1 /* ----------------------------------------------------------------------
2 * Project: CMSIS DSP Library
3 * Title: arm_cmplx_conj_f32.c
4 * Description: Floating-point 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 @defgroup cmplx_conj Complex Conjugate
37
38 Conjugates the elements of a complex data vector.
39
40 The <code>pSrc</code> points to the source data and
41 <code>pDst</code> points to the destination data where the result should be written.
42 <code>numSamples</code> specifies the number of complex samples
43 and the data in each array is stored in an interleaved fashion
44 (real, imag, real, imag, ...).
45 Each array has a total of <code>2*numSamples</code> values.
46
47 The underlying algorithm is used:
48 <pre>
49 for (n = 0; n < numSamples; n++) {
50 pDst[(2*n) ] = pSrc[(2*n) ]; // real part
51 pDst[(2*n)+1] = -pSrc[(2*n)+1]; // imag part
52 }
53 </pre>
54
55 There are separate functions for floating-point, Q15, and Q31 data types.
56 */
57
58 /**
59 @addtogroup cmplx_conj
60 @{
61 */
62
63 /**
64 @brief Floating-point complex conjugate.
65 @param[in] pSrc points to the input vector
66 @param[out] pDst points to the output vector
67 @param[in] numSamples number of samples in each vector
68 @return none
69 */
70
71 #if defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE)
72
arm_cmplx_conj_f32(const float32_t * pSrc,float32_t * pDst,uint32_t numSamples)73 void arm_cmplx_conj_f32(
74 const float32_t * pSrc,
75 float32_t * pDst,
76 uint32_t numSamples)
77 {
78 static const float32_t cmplx_conj_sign[4] = { 1.0f, -1.0f, 1.0f, -1.0f };
79 uint32_t blockSize = numSamples * CMPLX_DIM; /* loop counters */
80 uint32_t blkCnt;
81 f32x4_t vecSrc;
82 f32x4_t vecSign;
83
84 /*
85 * load sign vector
86 */
87 vecSign = *(f32x4_t *) cmplx_conj_sign;
88
89 /* Compute 4 real samples at a time */
90 blkCnt = blockSize >> 2U;
91
92 while (blkCnt > 0U)
93 {
94 vecSrc = vld1q(pSrc);
95 vst1q(pDst,vmulq(vecSrc, vecSign));
96 /*
97 * Decrement the blkCnt loop counter
98 * Advance vector source and destination pointers
99 */
100 pSrc += 4;
101 pDst += 4;
102 blkCnt--;
103 }
104
105 /* Tail */
106 blkCnt = (blockSize & 0x3) >> 1;
107
108 while (blkCnt > 0U)
109 {
110 /* C[0] + jC[1] = A[0]+ j(-1)A[1] */
111
112 /* Calculate Complex Conjugate and store result in destination buffer. */
113 *pDst++ = *pSrc++;
114 *pDst++ = -*pSrc++;
115
116 /* Decrement loop counter */
117 blkCnt--;
118 }
119
120 }
121
122 #else
arm_cmplx_conj_f32(const float32_t * pSrc,float32_t * pDst,uint32_t numSamples)123 void arm_cmplx_conj_f32(
124 const float32_t * pSrc,
125 float32_t * pDst,
126 uint32_t numSamples)
127 {
128 uint32_t blkCnt; /* Loop counter */
129
130 #if defined(ARM_MATH_NEON) && !defined(ARM_MATH_AUTOVECTORIZE)
131 float32x4_t zero;
132 float32x4x2_t vec;
133
134 zero = vdupq_n_f32(0.0f);
135
136 /* Compute 4 outputs at a time */
137 blkCnt = numSamples >> 2U;
138
139 while (blkCnt > 0U)
140 {
141 /* C[0]+jC[1] = A[0]+(-1)*jA[1] */
142 /* Calculate Complex Conjugate and then store the results in the destination buffer. */
143 vec = vld2q_f32(pSrc);
144 vec.val[1] = vsubq_f32(zero,vec.val[1]);
145 vst2q_f32(pDst,vec);
146
147 /* Increment pointers */
148 pSrc += 8;
149 pDst += 8;
150
151 /* Decrement the loop counter */
152 blkCnt--;
153 }
154
155 /* Tail */
156 blkCnt = numSamples & 0x3;
157
158 #else
159 #if defined (ARM_MATH_LOOPUNROLL) && !defined(ARM_MATH_AUTOVECTORIZE)
160
161 /* Loop unrolling: Compute 4 outputs at a time */
162 blkCnt = numSamples >> 2U;
163
164 while (blkCnt > 0U)
165 {
166 /* C[0] + jC[1] = A[0]+ j(-1)A[1] */
167
168 /* Calculate Complex Conjugate and store result in destination buffer. */
169 *pDst++ = *pSrc++;
170 *pDst++ = -*pSrc++;
171
172 *pDst++ = *pSrc++;
173 *pDst++ = -*pSrc++;
174
175 *pDst++ = *pSrc++;
176 *pDst++ = -*pSrc++;
177
178 *pDst++ = *pSrc++;
179 *pDst++ = -*pSrc++;
180
181 /* Decrement loop counter */
182 blkCnt--;
183 }
184
185 /* Loop unrolling: Compute remaining outputs */
186 blkCnt = numSamples % 0x4U;
187
188 #else
189
190 /* Initialize blkCnt with number of samples */
191 blkCnt = numSamples;
192
193 #endif /* #if defined (ARM_MATH_LOOPUNROLL) */
194 #endif /* #if defined (ARM_MATH_NEON) */
195
196 while (blkCnt > 0U)
197 {
198 /* C[0] + jC[1] = A[0]+ j(-1)A[1] */
199
200 /* Calculate Complex Conjugate and store result in destination buffer. */
201 *pDst++ = *pSrc++;
202 *pDst++ = -*pSrc++;
203
204 /* Decrement loop counter */
205 blkCnt--;
206 }
207
208 }
209 #endif /* defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE) */
210
211 /**
212 @} end of cmplx_conj group
213 */
214
