[opendtv] Re: Math of oversampling - a REALITY CHECK

  • From: Jeroen Stessen <jeroen.stessen@xxxxxxxxxxx>
  • To: opendtv@xxxxxxxxxxxxx
  • Date: Mon, 2 May 2005 16:46:39 +0200

Hi, 

Tom to Craig: 
> You sort of touched on it but let's also not forget the optimal 
> difference between the delivery and display resolutions.
> Most effectively you can hide the pixel/raster structure by having 
> a higher resolution display even though both the horizontal and 
> vertical display rez are maybe each 50% higher than the actual 
> images being broadcast. 

As proposed by John Watkinson. Very sensible indeed, because of 
the slow roll-off of the anti-aliasing filter of the camera and 
the anti-imaging filter of the display. You can get a better fill 
factor of the transmission channel if you apply steeper electronic 
filters that will dominate over the weak optical filters. 

> The results end up being similar to the various upscaled images of 
> Jeroen's.  Extra softness becomes noticeable in extreme conditions 

I bet that most display manufacturers can't resist the challenge 
of applying sharpness enhancement algorithms to the up-sampled 
images. While that typically won't increase the detail, it will 
at least repair some of the sharpness impression. We do it now 
for SD input signals, so why not do it for down-sampled HD ? 
(That is not to say that real HD can be replaced with up-converted 
SD, the difference will still be visible...) 

> but not, say, when viewing a 720p image on a 1080p display at a 
> viewing distance designed for the display instead of the image. 

But such a small lack of sharpness is easily repaired with a bit 
of transient improvement. Or a lot, depending on your taste. 
This trend is unstoppable anyway, that "we" will try to improve 
everything. I was almost tempted to apply a bit of LTI to the 
up-sampled images, but have decided not to do it. Not yet. 

By the way, did anybody notice how much ringing was added by my 
down-sampling filter ? Especially on the dark side of grey-white 
edges you can see it very well. What did you think of it ? 

> All this just adds a bit more wiggle room.

Read Watkinson: http://www.microsoft.com/whdc/archive/convrg1.mspx
It's more than mere wiggling, it actually makes a lot of sense. 

Greetings, 
-- Jeroen 

+-------------------------------+------------------------------------------+
| From:     Jeroen H. Stessen   | E-mail:  Jeroen.Stessen@xxxxxxxxxxx |
| Building: SFJ-5.22 Eindhoven  | Deptmt.: Philips Applied Technologies |
| Phone:    ++31.40.2732739     | Visiting & mail address: Glaslaan 2 |
| Mobile:   ++31.6.44680021     | NL 5616 LW Eindhoven, the Netherlands |
| Pager:    ++31.6.65133818     | Website: http://www.apptech.philips.com/ 
|
+-------------------------------+------------------------------------------+


 
 
----------------------------------------------------------------------
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: