1 /* Copyright (C) 2004 Jean-Marc Valin */
2 /**
3 @file cb_search_arm4.h
4 @brief Fixed codebook functions (ARM4 version)
5 */
6 /*
7 Redistribution and use in source and binary forms, with or without
8 modification, are permitted provided that the following conditions
9 are met:
10
11 - Redistributions of source code must retain the above copyright
12 notice, this list of conditions and the following disclaimer.
13
14 - Redistributions in binary form must reproduce the above copyright
15 notice, this list of conditions and the following disclaimer in the
16 documentation and/or other materials provided with the distribution.
17
18 - Neither the name of the Xiph.org Foundation nor the names of its
19 contributors may be used to endorse or promote products derived from
20 this software without specific prior written permission.
21
22 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
23 ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
24 LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
25 A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR
26 CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
27 EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
28 PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
29 PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
30 LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
31 NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
32 SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
33 */
34
35 #define OVERRIDE_COMPUTE_WEIGHTED_CODEBOOK
compute_weighted_codebook(const signed char * shape_cb,const spx_sig_t * r,spx_word16_t * resp,spx_word16_t * resp2,spx_word32_t * E,int shape_cb_size,int subvect_size,char * stack)36 static void compute_weighted_codebook(const signed char *shape_cb, const spx_sig_t *r, spx_word16_t *resp, spx_word16_t *resp2, spx_word32_t *E, int shape_cb_size, int subvect_size, char *stack)
37 {
38 int i, j, k;
39 //const signed char *shape;
40 for (i=0;i<shape_cb_size;i+=4)
41 {
42
43 //shape = shape_cb;
44 E[0]=0;
45 E[1]=0;
46 E[2]=0;
47 E[3]=0;
48
49 /* Compute codeword response using convolution with impulse response */
50 for(j=0;j<subvect_size;j++)
51 {
52 #if 1
53 spx_word16_t *res;
54 res = resp+j;
55 spx_word32_t resj0,resj1,resj2,resj3;
56 spx_word32_t dead1, dead2, dead3, dead4, dead5, dead6, dead7, dead8;
57 __asm__ __volatile__ (
58 "mov %0, #0 \n\t"
59 "mov %1, #0 \n\t"
60 "mov %2, #0 \n\t"
61 "mov %3, #0 \n\t"
62 ".weighted%=: \n\t"
63 "ldrsb %8, [%6] \n\t"
64 "ldr %10, [%5], #-4 \n\t"
65 "mov %9, %6 \n\t"
66 "ldrsb %11, [%9, %7]! \n\t"
67 "mla %0, %10, %8, %0 \n\t"
68 "ldrsb %8, [%9, %7]! \n\t"
69 "mla %1, %10, %11, %1 \n\t"
70 "ldrsb %11, [%9, %7]! \n\t"
71 "mla %2, %10, %8, %2 \n\t"
72 "subs %4, %4, #1 \n\t"
73 "mla %3, %10, %11, %3 \n\t"
74 "add %6, %6, #1 \n\t"
75 "bne .weighted%= \n\t"
76 : "=r" (resj0), "=r" (resj1), "=r" (resj2), "=r" (resj3),
77 "=r" (dead1), "=r" (dead2), "=r" (dead3), "=r" (dead4),
78 "=r" (dead5), "=r" (dead6), "=r" (dead7), "=r" (dead8)
79 : "4" (j+1), "5" (r+j), "6" (shape_cb), "7" (subvect_size)
80 : "cc", "memory");
81 #else
82 spx_word16_t *res;
83 res = resp+j;
84 spx_word32_t resj0=0;
85 spx_word32_t resj1=0;
86 spx_word32_t resj2=0;
87 spx_word32_t resj3=0;
88 for (k=0;k<=j;k++)
89 {
90 const signed char *shape=shape_cb+k;
91 resj0 = MAC16_16(resj0,*shape,r[j-k]);
92 shape += subvect_size;
93 resj1 = MAC16_16(resj1,*shape,r[j-k]);
94 shape += subvect_size;
95 resj2 = MAC16_16(resj2,*shape,r[j-k]);
96 shape += subvect_size;
97 resj3 = MAC16_16(resj3,*shape,r[j-k]);
98 shape += subvect_size;
99 }
100 #endif
101
102 #ifdef FIXED_POINT
103 resj0 = SHR(resj0, 11);
104 resj1 = SHR(resj1, 11);
105 resj2 = SHR(resj2, 11);
106 resj3 = SHR(resj3, 11);
107 #else
108 resj0 *= 0.03125;
109 resj1 *= 0.03125;
110 resj2 *= 0.03125;
111 resj3 *= 0.03125;
112 #endif
113
114 /* Compute codeword energy */
115 E[0]=ADD32(E[0],MULT16_16(resj0,resj0));
116 E[1]=ADD32(E[1],MULT16_16(resj1,resj1));
117 E[2]=ADD32(E[2],MULT16_16(resj2,resj2));
118 E[3]=ADD32(E[3],MULT16_16(resj3,resj3));
119 *res = resj0;
120 res += subvect_size;
121 *res = resj1;
122 res += subvect_size;
123 *res = resj2;
124 res += subvect_size;
125 *res = resj3;
126 res += subvect_size;
127 }
128 resp += subvect_size<<2;
129 shape_cb += subvect_size<<2;
130 E+=4;
131 }
132
133 }
134