Skip to content

Tags: iskunk/vgmstream

Tags

r1050

Toggle r1050's commit message
Updated for Visual Studio 2015

r1040

Toggle r1040's commit message
Merge pull request vgmstream#11 from soneek/master

Update bfstm for any header size

r1039

Toggle r1039's commit message
CSMP for Metroid Prime 3 and DKCR, from Antidote

r1038

Toggle r1038's commit message
IDSP as in SSB4 3DS

r1037

Toggle r1037's commit message
fix missing dependency when built with i586-mingw32msvc-gcc 4.8.2

r1036

Toggle r1036's commit message
add bfwav to MSVC 2008 project file

r1035

Toggle r1035's commit message
Eliminate C99isms so it compiles in MSVC, and comment out some unused…

… variables to silence warnings

r1034

Toggle r1034's commit message
clean up some warning-causing ugliness

r1033

Toggle r1033's commit message
Extend GENH to support little-endian "Gamecube" DSP ADPCM coefficient…

…s, for some 3DS titles.

This involves a reinterpretation of byte 0x30 (coef type). Formerly this byte took on only two values, to indicate how the ADPCM coefficients (aka codebook) were stored:
0 - normal coefs: all 16 coefs interleaved into one array, offset given at 0x24 for left, 0x28 for right
1 - split coefs: 8 coefs in the main array, additional offset to 2nd array given at 0x34 for left, 0x38 for right

Now I am considering this to be indicated only by bit 0 of the coef type. Bit 1 is taking on an additional interpretation, if it is set, we consider the coefficients to be little endian rather than the normal big endian.

This should maintain backwards compatibility with old GENH files, which should have only used the value 0 or 1.

Thus, in effect we have:
0: normal, big endian
1: split, big endian
2: normal, little endian
3: split, little endian

I don't know of any situation in which 3 would be used (yet), but I'm sure devs will continue to surprise me.

r1032

Toggle r1032's commit message
Merge pull request vgmstream#3 from soneek/master

hcs's update for IMA decoder