[MPlayer-dev-eng] Patch for gcc 3.4 against MPlayer-1.0pre5

[D Richard Felker III]

On Thu, Sep 02, 2004 at 08:12:10AM +0200, Torinthiel wrote:
> On Thu, Sep 02, 2004 at 01:20:23AM +0100, Martin Simmons wrote:
> > Based on what spec?  It is normal for floating point rounding of an exact half
> > to generate an even integer, i.e.
> 
> I don't know spec's, but ALWAYS if two directions are possible, then
> half is rounded up. So that round(x)=trunc(x+0.5)
> I don't know about negative numbers though.

says who? your high school math teacher? always rounding one way will
introduce a bias in the output which can often be visible/audible.
rounding to nearest even integer in the exact-half case tries to
alleviate this problem.

rich