>>>>> "S" == Sigbjrn Skjret <[EMAIL PROTECTED]> writes:
>> This feature is not default because it is deeply depending on
>> IEEE754 floating point calculation and there're some computers
>> NOT compatible to IEEE754. (but it is so rare :p)
S> When trying this hack the results got extremely noisy and of
S> very poor quality. This was on 68k, I've been unable to test on
S> PPC (read below).
Uum.....
S> The compiler also gave me several "type punning" warnings, this
S> is probably intentional, however it also gave me some warnings
S> about integer operations overflowing, this perhaps is not so
S> good...
"integer operations overflowing" !?
Uum....
I checked this hack on the SunOS ultrasparcII + SUN C(Workshop 4.2).
I got a message like this...
"quantize-pvt.c", line 1177: warning: integer overflow detected: op "<<"
"quantize-pvt.c", line 1178: warning: integer overflow detected: op "<<"
"quantize-pvt.c", line 1179: warning: integer overflow detected: op "<<"
"quantize-pvt.c", line 1180: warning: integer overflow detected: op "<<"
and the line 1177-1180 are
((float *)pi)[0] = (x0 + adj43asm[pi[0] - MAGIC_INT]);
((float *)pi)[1] = (x1 + adj43asm[pi[1] - MAGIC_INT]);
((float *)pi)[2] = (x2 + adj43asm[pi[2] - MAGIC_INT]);
((float *)pi)[3] = (x3 + adj43asm[pi[3] - MAGIC_INT]);
uum, there's no "<<".... ????
The binaries runs fine, and "make test" result is OK, too.
but the speed is slower than the no hacked version.... *sigh*
---
Takehiro TOMINAGA // may the source be with you!
--
MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
