[opendtv] Re: Digital vs. Analog Quality

  • From: Craig Birkmaier <craig@xxxxxxxxx>
  • To: opendtv@xxxxxxxxxxxxx
  • Date: Wed, 9 Jul 2008 08:41:12 -0400

At 1:56 PM -0400 7/8/08, Cliff Benham wrote:
Craig Birkmaier wrote:
...baseband NTSC could be
far better than transmitted NTSC. True for the luminance, but not true for the color difference signals.

I think there are several examples of wideband color difference signals being used in NTSC. As far back as the late 1970s, baseband NTSC has been capable of 120 lines of chroma resolution. This includes Laserdisk and later in the early 1980s, analog component recording formats, notably Betacam.

Neither of these are valid examples.

Laserdisc is an analog format with extended frequency response for the luminance component - in the range of 4.9 to 5.4 MHz. Laserdis uses the same subcarrier based system for the carriage of the color components. I found one reference that suggests that the internal recorded bandwidth for the subcarrier signal is 1.5 MHz, qwhich could allow for improved color recording in the Q channel. The system is said to provide 120 lines of color resolution, which is the spec cited for NTSC (but not for all colors because of the 0.5 MHz bandpass for Q). It may have had a bit more resolution internally, but the outputs are baseband are limited by the NTSC specification.

see: http://home.online.no/~espen-b/ld/

Betacam is an analog component format with R-Y and B-Y sampled at half the bandwidth for the luminance channel - i.e. 4:2:2. It provided a HUGE improvement over composite video recording formats for several reasons:

1. NO mixing of the color and luma components as with subcarrier-based systems. The color components are completely recoverable without the problems of separating luma and subcarrier via comb filtering techniques.

2. The R-Y and B-Y color difference signals are sampled at 1/2 the bandwidth of the luminance signal This is more than double the color information in NTSC (about 8x the bandwidth of the NTSC Q component.

Alas, when a Betacam machine encodes the analog component signals into a baseband NTSC output, the color signals are bandpass limited to NTSC specs. The good news is that the color difference signals are "oversampled" relative to NTSC and can be recovered accurately - thus the composite output is MUCH cleaner than for any heterodyne color recording system.

The BREAKTHRU with Betacam was the availability of analog component outputs, which could be used to feed an analog component video mixer that did not suffer from the limitations of composite video processing. I am reasonably knowledgeable about this, having built a prototype analog component mixer for Sony for the 1983 NAB. This was followed by an RGB 1600 series mixer, which both Sony and Grass Valley used at NAB in 1984 - This was the year we had two Eiodophor projectors displaying analog component video on huge screens above the NAB booth. It was also the year that we introduce the Model 100. After building the 1600 series mixer using RGB processing we realized that these were the wrong components to process - it was difficult to get three mixers to track in a perfectly linear fashion - any non-linearities showed up as shifts in the luminance level. So when we built the first commercial analog component switcher - the Model 100 CV - we change the components to Y, R-Y, B-Y. This allowed for excellent linearity in the luminance component - any tracking errors were shifted to the color difference components and were far less noticeable.

I might add that there were a number of improved NTSC encoders and NTSC enhancements developed in the same time frame as Betacam. Central Dynamics developed an excellent NTSC encoder that carefully limited the energy from the subcarrier in three dimensional space so that it did not overlap with the energy from the luminance signal. This allowed decoders to produce improved color components largely free from the luma/cross color issues in most NTSC decoders.

And Yve Farudja came up with several excellent NTSC enhancements with significantly improved performance, but the world was more interested in moving to HD.

[Betacam]"provides a crisp, true broadcast quality product with 300 lines of horizontal luma resolution, and 120 lines chroma resolution (versus ~30 for Betamax/VHS)." http://en.wikipedia.org/wiki/Betacam

Luma - 300 lines, which is better than the 240 line (typical) from NTSC

Chroma- 120 lines - this is the same as NTSC, but is misleading, as only certain colors actually have this resolution. Any color based on the ! component has far less resolution

Betamax/VHS - really poor color performance due to heterodyne color recording - FAR LESS color information and far more color distortion than a clean NTSC baseband signal.


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.

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: