[MPlayer-users] Comparison of different software scaler types

[Matthias Wieser]

Am Mittwoch, 15. Oktober 2003 12:16 schrieb Tuukka Toivonen:
> >I think, there is one.
> >A good downscaler should try to
> > -preserve as much as possible details
> >- not introduce artifacts or noise (as -sws 8 does)
> >After scaling up, the difference to the original imige should be
> > minimal.
>
> Since you don't specify upscaler,
                 I have thought of a "stupid" upscaler.
> you're actually talking about
> compression, not scaling. "make image smaller so that when restored
> to the original size there is as little differences as possible"
> is nothing else except compression.

That's right, call it compression if you want.

That's even the actual reason why people scale down:
To shift "compression work" from the videocodec to the downscaler.
You do it, because sometimes (when the framesize is quite large) 
downscaling produces less loss of quality than higher compression of the 
video codec.
So you could call downscaling "precompression". It's important to find a 
good ratio between "precompression" and video codec compression.

> The point is that the image should look as much to the original
> when scaled down as possible. But since human preferences and
> display devices vary, there isn't a single best algorithm or
> even a meter which could be used to judge "goodness" (except
> human viewers... but even then it's display device dependent).

That's exactly the reason why I'm showing the different scaled 
(=compressed) images opposite to the original image. I'm not going to 
give you any mathematical differences.
If you would write a downscaler that would take human eye into 
consideration, it would probably pass my comparison without problems.

> And the scaling ratio matters
> much, nearest neighbour scaler isn't that bad when scaling exactly 2x
> larger. 1.5x and it's much worse.

I used nearest neighbour scaler to scale the resized images exactly 2x 
back to original size.

Regards,
   Matthias