1;*****************************************************************************
2;* x86-optimized functions for volume filter
3;* Copyright (c) 2012 Justin Ruggles <justin.ruggles@gmail.com>
4;*
5;* This file is part of FFmpeg.
6;*
7;* FFmpeg is free software; you can redistribute it and/or
8;* modify it under the terms of the GNU Lesser General Public
9;* License as published by the Free Software Foundation; either
10;* version 2.1 of the License, or (at your option) any later version.
11;*
12;* FFmpeg is distributed in the hope that it will be useful,
13;* but WITHOUT ANY WARRANTY; without even the implied warranty of
14;* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
15;* Lesser General Public License for more details.
16;*
17;* You should have received a copy of the GNU Lesser General Public
18;* License along with FFmpeg; if not, write to the Free Software
19;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
20;******************************************************************************
21
22%include "libavutil/x86/x86util.asm"
23
24SECTION_RODATA 32
25
26pd_1_256:     times 4 dq 0x3F70000000000000
27pd_int32_max: times 4 dq 0x41DFFFFFFFC00000
28pw_1:         times 8 dw 1
29pw_128:       times 8 dw 128
30pq_128:       times 2 dq 128
31
32SECTION_TEXT
33
34;------------------------------------------------------------------------------
35; void ff_scale_samples_s16(uint8_t *dst, const uint8_t *src, int len,
36;                           int volume)
37;------------------------------------------------------------------------------
38
39INIT_XMM sse2
40cglobal scale_samples_s16, 4,4,4, dst, src, len, volume
41    movd        m0, volumem
42    pshuflw     m0, m0, 0
43    punpcklwd   m0, [pw_1]
44    mova        m1, [pw_128]
45    lea       lenq, [lend*2-mmsize]
46.loop:
47    ; dst[i] = av_clip_int16((src[i] * volume + 128) >> 8);
48    mova        m2, [srcq+lenq]
49    punpcklwd   m3, m2, m1
50    punpckhwd   m2, m1
51    pmaddwd     m3, m0
52    pmaddwd     m2, m0
53    psrad       m3, 8
54    psrad       m2, 8
55    packssdw    m3, m2
56    mova  [dstq+lenq], m3
57    sub       lenq, mmsize
58    jge .loop
59    REP_RET
60
61;------------------------------------------------------------------------------
62; void ff_scale_samples_s32(uint8_t *dst, const uint8_t *src, int len,
63;                           int volume)
64;------------------------------------------------------------------------------
65
66%macro SCALE_SAMPLES_S32 0
67cglobal scale_samples_s32, 4,4,4, dst, src, len, volume
68%if ARCH_X86_32 && cpuflag(avx)
69    vbroadcastss   xmm2, volumem
70%else
71    movd           xmm2, volumed
72    pshufd         xmm2, xmm2, 0
73%endif
74    CVTDQ2PD         m2, xmm2
75    mulpd            m2, m2, [pd_1_256]
76    mova             m3, [pd_int32_max]
77    lea            lenq, [lend*4-mmsize]
78.loop:
79    CVTDQ2PD         m0, [srcq+lenq         ]
80    CVTDQ2PD         m1, [srcq+lenq+mmsize/2]
81    mulpd            m0, m0, m2
82    mulpd            m1, m1, m2
83    minpd            m0, m0, m3
84    minpd            m1, m1, m3
85    cvtpd2dq       xmm0, m0
86    cvtpd2dq       xmm1, m1
87%if cpuflag(avx)
88    vmovdqa [dstq+lenq         ], xmm0
89    vmovdqa [dstq+lenq+mmsize/2], xmm1
90%else
91    movq    [dstq+lenq         ], xmm0
92    movq    [dstq+lenq+mmsize/2], xmm1
93%endif
94    sub            lenq, mmsize
95    jge .loop
96    REP_RET
97%endmacro
98
99INIT_XMM sse2
100%define CVTDQ2PD cvtdq2pd
101SCALE_SAMPLES_S32
102%if HAVE_AVX_EXTERNAL
103%define CVTDQ2PD vcvtdq2pd
104INIT_YMM avx
105SCALE_SAMPLES_S32
106%endif
107%undef CVTDQ2PD
108
109; NOTE: This is not bit-identical with the C version because it clips to
110;       [-INT_MAX, INT_MAX] instead of [INT_MIN, INT_MAX]
111
112INIT_XMM ssse3, atom
113cglobal scale_samples_s32, 4,4,8, dst, src, len, volume
114    movd        m4, volumem
115    pshufd      m4, m4, 0
116    mova        m5, [pq_128]
117    pxor        m6, m6
118    lea       lenq, [lend*4-mmsize]
119.loop:
120    ; src[i] = av_clipl_int32((src[i] * volume + 128) >> 8);
121    mova        m7, [srcq+lenq]
122    pabsd       m3, m7
123    pshufd      m0, m3, q0100
124    pshufd      m1, m3, q0302
125    pmuludq     m0, m4
126    pmuludq     m1, m4
127    paddq       m0, m5
128    paddq       m1, m5
129    psrlq       m0, 7
130    psrlq       m1, 7
131    shufps      m2, m0, m1, q3131
132    shufps      m0, m0, m1, q2020
133    pcmpgtd     m2, m6
134    por         m0, m2
135    psrld       m0, 1
136    psignd      m0, m7
137    mova  [dstq+lenq], m0
138    sub       lenq, mmsize
139    jge .loop
140    REP_RET
141