Define 4:1:1

[james_hutchison]

I'm looking for a definition of the "4:1:1" term I've seen used on

the site, eg. someone mentioned broadcast quality being 4:4:4, and

MiniDV being 4:1:1.

 

Can someone define what the numbers mean please?

 

Thanks

 

Jim

[adrian douglas]

These numbers represent the ratio of brightness(luminance) to colour information(chrominance) broken down into three tracks. The first 4 represents 13.5Mhz luminance(brightness and green) and is the same for all DV formats. The second and third numbers represent the ratio of colour information in the two track compared to the luminance track. So for 4:2:2 you get luminance sampled at 13.5Mhz, and chrominance sampled at 6.75Mhz.

 

What that means is that 4:2:2 video, the broadcast standard, displays 360 colours per TV scanline, or in the case of NTSC DV 180 colours. Sampling at 4:2:2 is used in D-1, DigiBeta, DVCPRO50, BetaSX, and Digital-S. These formats also use 10bit encoding compared to MiniDVs 8bit.

 

To get 4:4:4 you would be looking at film, or Sony CineAltas, and unless you have a spare $100k, MiniDV will have to do.

[ben_gervais]

In case the last answer was too techy, here's a layman's version:

 

4:1:1 and 4:2:2, etc have to do with compression. In a perfect world, when we have all the storage space we could ever need, every pixel would have it's own set of information stored for it. In video, we store information as 3 pieces of information, 1 that tells us the brightness of the pixel, and 2 others that tell us the colour.

 

So, in a perfect world, pixel A would have information x1, y1, and z1. Pixel B next to it would have information x2, y2, z2.

 

Since that is alot of information to store, engineers have noticed that they can 'cheat' with the colour information, and only store one colour for every 2 pixels. Thus, pixel A has information x1, y1, z1 and pixel B has information x2, y1, z1. thus saving a significant amount of storage. This is called 4:2:2, since for every 1 pixel we have about half the amount of colour information.

 

Now taking it further for consumer DV devices, introduce pixels C and D, adjacent to but two lines below (because NTSC is interlaced) A and B. And use the same colour information for them. So C has x3, y1, z1 and D has x4, y1, z1. We've saved even more space! but at the expense of colour or 'chroma' resolution (note that we have still kept all our brightness or 'luma' resolution though). This is called 4:1:1 and is just fine for most consumer uses, but not very good if you want to have a very detailed picture.

 

As noted above, the actual numbers refer to the number of samples, and will be called into account if you are comparing different high end formats. But the ratio is by no means mathematical, and becomes meaningless if you really get into the topic. IE some pro stuff claims 4:4:4, and some 8:8:8 - if it's just a ratio, then these would be the same, but they're not the same. You'll also note that 4:1:1 does not have any indication that, in fact, the next two pixels are below, and not beside the first two. All you really need to know is all other factors being equal, higher numbers are better here.

 

Hope that helps,

-ben