[MEncoder-users] High quality encoding of anime and broken mencoder options

[Jorge Peixoto de Morais Neto]

On Sat, 8 Mar 2008 01:49:25 +0100
Nicolas George <nicolas.george at normalesup.org> wrote:

> L'octidi 18 ventôse, an CCXVI, Jorge Peixoto de Morais Neto a écrit :
> > Yes I am. I know that PSNR numbers are not a perfect way of
> > measuring visual quality, but I don't know of any other metric.
> 
> It is possible to compute the linear correlation coefficient between the
> original and the encoded result, plus factors to ensure that the correlation
> function is actually the identity. I took the idea from <URL:
> http://www.cns.nyu.edu/~zwang/files/papers/quality_2c.pdf
>  >, where the experiments seemed to indicate this was a better estimation of
> a subjective quality than the PSNR; but of course, they will not say that
> their work is lousy.
What is the easiest way I can calculate such a correlation?


-- 
Software is like sex: it is better when it is free. --Linus Torvalds