1 /* ----------------------------------------------------------------------
2 * Project: CMSIS DSP Library
3 * Title: arm_cmplx_conj_q15.c
4 * Description: Q15 complex conjugate
5 *
6 * $Date: 23 April 2021
7 * $Revision: V1.9.0
8 *
9 * Target Processor: Cortex-M and Cortex-A cores
10 * -------------------------------------------------------------------- */
11 /*
12 * Copyright (C) 2010-2021 ARM Limited or its affiliates. All rights reserved.
13 *
14 * SPDX-License-Identifier: Apache-2.0
15 *
16 * Licensed under the Apache License, Version 2.0 (the License); you may
17 * not use this file except in compliance with the License.
18 * You may obtain a copy of the License at
19 *
20 * www.apache.org/licenses/LICENSE-2.0
21 *
22 * Unless required by applicable law or agreed to in writing, software
23 * distributed under the License is distributed on an AS IS BASIS, WITHOUT
24 * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
25 * See the License for the specific language governing permissions and
26 * limitations under the License.
27 */
28
29 #include "dsp/complex_math_functions.h"
30
31 /**
32 @ingroup groupCmplxMath
33 */
34
35 /**
36 @addtogroup cmplx_conj
37 @{
38 */
39
40 /**
41 @brief Q15 complex conjugate.
42 @param[in] pSrc points to the input vector
43 @param[out] pDst points to the output vector
44 @param[in] numSamples number of samples in each vector
45
46 @par Scaling and Overflow Behavior
47 The function uses saturating arithmetic.
48 The Q15 value -1 (0x8000) is saturated to the maximum allowable positive value 0x7FFF.
49 */
50
51
52 #if defined(ARM_MATH_MVEI) && !defined(ARM_MATH_AUTOVECTORIZE)
arm_cmplx_conj_q15(const q15_t * pSrc,q15_t * pDst,uint32_t numSamples)53 ARM_DSP_ATTRIBUTE void arm_cmplx_conj_q15(
54 const q15_t * pSrc,
55 q15_t * pDst,
56 uint32_t numSamples)
57 {
58 uint32_t blockSize = numSamples * CMPLX_DIM; /* loop counters */
59 uint32_t blkCnt;
60 q31_t in1;
61
62 q15x8x2_t vecSrc;
63 q15x8_t zero;
64
65 zero = vdupq_n_s16(0);
66
67 /* Compute 8 real samples at a time */
68 blkCnt = blockSize >> 4U;
69 while (blkCnt > 0U)
70 {
71 vecSrc = vld2q(pSrc);
72 vecSrc.val[1] = vqsubq(zero, vecSrc.val[1]);
73 vst2q(pDst,vecSrc);
74 /*
75 * Decrement the blkCnt loop counter
76 * Advance vector source and destination pointers
77 */
78 pSrc += 16;
79 pDst += 16;
80 blkCnt --;
81 }
82
83 /* Tail */
84 blkCnt = (blockSize & 0xF) >> 1;
85
86 while (blkCnt > 0U)
87 {
88 /* C[0] + jC[1] = A[0]+ j(-1)A[1] */
89
90 /* Calculate Complex Conjugate and store result in destination buffer. */
91 *pDst++ = *pSrc++;
92 in1 = *pSrc++;
93 *pDst++ = __SSAT(-in1, 16);
94
95 /* Decrement loop counter */
96 blkCnt--;
97 }
98 }
99 #else
arm_cmplx_conj_q15(const q15_t * pSrc,q15_t * pDst,uint32_t numSamples)100 ARM_DSP_ATTRIBUTE void arm_cmplx_conj_q15(
