dvx at 2:4:4:12:45:67:32

PaPa

PaPa

Major Contributor
guys, what does all this mean whenevre i see 2:4:4:4:4:4. please explain it in a way for someone who is not technically literate.

is this the same thing as a pulldown of 3:2:23:4345 or whatever

if there are more than one uses for this terminology, please explain what they are :)!

-Jon PaPa
 
No, 2:3:3:2 or 2:3 pulldown has NOTHING to do with 4:4:4, 4:2:2. 4:2:0, 4:1:1, or any other combination involving 4's. :D It's basically the ratio of color to brightness values in a given video stream. Basically you can throw out color information (thus lowering the bandwidth of the video stream) without losing much quality. But if you throw out brightness values the human eye will perceive a larger degradation of quality. So, the ratio is just that. Brightness (the beginning 4) compared to the amount of chroma information. DV is 4:1:1, meaning the color information compared to the brightness values is really low. 4:4:4 is the best quality you can get, as all of the values are equal in comparison to each other. That's probably still a little technical, but it's something to start on.


With all of that said, the 2:3:3:2 or just 2:3 pulldown is related to the framerate (specifically 24P). The DVX has to use the pulldown in order to make the frames record at the right rate in a 60i video stream. That's a whole other topic.

Read this to get a better understanding of the 4:4:4, 4:1:1 ratios.
 
Last edited:
4:4:4 = RGB color space. Individual color is recorded for every pixel. Rarely ever used in any sort of video recording system. For the purposes of this discussion, this will be the sample image we start with:
Color-Sampling-4-4-4.JPG


In that image, every pixel has its own unique color recorded. You would normally never see a video image with that much individual color detail; normally you'd only see something like that in a computer graphics program.

For video, the eye is much less sensitive to color than it is to brightness. So video engineers long ago figured out that they could "cheat" on the color and save a lot of bandwidth and nobody would really notice. So they invented color subsampling.

First up is 4:1:1. This is the sampling system used in NTSC DV. In 4:1:1, each block of four pixels are all set to the same color. Brightness information is recorded for each pixel individually, but the color information is forced to all be the same (an average of the color of each of the four pixels). So if you took our original image here:
Color-Sampling-4-4-4.JPG

and processed it through a 4:1:1 compression system, you'd get this result:

4-1-1-Illustration.JPG


See how that works? Each group of four pixels is forced to be the same color. You can differentiate between the pixels because they have different brightness values (the pixels in the upper right are all green, but they're different shades of green because the brightness values are different per pixel). That's how DV works. That's why chroma keying is not so easy on DV, and why if you shoot video where red and green edges meet, it looks blocky; because the DV color system uses large 4-pixel-wide color samples. But, in 4:1:1, color samples are recorded individually on each line. So 4:1:1 ends up making 4x1 color samples.


Now, 4:2:0 is how DV is encoded in PAL. It also uses one color sample per block of four pixels, but instead of doing 4 wide by 1 tall, 4:2:0 does 2 wide by 2 tall (2x2). It looks like this:

4-2-0-Illustration.JPG


You end up with about the same overall color resolution, just in a different way. 4:2:0 is used in PAL DV, and in DVD, and in HDV. It works okay in progressive, but it's fairly ghastly in interlaced encoding. And 4:2:0 gives you the same chroma keying problems that 4:1:1 does, just in a different way -- you have blocky fat pixels vertically, vs. 4:1:1's fine resolution vertically but blocky fat pixels horizontally. Either way, 4:1:1 and 4:2:0 both deliver the same amount of chroma resolution.


Then there's 4:2:2. 4:2:2 gives you twice as much color resolution as the other systems. It sets one color sample for each two pixels (2x1). Much more color accuracy, much better chroma keying, better sharpness, better everything. 4:2:2 is used in professional formats like Digital Betacam and DVCPRO-HD and DVCPRO50.
4-2-2-Illustration.JPG
 
Last edited:
... i havent read your post yet, i will, i promise, have to go back on the floor.

but wow, barry posted in my thread.. lol
 
You must log in or register to reply here.
Back
Top