you get 67.57 MHz, which is not divisible
by 2.25 Mhz. The closest that you get is 74.25 MHz, while still keeping the horizontal
size a whole number. In this case it happens to be 2200.

Therefore, 74.25 MHz is the sampling rate for HDTV 1080p at 30fps, and 720p at
60fps. Out of 2200 pixels, we only need 1920 pixels for the image.
For SMPTE 292M (HD-SDI), i.e., 10-bit 4:2:2 1080p29.97 HDTV, the luma is sampled
at 74.25 Mhz, while each chrominance value is sampled at half that, 37.125 Mhz.

For SMPTE 372M (Dual HD-SDI) or 424M (3G-SDI), the sampling rate is doubled, to
148.5MHz, to account for the increased frame rate to 60fps.

Why 4?

Courtesy: Del arte

You’ve all seen 4:4:4, 4:2:0 and 4:2:2. First of all, just for the purposes of this article,
let’s call this method of notation the ’4-sytem’. The 4-system is essentially a digital
notation, which isn’t supposed to have a direct analog counterpart (but it can be
equated or derived from it).

While defining color bit depth, we understood that in order to make digital words, the
formula for the number of combinations with n letters is 2 n. This is the hallmark of
digital numbers.

Let’s assume n = 1.

In this case, in the Y’CbCr model, Y’ can have only 21 = 2 combinations or values. It
can be a 0 or 1. Cb and Cr, too, can have only 2 values each – 0 and 1.

But if Cb and Cr are to be compressed (sampled at a lower rate than Y’ to save

bandwidth), the compressed values will have to be represented by a fraction  – which
as we know in computer land, is a big boo boo.

So n=1 won’t do. The next in line is 2. When n=2, 2 n = 4.

This gives Y’ Cb Cr 4 values each, represented as 4:4:4 when not compressed. If you
want to sample chroma at half rate, you can represent them by 4:2:2. A quarter rate
will give us 4:1:1 or 4:2:0. It gives us four options for each channel in the luma-
chroma model.

Is the number 4 good enough? It works. Like all things in broadcast, this too, is a
compromise. Tomorrow, if someone feels the sampling rate will need to be changed,
this system will become obsolete.

Here are some of the digital chroma sub-sampling values of popular formats:
 HDCAM 3:1:1

NTSC 4:1:1

PAL, DV, DVCAM, HDTV, etc 4:2:0

Internet video 4:2:0

HDTV Broadcast quality 4:2:2

Uncompressed (Full information) 4:4:4

Is 3:1:1 better than 4:2:2?

The old 1080p HDCAM format was 3:1:1, and at the time, broadcast 720p was (still is)
4:2:2. Which one was better?

On the surface, it looks easy. 4:2:2 is double 3:1:1, as far as color sampling is
concerned, so if you’re like me you’d say 4:2:2 is the winner, hands down.

But it’s not that simple. The sampling numbers in the 4-system don’t take into account
the size of the image. Here’s something to ponder over:

Color information in 1080p 3:1:1

The 1080p image size is 1920×1080 = 2,073,600 pixels.

That’s 2,073,600 possible colors.

Since color values are only sampled at one-fourth the maximum frequency,

Cb = 1/4 * 2,073,600 * 1/2 = 259,200 colors

Cr = 1/4 * 2,073,600 * 1/2 = 259,200 colors

Total colors = Cb+Cr = 518,400

Color Density per pixel = 518,400 / 2,073,600 = 0.25 or 25%

Color information in 720p 4:2:2

The 720p image size is 1280×720 = 921,600 pixels.

That’s 921,600 possible colors

Chroma sampling is half the maximum frequency, therefore

Cb = 1/2 * 921,600 * 1/2 = 230,400 colors

Cr = 1/2 * 921,600 * 1/2 = 230,400 colors

Total colors = Cb+Cr = 460,800

Color Density per pixel = 460,800 / 921,600 = 0.5 or 50%

What do you see? The 3:1:1 1080p image has more color information (518,400) than
the 4:2:2 720p image (230,400). But in quality (color density per pixel), obviously
4:2:2 is better than 3:1:1.

So, which is better? An image with more color information, or an image with better
sampled color but lesser color information? I don’t know. The point of this little
exercise was to show you exactly this, that there is more to these things than meets
the eye.

Obviously, if you sample an image at 4:4:4 that gives you the best possible quality
based on your sampling frequency.

Okay, if everything else is equal, is 4:2:2 better than 4:2:0?

