[opendtv] Re: Get rid of Interlaced Media?

  • From: "Allen Le Roy Limberg" <allimberg@xxxxxxxxxxxx>
  • To: <opendtv@xxxxxxxxxxxxx>
  • Date: Tue, 23 Jan 2007 22:27:43 -0500

Columbia University owns the patent on handling interlace correctly for MPEG-2 
type video compression.  Everybody avoids using it to avoid the royalty.

Al Limberg
  ----- Original Message ----- 
  From: dan.grimes@xxxxxxxx 
  To: opendtv@xxxxxxxxxxxxx 
  Sent: Tuesday, January 23, 2007 7:54 PM
  Subject: [opendtv] Re: Get rid of Interlaced Media?


  Mark, I agree that interlaced 1080i29.97 media doesn't look too bad in 
certain uses. No doubt the work at the Met. Opera House looks good.

  I've been working with 1080i29.97 for image magnification (remote playback at 
satellite campuses) using Doremi V1-HD recorders, which use JPEG2000, at rates 
of 300Mb/s and Christie projectors. The picture is very decent at reasonable 
distances. The weak link in my case is the poor SNR and native 720 pixels 
(spatially offset) in the Panasonic AK-HC1500G camera.

  With broadcast TV, the interlace appears to work fine for dramas where the 
images are close up. But with fine detail and fast motion such as sports, it 
certainly doesn't hold up at 19Mb/s. Some movies have detail problems as well.

  The project I am currently doing is encoded at 6-8 Mb/s and requires the 
video to be displayed on a computer. Using players such as WMP and QT, the 
interlaced MPEG-2 video looks terrible on the computer monitor. There are a lot 
of steps one must take to make interlaced video look good on a typical computer 
display and not one software perform each step necessary. There are a lot of 
hurdles to jump to get there.

  For the big media networks that distribute over cable, sat., OTA, etc. with 
media destined to dedicated displays that incorporate hardware to handle 
interlaced media, I suppose interlace works alright. But for internet 
distribution and displaying media on typical computers, the interlaced media 
can be more problematic.

  I guess my point is if interlacing causes so many problems in the media 
production world that could be simply overcome by staying with progressive 
media, doesn't it make sense to work towards a production and distribution 
model that avoids it?

  Dan Grimes

Other related posts: