[FFmpeg-devel] a64 encoder
Fri Jan 16 11:15:41 CET 2009
Vitor Sessak wrote:
> Here you could use just int [instead of uint32_t] (it will use just
> as much mem, since it will be mapped to registers anyway). Also, C99
> warrants that int has at least 32 bits.
This is not quite correct.
INT_MAX >= 32767
INT_MIN <= -32767
i.e. int must be at least ln(65535) / ln(2) = 15.999977986-bits wide.
POSIX however requires
INT_MAX >= 2147483647
INT_MIN <= -2147483647
i.e. almost 32 bits but most two's-complement systems will define
INT_MIN to -2147483648
More information about the ffmpeg-devel