Hi,
I've compiled lame with the free Borland C++ compiler (for Windows).
Some code changes were necessary. In ts_process_time,
LARGE_INTEGER Kernel = { KernelTime.dwLowDateTime,
KernelTime.dwHighDateTime };
LARGE_INTEGER User = { UserTime.dwLowDateTime,
UserTime.dwHighDateTime };
needs to be changed to:
LARGE_INTEGER Kernel;
LARGE_INTEGER User;
Kernel.LowPart = KernelTime.dwLowDateTime;
Kernel.HighPart = KernelTime.dwHighDateTime;
User.LowPart = UserTime.dwLowDateTime;
User.HighPart = UserTime.dwHighDateTime;
To be able to link nasm object modules, I had to change
section_code
to
section code class=code use32
and corresponding for "section_data", since the Borland linker expects
32-bit "-f obj" format. If "code class=code" is missing, a module
definition file is required, and I didn't want to deal with that (I had
enough problems getting to terms with the compiler as it was :).
When I tried using TAKEHIRO_IEEE754_HACK, I had to create some new
definitions in machine.h, since Borland doesn't support "long long".
Instead, it has an "__int64" type. But it seemed like this hack requires
the inline assembly in vbrquantize.c; _adj43 wasn't found during link.
Why not move that assembler to a .nas file, so that more compilers can
gain from it?
Other than that, nothing special was required really. I also compiled
and used mpglib and libsndfile. Some quick tests (with 3.83) seemed to
work - I didn't examine the output closely - though it was very slow (no
assembler used at all then).
Btw, the FFT code in i386/, is that safe to use? If so, should I simply
change the call in fft.c?
Before sending in the makefile, I'd like to play around with it some
more and try to make the executable faster (I compared it with a
lame.exe I downloaded somewhere (same version), and the other version
was about twice as fast - and that was much slower than the 3.80 dll I
use with CDEx in VBR mode. I tried to make the settings equal in all
cases, but...). However, I won't have the possibility to do that for a
little more than a month...
--
Magnus Holmgren
--
MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
