[opendtv] Re: Analysis: Should Apple Buy Hollywood?

  • From: Craig Birkmaier <craig@xxxxxxxxx>
  • To: opendtv@xxxxxxxxxxxxx
  • Date: Wed, 29 Feb 2012 16:55:24 -0500

At 3:22 PM -0600 2/28/12, Manfredi, Albert E wrote:
So, do you need 4:4:4 to do a lot better than NTSC, or is 4:1:1 or 4:2:0 also a lot better?
These notations refer to the sampling rates for the three components of digital component video. They do not tell you anything about the sampling system - i.e. interlaced or progressive scan. 

There are two major issues that affect the perception of quality here:

Interlace versus progressive scanning

AND
The bandwidth of the color difference signals and the ability of the digital encoding system to preserve color detail.
I think we can agree now that progressive scan is better than interlace. Interlace is a major contributor to the problems with NTSC, both in terms of the amount of vertical detail delivered and the quality of the color encoding.
In terms of color bandwidth, 4:1:1 is slightly better than NTSC. 4:2:2 and 4:2:0 are MUCH BETTER. Note that the theory behind color subsampling is that we do not perceive as much color detail; studies have found that sampling the color difference signals at 1/2 the bandwidth of the luminance is more than adequate.
Since NTSC is interlaced, the easiest way to capture 1/2 the color bandwidth is to sample EVERY line of both fields at half the bandwidth of luminance. Thus we see 4:2:2 commonly associated with high quality digital component gear. It is far more difficult to sub-sample NTSC in the vertical domain as the color information from two fields must be combined; motion between fields results in a loss of color detail.
If the source is progressive scan it is easy to subsample the color both horizontally AND vertically, hence the somewhat misleading 4:2:0 designation.
MPEG-2 MP@ML specifies 4:2:0 sampling which tends to limit the accuracy of the color samples when there is rapid motion.
Not to stray from the point here, but you don't need as much chrominance bandwidth as luminance bandwidth. That's why S-VHS looked so good. And yet, it didn't create the HDTV market, nor did DVDs. DVDs merely exploited the existence of HDTVs, by introducing, after the fact, after HDTVs had been around for some years, component analog interfaces. Early DVD players did not have component outputs.
Not true. The DVD Video format was first introduced by Toshiba in Japan in November 1996, and in the United States in March 1997. The first (non-test) HDTV broadcasts in the U.S. began in November of 1998.
Toshiba had HDTV home theater set-ups in stores featuring DVD players more than a year before the first regular HDTV broadcasts in the U.S. These sets had analog component inputs and the DVD players had analog component outputs. The ability to display component rather than composite video made a huge difference in image quality. 

Regards
Craig



----------------------------------------------------------------------
You can UNSUBSCRIBE from the OpenDTV list in two ways:
- Using the UNSUBSCRIBE command in your user configuration settings at FreeLists.org 
- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word 
unsubscribe in the subject line.

Other related posts:

  1. Home
  2. opendtv
  3. Archive
  4. 02-2012