On Monday, 17 September 2001 at 17:37, Mark Taylor wrote:
>
> The code related to this reduces to:
>
> offset = x-(floor(x)+.5)
> assert(fabs(offset)<.50001))
>
> so it is either just a precision problem (and then I wonder
> why it would be showing up now?) or a rather obscure bug.
> I changed it to .501.
I wonder if the problem is occuring on a Redhat 7 system? I believe the
gcc distributed with that redhat has some serious floating point
problems, which may also affect the quality of your mp3s. Try kgcc on
that system and see if you still have that problem...
--
MP3 ENCODER mailing list archive is at:
http://www.mail-archive.com/mp3encoder%40minnie.cs.adfa.edu.au/
