[opendtv] Re: Learning From the Veterans - local news in HD

  • From: Craig Birkmaier <craig@xxxxxxxxx>
  • To: opendtv@xxxxxxxxxxxxx
  • Date: Thu, 29 Apr 2010 07:03:04 -0400

At 9:00 AM -0400 4/28/10, Leonard Caillouet wrote:
How does orthogonal sampling require that the pixels be square? Similarly, having square pixels does not necessarily imply that the sampling is orthogonal, does it? Orthogonal sampling is a statistical notion that requires equal sampling in both directions, but does not require that each sample have a particular characteristic, does it? I don't see the connection, nor the relevance of throwing this term into the discussion. Can you clarify or show me what I am missing?

You make a good point!

We used this term in the '90s during the DTV wars because it better expressed the most important point, which you stated above:

Orthogonal sampling is a statistical notion that requires equal sampling in both directions.

We used the term in part because "square pixels" became a fighting term between the computer industry and the video industry.

Assuming the goal is orthogonal sampling, while the actual pixel geometry of a sampling device may vary, by the time we resample an image and encode it for emission, we are left with sample data that has equivalent H & V content. And in virtually all cases today, these samples are displayed on devices where each sample is essentially square.

The important take-away here is that video information should be delivered using the same pixel geometry as is used for all other types of visual information in an infrastructure (i.e. the Internet) that assumes square pixels for display.

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: