On 20 Feb 2000 14:46:35 -0500, Nieminen Juha
<war### [at] sarakerttunencstutfi> wrote:

>(I don't know if floating point numbers has to used for this process, but
>specially if they have, then inaccuracies are most likely to occur).

I have programs that allow selecting floating point or integer
routines for jpeg decompression. The floating point routines are
considered slower, but more accurate by these programs.

for what it's worth...

Glen Berry

Post a reply to this message