[mythtv] ffmpeg SWSCALE!

Daniel Kristjansson danielk at cuymedia.net
Thu Aug 31 02:13:53 UTC 2006

On Wed, 2006-08-30 at 21:54 -0400, Yeasah Pell wrote:
> Michael T. Dean wrote:
> But a monitor doesn't really have any such low pass filtering -- it's an 
> array of square (or rectangular) pixels, with hard edges.
You mean an LCD, a CRT monitor has many RGB components for each pixel,
and the electron beam is in fact filtered.

> As a result, 
> the high-frequency aliasing is (and must be) fully present in the 
> displayed image. Depending on how high the resolution is, you can see 
> this aliasing easily -- but even though it is more noticeable in lower 
> resolution images, there's still just as much aliasing in higher 
> resolution images, it's just harder to see (use a loupe if you need :-)

You are confusing aliasing with other properties of the display.
Aliasing is seen the _low_ frequency you see when the number of
samples is insufficient for the high frequencies in the signal.
If there is no aliasing present in the transmitted signal, you
will never get aliasing when you display the image so long as
you do not re-sample the image. Of course if you do re-sample
the image by theory you need to double the display resolution
wrt to the source image to fully resolve the image, and in
practice you need to more than double the image. If you don't
have such a high resolution display you need to bandlimit
the image before scaling, i.e. you need to blur it.

This is one of the reasons why a 1920x1080i CRT display looks so
much better than a 3840x2160p LCD or DLP display (In practice
the DLP's are 4096x4096 and has a lot of other artifacts that
contribute to the poor image quality, and the only high-res
LCD I know of has a 4x3 aspect ratio).

BTW This is also why people get worse image quality when they
configure their nVidia TV-out resolutions to something like
800x600, rather than matching the resolution exactly to the
TV and to the capture resolution.

-- Daniel

More information about the mythtv-dev mailing list