4:1:1 vs 4:2:0 at DVinfo.net

Dennis Davidoff · April 11th, 2005, 02:14 PM

Recently I came across a curious info regarding different compressions in XL2. It seems that XL2 NTSC uses 4:1:1 compression, while PAL version uses 4:2:0.

So, what does that mean for the visible picture quality and how different is it, really? Which is "better"?

Thanks

Joshua Provost · April 11th, 2005, 02:47 PM

Dennis,

Both formats are storing the same amount of color data (1/4 the luma data). However, the store it in differents shapes, if you will. 4:1:1 in long, rectangular blocks, 1x4 pixels. 4:2:0 on square blocks, 2x2 pixels. The same amount of data, in a different configuration.

I don't think that one is necessarily better than the other. However, you must realize that they will result in different types of artifacts. Banding artifacts versus block artifacts.

Now, I have a theory about 4:1:1 being a better format for interlaced video, because for any motion where there will be interlace combing, 4:1:1 is accurate (as it can be) for each line. 4:2:0 encoding of interlace coding could end up being highly inaccurate.

For progressive video, 4:2:0 is likely better, as it distributes it's artifacts equally horizontally and vertically.

With that in mind, XL2 PAL 25p would seem to be a great format for eventual film transfer.

Josh

April 11th, 2005, 02:14 PM	#1
Dennis Davidoff New Boot Join Date: Dec 2004 Location: Vancouver, BC, Canada Posts: 23	4:1:1 vs 4:2:0 Recently I came across a curious info regarding different compressions in XL2. It seems that XL2 NTSC uses 4:1:1 compression, while PAL version uses 4:2:0. So, what does that mean for the visible picture quality and how different is it, really? Which is "better"? Thanks

April 11th, 2005, 02:47 PM	#2
Joshua Provost Major Player Join Date: Sep 2004 Location: Phoenix, AZ Posts: 493	Dennis, Both formats are storing the same amount of color data (1/4 the luma data). However, the store it in differents shapes, if you will. 4:1:1 in long, rectangular blocks, 1x4 pixels. 4:2:0 on square blocks, 2x2 pixels. The same amount of data, in a different configuration. I don't think that one is necessarily better than the other. However, you must realize that they will result in different types of artifacts. Banding artifacts versus block artifacts. Now, I have a theory about 4:1:1 being a better format for interlaced video, because for any motion where there will be interlace combing, 4:1:1 is accurate (as it can be) for each line. 4:2:0 encoding of interlace coding could end up being highly inaccurate. For progressive video, 4:2:0 is likely better, as it distributes it's artifacts equally horizontally and vertically. With that in mind, XL2 PAL 25p would seem to be a great format for eventual film transfer. Josh