101 const q15_t * pSrc,
102 q15_t * pDst,
103 uint32_t numSamples)
104 {
105 uint32_t blkCnt; /* Loop counter */
106 q31_t in1; /* Temporary input variable */
107
108 #if defined (ARM_MATH_LOOPUNROLL) && defined (ARM_MATH_DSP)
109 q31_t in2, in3, in4; /* Temporary input variables */
110 #endif
111
112
113 #if defined (ARM_MATH_LOOPUNROLL)
114
115 /* Loop unrolling: Compute 4 outputs at a time */
116 blkCnt = numSamples >> 2U;
117
118 while (blkCnt > 0U)
119 {
120 /* C[0] + jC[1] = A[0]+ j(-1)A[1] */
121
122 /* Calculate Complex Conjugate and store result in destination buffer. */
123
124 #if defined (ARM_MATH_DSP)
125 in1 = read_q15x2_ia (&pSrc);
126 in2 = read_q15x2_ia (&pSrc);
127 in3 = read_q15x2_ia (&pSrc);
128 in4 = read_q15x2_ia (&pSrc);
129
130 #ifndef ARM_MATH_BIG_ENDIAN
131 in1 = __QASX(0, in1);
132 in2 = __QASX(0, in2);
133 in3 = __QASX(0, in3);
134 in4 = __QASX(0, in4);
135 #else
136 in1 = __QSAX(0, in1);
137 in2 = __QSAX(0, in2);
138 in3 = __QSAX(0, in3);
139 in4 = __QSAX(0, in4);
140 #endif /* #ifndef ARM_MATH_BIG_ENDIAN */
141
142 in1 = ((uint32_t) in1 >> 16) | ((uint32_t) in1 << 16);
143 in2 = ((uint32_t) in2 >> 16) | ((uint32_t) in2 << 16);
144 in3 = ((uint32_t) in3 >> 16) | ((uint32_t) in3 << 16);
145 in4 = ((uint32_t) in4 >> 16) | ((uint32_t) in4 << 16);
146
147 write_q15x2_ia (&pDst, in1);
148 write_q15x2_ia (&pDst, in2);
149 write_q15x2_ia (&pDst, in3);
150 write_q15x2_ia (&pDst, in4);
151 #else
152 *pDst++ = *pSrc++;
153 in1 = *pSrc++;
154 *pDst++ = (in1 == (q15_t) 0x8000) ? (q15_t) 0x7fff : -in1;
155
156 *pDst++ = *pSrc++;
157 in1 = *pSrc++;
158 *pDst++ = (in1 == (q15_t) 0x8000) ? (q15_t) 0x7fff : -in1;
159
160 *pDst++ = *pSrc++;
161 in1 = *pSrc++;
162 *pDst++ = (in1 == (q15_t) 0x8000) ? (q15_t) 0x7fff : -in1;
163
164 *pDst++ = *pSrc++;
165 in1 = *pSrc++;
166 *pDst++ = (in1 == (q15_t) 0x8000) ? (q15_t) 0x7fff : -in1;
167
168 #endif /* #if defined (ARM_MATH_DSP) */
169
170 /* Decrement loop counter */
171 blkCnt--;
172 }
173
174 /* Loop unrolling: Compute remaining outputs */
175 blkCnt = numSamples % 0x4U;
176
177 #else
178
179 /* Initialize blkCnt with number of samples */
180 blkCnt = numSamples;
181
182 #endif /* #if defined (ARM_MATH_LOOPUNROLL) */
183
184 while (blkCnt > 0U)
185 {
186 /* C[0] + jC[1] = A[0]+ j(-1)A[1] */
187
188 /* Calculate Complex Conjugate and store result in destination buffer. */
189 *pDst++ = *pSrc++;
190 in1 = *pSrc++;
191 #if defined (ARM_MATH_DSP)
192 *pDst++ = __SSAT(-in1, 16);
193 #else
194 *pDst++ = (in1 == (q15_t) 0x8000) ? (q15_t) 0x7fff : -in1;
195 #endif
196
197 /* Decrement loop counter */
198 blkCnt--;
199 }
200
201 }
202 #endif /* defined(ARM_MATH_MVEI) */
203
204 /**
205 @} end of cmplx_conj group
206 */
207
