diff options
| author | Michael Niedermayer <michaelni@gmx.at> | 2008-04-17 03:00:08 +0000 |
|---|---|---|
| committer | Michael Niedermayer <michaelni@gmx.at> | 2008-04-17 03:00:08 +0000 |
| commit | c49c5e23dc3fb273b46440656f74d416bcc9ab41 (patch) | |
| tree | 5c8c032c1172879b966f520f6d85821e1b255d76 /libavcodec/alac.c | |
| parent | 59b377ac86589e3c11a613e9a520a7ca76935ea4 (diff) | |
| download | ffmpeg-c49c5e23dc3fb273b46440656f74d416bcc9ab41.tar.gz | |
Change k limiting code, i think the code was buggy.
If you have ALAC files TEST them! Mine produce the same md5 but the new
code is not identical if limiting does happen.
Originally committed as revision 12872 to svn://svn.ffmpeg.org/ffmpeg/trunk
Diffstat (limited to 'libavcodec/alac.c')
| -rw-r--r-- | libavcodec/alac.c | 12 |
1 files changed, 7 insertions, 5 deletions
diff --git a/libavcodec/alac.c b/libavcodec/alac.c index b143d3a9d4..e99b12b0e0 100644 --- a/libavcodec/alac.c +++ b/libavcodec/alac.c @@ -226,18 +226,20 @@ static void bastardized_rice_decompress(ALACContext *alac, k = count_leading_zeros(history) + ((history + 16) >> 6 /* / 64 */) - 24; - extrabits = show_bits(&alac->gb, k); + if (k >= rice_kmodifier) + k = rice_kmodifier; - block_size = (((1 << k) - 1) & rice_kmodifier_mask) * x - + extrabits - 1; + x = (x << k) - x; + + extrabits = show_bits(&alac->gb, k); if (extrabits < 2) { - x = 1 - extrabits; - block_size += x; skip_bits(&alac->gb, k - 1); } else { + x += extrabits - 1; skip_bits(&alac->gb, k); } + block_size = x; } if (block_size > 0) { |