View Full Version : 4:2:2 vs 4:1:1 etc.
does anybody have a link or something where these things are explained formally?
07-01-2004, 03:46 AM
I don't have a link off hand (edit: not true, see the bottom), but I do know a tiny bit on the subject...
It's a ratio between the sampling of the different channels in LAB mode. For example, in 4:1:1 compression, the luminance is sampled 4 times the amount that the chrominance is sampled... You get 4 times the luminance info than the colour info.
This is because the human eye contains more of the cell (cone or rod, can't remember which) that is monochrome... So, the movie is compressed to accentuate this property, cause our eyes can discern much more luminance detail than that of colour.
So, if you look close at 4:1:1 video (ie DV) you can see crisp edges with respect to luminance, but the colours will bleed over this crisp edge in a blocky pattern...
Hope that's a good start...
edit: I found you a nice PDF that explains it perfectly... :thumbsup:
Thanks for posting that link - definately an interesting read.... I'm going to have to bookmark it for future reference (I've lost count of the number of times I've seen this question asked...)
<Edited this section of my post, cos I found out that I was talking rubbish.... My apologies for doubting you.... And if you have no idea what I was doubting you about, then I'm not about to tell you ;) >
07-01-2004, 09:53 AM
A ratio used to describe the sampling frequency of a digitized signal. The ratio describes luminance as being sampled 4 times at 3.37 MHz, while color is sampled 1 time at 3.37 MHz in each of it's separate parts. DV, DVCAM and DVCPRO25 use 4:1:1 color sampling. Formulated as: Y (luminance) is sampled at 13.5 MHz (or 3.37 x 4), R-Y (color) is sampled at 3.37 MHz (or 3.37 x 1), B-Y (color) is sampled at 3.37 MHz (or 3.37 x 1) equals 4:1:1.
4:2:2, 4:4:4 and 4:4:4:4 (http://www.videohelp.com/glossary?0#4:2:2,%204:4:4%20and%204:4:4:4)
Put simply 4:4:4 and 4:2:2 terms are descriptions of the sample formats used in digital video. In the early 80's tests were done to determine the sample formats and rates for digital video. The eventual sample structure used for SDI video ended up being 4 times the base sample rate chosen.
The first 4 in the 4:2:2 term is for luminance or the black and white information, and this is where most of the picture detail is. Early tests in television human vision discovered a greater sensitivity to black and white information, while the color is filled in with less detailed areas of the human eye. This means you can reduce the color information and your eye cannot really tell. This is what the 2:2 part of 4:2:2 is for. It means the red and blue channels of the video signal are half the bandwidth of the luminance information. Green is not sent, as you can calculate green from red, blue and luminance information.
This color bandwidth reduction has been used for years in broadcast color television, and in fact the color bandwidth of 4:2:2 is much higher than composite video. This all adds up to 4:2:2 being compatible with black and white or composite television, as the color and luminance information is sent separately, while only 2/3 of the data rate is required for about the same visual quality video.
4:4:4 video is similar, but this time all the color information is sent. RGB computer graphics are really 4:4:4. The 4:4:4:4 format adds a key channel.
*Courtesy of Videoehelp.com
One time I had to correct a terrible compositing for a tv commercial. To save some money they filmed in Beta but did a foolish transfer to mini-DV in order to get it to digital. I did not know why the chroma wasn't working... Advantedge Ultimate saved my life and that commercial... Anyway... later on someone told me the reason why the chroma came out like trash was because the miniDV format has that 4:1:1 compresion. Is that true? Was that the reason? Everything else seemed to be ok.
Is DV and miniDV a format not usable for conventional chorma shots? (I say conventional because with Ultimate you can do everything... not the simple plug-in but the plug-in that opens a software and returns the chroma results)
**chroma = green screen replacement
07-02-2004, 04:05 PM
Is that true? Was that the reason?
Pretty much, yeah... There just isn't enough chroma info.
While the frame may look in tact, that's just because all the luma values are accurate. The resolution of the chroma is (usually) just too low in comparison to get a good key.
This is also why it is suggested to use a green screen when working with DV, because much more of the green information is retained in comparison to blue. Nearly (if not) twice as much info, in fact.
01-18-2006, 03:00 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.
vBulletin v3.0.5, Copyright ©2000-2013, Jelsoft Enterprises Ltd.