Can’t say! You see, in the digital world, you can play a lot of tricks. A manufacturer
can sample an image at 4:2:0, and then repackage it as 4:2:2, just to make up the
numbers.

Nobody can prove them wrong, because they are giving you 4:2:2, just that this might
be an ‘empty shell’, sort of like a Ferrari chassis which is running on a crappy third-
party engine.

The only way you can know for sure whether you’ve been conned or not is to make
tests between the two.

Using chroma sampling numbers to calculate data size

I personally calculate how much information is lost as follows:

Maximum possible quality is 4+4+4 = 12

A full color image is 4:4:4 = 4+4+4 = 12, or 100% of maximum possible quality.
From this, you can derive the rest:
 4:2:2 = 4+2+2 = 8, which is 66.7% of 4:4:4 (12)
 4:2:0 = 4+2+0 = 6, which is 50% of 4:4:4 (12)
 4:1:1 = 4+1+1 = 6, which is 50% of 4:4:4 (12)
 3:1:1 = 3+1+1 = 5, which is 42% of 4:4:4 (12)
 
So, if a 4:4:4 uncompressed frame is 24 MB, then a 4:2:2 frame will reduce to 16 MB,
a 4:2:0 or 4:1:1 image will be 12 MB, and a 3:1:1 image will be 10 MB. Now you know
why chroma sub-sampling is still around. For television and internet video, it reduces
the file size by half, even before any compression has been applied.

I’ve used this calculation methodology in the costs of working with 4K and 2K. It’s not
a foolproof method and only should be used as a general guideline for quick
calculations; but it gets you damn close.

Understanding Terminology – Aspect Ratio and

Pixel Aspect Ratio
By Sareesh Sudhakaran

Disclosure: Links in this post may be to our affiliates; sales through affiliate links may benefit this site.

Please help support wolfcrow and buy from Amazon . It won’t cost you anything extra.

  High Definition (HD) and Standard Definition (SD)

The most common standard found in televisions worldwide is Standard Definition (SD).

Depending on your location on Earth, you might have access to one of these two major SD standards:

 PAL (720×576)

 NTSC (720×480)

In order to enhance the viewer experience by providing greater clarity and visual detail, high definition

systems were introduced, with greater resolutions. There are two broad flavors of high-definition (HD):

 1280×720 (sometimes called HD-ready for who knows what reason)

 1920×1080 (also called Full HD to distinguish it from 1280×720)

Is there anything better than HD? Of course there is. There’s 2K (with a horizontal resolution of about
2048), 4K (horizontal resolution of about 4096), and so on. Digital photo cameras can reach up to 10328

x 7760 (80 MP) in a single sensor, as of 2012.

The future might even bring 8K to homes, with a standard called UHDTV, or ultra high definition

television. Stay tuned, but don’t hold your breath.

Pixel Aspect Ratio

Pixels can be rectangular or square.

Pixel Aspect Ratio = Length of pixel ÷ Breadth of pixel

It is notated by the form a:b. E.g. 1.78:1, 1.33:1, etc. Sometimes, people drop

the ‘:1’ for brevity. It can also be less than one, as in 0.9 for NTSC 4:3.

Why is the pixel aspect ratio important?

The pixel aspect ratio changes the size of an image, and this can be illustrated

with an example.

If an image with the resolution 720×576 has a pixel aspect ratio of 1.0 (square

pixels), then the size of my image is 720×576.

However, an image with the resolution 720×576 but with a pixel aspect ratio of

1.422:1 (length of the pixel is 1.422 times the breadth) will have a size of:

(720 x 1.422) x 576 = 1024×576

This means that even though the actual pixel count horizontally is only 720, by

stretching the pixel a manufacturer can claim it is 1024!

In the above example, it is important to note that the resolution of the image is

still 720×576, but the size of the image has changed to 1024×576 – there is no

additional detail, just a stretched image.

Squashed or stretched footage has its genesis in the selection of incorrect pixel

aspect ratios.

Image Size

Image Size = (Horizontal Resolution × Pixel Aspect Ratio) × Vertical Resolution

The size of an image is not the same as its resolution, but both are notated in the

form HxV, so beware!

 Aspect Ratio

Aspect Ratio = (Horizontal Resolution / Vertical Resolution) × Pixel Aspect Ratio

The horizontal size of the image divided by its vertical size is the aspect ratio of

the image, notated in the form a:b. Aspect ratios are usually rounded off to two

decimal places.

