[opendtv] Re: Finally anamorphically compressed 480i

  • From: Craig Birkmaier <craig@xxxxxxxxx>
  • To: opendtv@xxxxxxxxxxxxx
  • Date: Wed, 2 Feb 2011 08:44:14 -0500

At 4:01 PM -0600 2/1/11, Manfredi, Albert E wrote:
John Shutt wrote:

 Very cool, Ron.  Thanks!

+1. It was cool indeed.

So, that whole discussion started because we were talking about downconverting 1080/60i to 480/60i, without having to deinterlace. Of all the conversions, the 1080i to 480i seemed to be the messiest, *if* you need to deinterlace and then reinterlace. If those ugly steps can be avoided, maybe it's not such a big deal (and you should be able to extract a wide screen copy of the HD original).

It is a question of quality. You can do the conversion cheap and dirty or do it well.

Keep in mind that this has NOTHING to do with MPEG or the encoding. You are going from one MPEG Profile to another (MP@HL to MP@ML). The downconversion takes place before re-encoding.


But the underlying premise is that the 480i is needed to provide the lowest possible bit rate, for use in the SD subchannels.

More like the standard broadcast industry practice. Services that want to deliver good quality at lower bitrates over the Internet have, for the most part, moved to encoding progressive source at 360P (or less).

For broadcasters this is mostly about the tools they have in their facilities, which for many broadcasters are still optimized for interlaced SD. MANY stations have not upgraded their plants to support ANY kind of HD other than for pass through of HD network signals. As John pointed out, this is simple pragmatism and economics.


Now that TV cameras can output progressive, and that all US TV receivers out there are perfectly happy receiving progressive (e.g. 480/24p, 480/30p, and 480/60p), why don't the anti-interlace folk out there put their money where their mouths have been?

The anti interlace folks are not the problem. They have moved on. It is the folks who are heavily invested in an infrastructure that was built around interlaced SD that are the problem. That and the difficulty that many independent and small market network affiliates are having just surviving. They simply are unwilling to make the investment needed to move to an all progressive work flow, despite the fact that it is now cheaper to build a 720P infrastructure than it was to build a interlaced SD (601) infrastructure (and keep in mind that many stations never upgraded to DIGITAL SD, and are still using NTSC analog acquisition and processing gear.


Is it not true that progressive compresses more efficiently than interlace? Why are the SD subchannels universally using 480i? I haven't ever seen any that use 480p. How come? Even without a 360p option, can't prefiltering of the image result in a bit rate constrained SD stream, say 480/24p?

The major limitation is ATSC Table 3. As you note, there is no 360P option. And since the ATSC does not support any line length other than 704 for interlaced SD (640 for progressive 4:3 480P), a broadcaster cannot save bits with a lower number of samples per line. So the most common bandwidth conserving technique is to pre-filter the 704 x 480i source to remove detail. You will often see that these sub-channels are very soft because of this pre-filtering.

If the ATSC had simply adopted the entire MP@ML and MP@HL profiles, instead of imposing format restrictions, broadcasters would have many additional tools to work with to improve delivered quality at lower bit rates.

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: