[vlc-commits] aout: always use S16N (rather than FI32) in absence of FPU

[Rémi Denis-Courmont]

vlc | branch: master | Rémi Denis-Courmont <remi at remlab.net> | Wed Jul  4 22:43:39 2012 +0300| [a68389199363c5d16a384960b677f4d2861359e7] | committer: Rémi Denis-Courmont

aout: always use S16N (rather than FI32) in absence of FPU

None of the real audio outputs support FI32; they already typically
converted down to S16 internally. Moreover, mixing FI32 requires
64-bits (signed) multiplication, which non-FPU processors typically
do not support natively.

> http://git.videolan.org/gitweb.cgi/vlc.git/?a=commit;h=a68389199363c5d16a384960b677f4d2861359e7
---

 src/audio_output/output.c |    5 -----
 1 file changed, 5 deletions(-)

diff --git a/src/audio_output/output.c b/src/audio_output/output.c
index 9bb276c..98d829c 100644
--- a/src/audio_output/output.c
+++ b/src/audio_output/output.c
@@ -234,11 +234,6 @@ int aout_OutputNew( audio_output_t *p_aout,
     if( HAVE_FPU )
         owner->mixer_format.i_format = VLC_CODEC_FL32;
     else
-    /* Otherwise, audio filters will not work. Use fixed-point if the input has
-     * more than 16-bits depth. */
-    if( p_format->i_bitspersample > 16 || !AOUT_FMT_LINEAR(p_format))
-        owner->mixer_format.i_format = VLC_CODEC_FI32;
-    else
     /* Fallback to 16-bits. This avoids pointless conversion to and from
      * 32-bits samples for the sole purpose of software mixing. */
         owner->mixer_format.i_format = VLC_CODEC_S16N;