E.g., a 1920×1080 image has an aspect ratio of 1920/1080 = 1.78:1, if it has

square pixels (pixel aspect ratio is 1:1).

What about a 720×576 image with a pixel aspect ratio of 1.422:1? What’s the

aspect ratio of this image?

It’s NOT 720/576=1.25. The answer is calculated as follows:

Horizontal size = (720 x 1.422) = 1024

Vertical size = (576 x 1) = 576

Aspect Ratio = 1024/576 = 1.78:1

Bits vs Bytes, and the story behind Disk Drive

Sizes
By Sareesh Sudhakaran

Disclosure: Links in this post may be to our affiliates; sales through affiliate links may benefit this site.

Please help support wolfcrow and buy from Amazon . It won’t cost you anything extra.

Bits
In computer terminology, the bit is the smallest possible thingy – it’s either a 0 or a 1.

A bit is always represented by the small letter ‘b’.

If I’m shooting 8-bit 4:4:4 1080p, then the total number of bits in one frame is 1920 x 1080 x 3 x 8 =

49,766,400 bits. If I’m shooting at 24 frames per second, then the total bits per second = 1,194,393,600

bits per second (b/s or bps).

One can see how talking in bits is tedious. The numbers are too large. What’s next?

Bytes
Scientists discovered pretty early that bits were too small to manage, even back in the early days of

computing. For better or for worse, they decided to group 8 bits together, which they called a Byte.
A byte is always represented by the capital letter ‘B’.

One Byte (B) = 8 bits (b)

So, our 8-bit 4:4:4 1080p frame which is 49,766,400 bits is also 6,220,800 bytes. At 24 fps, the data rate

is 149,299,200 bytes per second (B/s).

Bits vs Bytes
You’ll find many people using the letters b and B interchangeably, without realizing the ramifications of

their ignorance. Even eminent professionals on the internet do this without realizing their mistake.

MBps is not Mbps – They’re two totally different things. If you write one when you mean the other, you

are only making it worse for everyone. On wolfcrow, I strictly use the notation MB/s to mean Megabytes

per second, while I use Mbps to mean Mega bits per second. I use the ‘/’ notation because some

programs (or humans) might be inclined to make everything lower case or upper case while copying or

quoting me. At least the ‘/’ will still remain!

Just for the record, if you want to know, in traditional science, one uses the ‘/’ to indicate ‘per’, and avoids

the letter ‘p’. On a personal note I hate it when even camera manufacturers and others who should know

better continue to use notations like a fifth-grader. It makes their own engineers look bad.

A long time ago bytes were good enough. However, advancements in computer technology ensured bytes

were just as ‘unwieldy’ as bits.

Kilobytes
In traditional computing, everything is based on the number 2. To get multiples, the formula is 2 n. Even
the byte is 8 bits (23 = 8), and not 10.

Traditional science encourages the use of the word ‘kilo’ to mean 1,000. E.g., if 1000 grams (g) is a

kilogram (kg), in computing terms, they had to choose a number that could be the result of a direct

power of 2.

29 = 512 and 210 = 1024. 1024 was chosen as notation for kilo. Therefore, a kilobyte is not 1000 bytes,

but 1024 bytes.

One Kilobyte = 1024 bytes

So, our 8-bit image which is 6,220,800 bytes is also 6,075 kilobytes. At 24 fps, our data rate is 145,800

KB/s. Remember, KB is kilobytes. If you write Kb, it means kilobits.

Megabytes
A megabyte is not 1000 kilobytes, but 1024 kilobytes.
 One megabyte (MB) = 1024 kilobytes (KB)

So, our 8-bit image which is 6,075 kilobytes is also 5.93 Megabytes. At 24 fps, our data rate is 142.4

MB/s.

Gigabytes
One gigabyte (GB) = 1024 megabytes (MB)

So, our 8-bit image which is 5.93 Megabytes is also 0.0058 Gigabytes. Obviously, for our purposes, it’s

much easier to stop at Megabytes. At 24 fps, our data rate is 0.139 GB/s.

A minute of footage is 8.34 GB. This is how the sizes are used – based on how easy it is to remember or

talk about.

Terabytes
One terabyte (TB) = 1024 gigabyte (GB)

So, two and a half hours of our 8-bit 1080p video will need 8.24 x 150 = 1236 GB = 1.2 TB.

All this seems so easy, right? But just as marketing idiots who print brochures with disregard to correct

notations are guilty of crimes of omission, some manufacturers could be said to be guilty of crimes of

commission.

