1;******************************************************************************
2;* Opus SIMD functions
3;*
4;* This file is part of FFmpeg.
5;*
6;* FFmpeg is free software; you can redistribute it and/or
7;* modify it under the terms of the GNU Lesser General Public
8;* License as published by the Free Software Foundation; either
9;* version 2.1 of the License, or (at your option) any later version.
10;*
11;* FFmpeg is distributed in the hope that it will be useful,
12;* but WITHOUT ANY WARRANTY; without even the implied warranty of
13;* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
14;* Lesser General Public License for more details.
15;*
16;* You should have received a copy of the GNU Lesser General Public
17;* License along with FFmpeg; if not, write to the Free Software
18;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19;******************************************************************************
20
21%include "libavutil/x86/x86util.asm"
22
23SECTION_RODATA
24
25         ; 0.85..^1    0.85..^2    0.85..^3    0.85..^4
26tab_st: dd 0x3f599a00, 0x3f38f671, 0x3f1d382a, 0x3f05a32f
27
28SECTION .text
29
30INIT_XMM fma3
31%if UNIX64
32cglobal opus_deemphasis, 3, 3, 8, out, in, len
33%else
34cglobal opus_deemphasis, 4, 4, 8, out, in, coeff, len
35%endif
36%if ARCH_X86_32
37    VBROADCASTSS m0, coeffm
38%elif WIN64
39    shufps m0, m2, m2, 0
40%else
41    shufps m0, m0, 0
42%endif
43
44    movaps m4, [tab_st]
45    VBROADCASTSS m5, m4
46    shufps m6, m4, m4, q1111
47    shufps m7, m4, m4, q2222
48
49.loop:
50    movaps  m1, [inq]                ; x0, x1, x2, x3
51
52    pslldq  m2, m1, 4                ;  0, x0, x1, x2
53    pslldq  m3, m1, 8                ;  0,  0, x0, x1
54
55    fmaddps m2, m2, m5, m1           ; x + c1*x[0-2]
56    pslldq  m1, 12                   ;  0,  0,  0, x0
57
58    fmaddps m2, m3, m6, m2           ; x + c1*x[0-2] + c2*x[0-1]
59    fmaddps m1, m1, m7, m2           ; x + c1*x[0-2] + c2*x[0-1] + c3*x[0]
60    fmaddps m0, m0, m4, m1           ; x + c1*x[0-2] + c2*x[0-1] + c3*x[0] + c*s
61
62    movaps [outq], m0
63    shufps m0, m0, q3333             ; new state
64
65    add inq,  mmsize
66    add outq, mmsize
67    sub lend, mmsize >> 2
68    jg .loop
69
70%if ARCH_X86_64 == 0
71    movss r0m, m0
72    fld dword r0m
73%endif
74    RET
75
76
77INIT_XMM fma3
78cglobal opus_postfilter, 4, 4, 8, data, period, gains, len
79    VBROADCASTSS m0, [gainsq + 0]
80    VBROADCASTSS m1, [gainsq + 4]
81    VBROADCASTSS m2, [gainsq + 8]
82
83    shl periodd, 2
84    add periodq, 8
85    neg periodq
86
87    movups  m3, [dataq + periodq]
88    mulps   m3, m2
89
90.loop:
91    movups  m4, [dataq + periodq +  4]
92    movups  m5, [dataq + periodq +  8]
93    movups  m6, [dataq + periodq + 12]
94    movups  m7, [dataq + periodq + 16]
95
96    fmaddps m3, m7, m2, m3
97    addps   m6, m4
98
99    fmaddps m5, m5, m0, [dataq]
100    fmaddps m6, m6, m1, m3
101
102    addps   m5, m6
103    mulps   m3, m7, m2
104
105    movaps  [dataq], m5
106
107    add dataq, mmsize
108    sub lend,  mmsize >> 2
109    jg .loop
110
111    RET
112