> >>>>> "M" == Mark Taylor <[EMAIL PROTECTED]> writes:
>
> M> A more serious problem is a few lines down: The MS switching
> M> criterion has the same mistake.
>
> Uum, I think this is OK. no mistake is found.
> The code is
>
> x1 = minimum(thm_save[0][sb],thm_save[1][sb]);
> x2 = maximum(thm_save[0][sb],thm_save[1][sb]);
> /* thresholds difference in db */
> if (x2 >= 1000*x1) db=30;
> else db = 10*log10(x2/x1);
>
> It will be always x2 > x1,
> and "thresholds difference in db" is always positive value,
> which means the absolute value of difference.
> ---
> Takehiro TOMINAGA // may the source be with you!
> --
Takehiro is of course right again :-). That's a relief, because
it means the only bug was that the side channel maskings
(and more unlikely, the mid channel) were too large.
Mark
--
MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