Imagine this: If I owed you $1,024 but only wanted to give you $1,000, I could call 1,000 1,024. Crazy?

That’s what our next topic is about.

The birth of confusion

First of all, let me make it clear that the full blame for today’s confusion lies both with manufacturers as

well as the scientists who ratified these notations.

Most people couldn’t be expected to know or care that computer engineers loved 1024 more than 1000,

so people began to use the term kilo like they do with kilograms. Many manufacturers added to this

confusion with non-standardized marketing.

Scientists shouldn’t have used the word kilo if they didn’t mean 1000 exactly. Now, half the world cannot

talk to the other half because one side means 1000 while the other 1024.

Did somebody try to correct the problem? Sure. They invented new names. Here’s the modern way, also

endorsed by the IEEE:

 Bit is represented as b (lower case).

 Byte = 8 bits is represented as B (upper case)

 Kilobyte = 1000 bytes and is represented as kB or KB

 Megabyte = 1000 kilobytes and is represented as MB

 Gigabyte = 1000 megabytes and is represented as GB

 Terabyte = 1000 gigabytes and is represented as TB

 Petabyte = 1000 terabytes and is represented as PB

So, what about the 1024 faction? Computers still can only deal with powers of 2. Therefore, we have a

new system for them:

 1024 bytes is now a kibibyte (KiB)

 1024 kilobytes is now a mebibyte (MiB)

 1024 megabytes is now called a gigibyte (GiB)

 1024 gigabytes is a tebibyte (TiB)

 1024 tebibytes is a pebibyte (PiB)

Does this improve matters? Nope. In popular usage, manufacturers and users continue to use the terms

interchangeably. It’s hopeless.

Look at what they’ve done. They’ve called a mile a ‘mili’, while what we all know as a mile is now a new

value because a few corporations feel that will better their profit margins. Now, they can claim their cars

give more miles per gallon, because the length of a mile has been reduced – officially!

Thankfully, this crazy scheme hasn’t been widely accepted. There’s still some common sense left in the

world.

Disk Drive Sizes

Next time you see a hard disk labeled as 4 TB, read the fine print to know whether the manufacturer

means 4 TB or 4 TiB.

What’s the big deal? Look at this:

4 TB = 32,000,000,000,000 bits

4 TiB = 35,184,372,088,832 bits

The difference?

3,184,372,088,832 bits – or ‘only’ about 398 GB!!

Can you see the problem? As disk sizes increase, the seriousness of this difference in terminology

increases in proportion. Stop taking it lightly. Write to your drive manufacturers asking them to clarify

their terms. At the very least, keep this in mind while estimating your disk drive sizes.

As for this website, I avoid the IEEE notification scheme, and stick to the traditional scheme. Whenever I

use kilobytes, megabytes, gigabytes and so on, I’m always using the 1024 system as explained earlier.

How to calculate the size of an Image

Let’s say we have an uncompressed TIFF file (or any other raster image/bitmap file). How do we calculate

the size per frame?

Here’s the formula for file/frame size for uncompressed footage:

Frame size (in bits) = H × V × c × b × s

Where

H = horizontal pixels

V = vertical pixels

c = number of color channels, almost always 3. Monochrome is 1.

b = color bit depth in bits

s = chroma sub-sampling percentage calculated using my formula, expressed as a fraction

Example One: 1920×1080 16-bit 4:4:4

 Image size

= 1920(H)x1080(V)x3(c)x16(b)x1(s)

= 99,532,800 bits

= 99,532,800 / 8 bytes = 12,441,600 bytes

= 12.44 MB or 11.87 MiB

Example Two: 1920×1080 8-bit 4:2:0

Image size

= 1920x1080x3x8x0.5

= 24,883,200 bits

= 3,110,400 bytes

= 3.11 MB or 2.67 MiB.

Example Three: 4096×1716 12-bit 4:2:2

Image size

= 4096x1716x3x12x0.667

= 168,774,009 bits

= 21,096,751 bytes

= 21.1 MB or 20.1 MiB

As you can see, GB is always the bigger number than GiB. Now you understand why manufacturers love

them so much.

Once you have the image size per frame, calculating the data rate is simple – just multiply by the frame

rate. To know how RAW files are calculated, check out Deconstructing RAW. To learn how image sizes

impacts data rate, take a look at Costs of working with 2K and 4K footage. Knowing this is the first step

towards calculating the right disk drive sizes.

Please support wolfcrow and buy from Amazon . It won't cost you extra.

If this post helped you, please buy me a