har0ke/FFmpeg
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

ac3enc: NEON optimised ac3_max_msb_abs_int16 and ac3_exponent_min

Browse Source
This commit is contained in:
Mans Rullgard 2011-03-09 19:51:50 +00:00
parent f711424969
commit f4855a904e
5 changed files with 100 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
libavcodec/ac3dsp.c
View File

@ -109,6 +109,8 @@ av_cold void ff_ac3dsp_init(AC3DSPContext *c, int bit_exact)
c->ac3_rshift_int32 = ac3_rshift_int32_c;
c->float_to_fixed24 = float_to_fixed24_c;
if (ARCH_ARM)
ff_ac3dsp_init_arm(c, bit_exact);
if (HAVE_MMX)
ff_ac3dsp_init_x86(c, bit_exact);
}

1
libavcodec/ac3dsp.h
View File

@ -84,6 +84,7 @@ typedef struct AC3DSPContext {
} AC3DSPContext;
void ff_ac3dsp_init (AC3DSPContext *c, int bit_exact);
void ff_ac3dsp_init_arm(AC3DSPContext *c, int bit_exact);
void ff_ac3dsp_init_x86(AC3DSPContext *c, int bit_exact);
#endif /* AVCODEC_AC3DSP_H */

3
libavcodec/arm/Makefile
View File

@ -1,3 +1,4 @@
OBJS-$(CONFIG_AC3DSP) += arm/ac3dsp_init_arm.o
OBJS-$(CONFIG_DCA_DECODER) += arm/dcadsp_init_arm.o \
OBJS-$(CONFIG_VP5_DECODER) += arm/vp56dsp_init_arm.o
@ -46,6 +47,8 @@ NEON-OBJS-$(CONFIG_H264DSP) += arm/h264dsp_neon.o \
NEON-OBJS-$(CONFIG_H264PRED) += arm/h264pred_neon.o \
NEON-OBJS-$(CONFIG_AC3DSP) += arm/ac3dsp_neon.o
NEON-OBJS-$(CONFIG_DCA_DECODER) += arm/dcadsp_neon.o \
arm/synth_filter_neon.o \

35
libavcodec/arm/ac3dsp_init_arm.c Normal file
View File

@ -0,0 +1,35 @@
/*
* Copyright (c) 2011 Mans Rullgard <mans@mansr.com>
*
* This file is part of Libav.
*
* Libav is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* Libav is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with Libav; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include <stdint.h>
#include "libavutil/attributes.h"
#include "libavcodec/ac3dsp.h"
#include "config.h"
void ff_ac3_exponent_min_neon(uint8_t *exp, int num_reuse_blocks, int nb_coefs);
int ff_ac3_max_msb_abs_int16_neon(const int16_t *src, int len);
av_cold void ff_ac3dsp_init_arm(AC3DSPContext *c, int bit_exact)
{
if (HAVE_NEON) {
c->ac3_exponent_min = ff_ac3_exponent_min_neon;
c->ac3_max_msb_abs_int16 = ff_ac3_max_msb_abs_int16_neon;
}
}

59
libavcodec/arm/ac3dsp_neon.S Normal file
View File

@ -0,0 +1,59 @@
/*
* Copyright (c) 2011 Mans Rullgard <mans@mansr.com>
*
* This file is part of Libav.
*
* Libav is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* Libav is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with Libav; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "asm.S"
function ff_ac3_max_msb_abs_int16_neon, export=1
vmov.i16 q0, #0
vmov.i16 q2, #0
1: vld1.16 {q1}, [r0,:128]!
vabs.s16 q1, q1
vld1.16 {q3}, [r0,:128]!
vabs.s16 q3, q3
vorr q0, q0, q1
vorr q2, q2, q3
subs r1, r1, #16
bgt 1b
vorr q0, q0, q2
vorr d0, d0, d1
vpmax.u16 d0, d0, d0
vpmax.u16 d0, d0, d0
vmov.u16 r0, d0[0]
bx lr
endfunc
function ff_ac3_exponent_min_neon, export=1
cmp r1, #0
bxeq lr
push {lr}
mov r12, #256
1:
vld1.8 {q0}, [r0,:128]
mov lr, r1
add r3, r0, #256
2: vld1.8 {q1}, [r3,:128], r12
subs lr, lr, #1
vmin.u8 q0, q0, q1
bgt 2b
subs r2, r2, #16
vst1.8 {q0}, [r0,:128]!
bgt 1b
pop {pc}
endfunc