ARM: NEON optimised vector_clipf

Originally committed as revision 20031 to svn://svn.ffmpeg.org/ffmpeg/trunk

1 files changed, 4 insertions, 0 deletions

diff --git a/libavcodec/arm/dsputil_neon.c b/libavcodec/arm/dsputil_neon.c
index 024bb9727c..efe07e1172 100644
--- a/libavcodec/arm/dsputil_neon.c
+++ b/libavcodec/arm/dsputil_neon.c
@@ -174,6 +174,8 @@ void ff_int32_to_float_fmul_scalar_neon(float *dst, const int *src,
 void ff_vector_fmul_reverse_neon(float *dst, const float *src0,
                                  const float *src1, int len);
 
+void ff_vector_clipf_neon(float *dst, const float *src, float min, float max,
+                          int len);
 void ff_float_to_int16_neon(int16_t *, const float *, long);
 void ff_float_to_int16_interleave_neon(int16_t *, const float **, long, int);
 
@@ -297,6 +299,8 @@ void ff_dsputil_init_neon(DSPContext *c, AVCodecContext *avctx)
     c->sv_fmul_scalar[0] = ff_sv_fmul_scalar_2_neon;
     c->sv_fmul_scalar[1] = ff_sv_fmul_scalar_4_neon;
 
+    c->vector_clipf = ff_vector_clipf_neon;
+
     if (!(avctx->flags & CODEC_FLAG_BITEXACT)) {
         c->float_to_int16 = ff_float_to_int16_neon;
         c->float_to_int16_interleave = ff_float_to_int16_interleave_neon;