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 */
69
70 #if defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE)
71
arm_cmplx_conj_f32(const float32_t * pSrc,float32_t * pDst,uint32_t numSamples)72 void arm_cmplx_conj_f32(
73 const float32_t * pSrc,
74 float32_t * pDst,
75 uint32_t numSamples)
76 {
77 static const float32_t cmplx_conj_sign[4] = { 1.0f, -1.0f, 1.0f, -1.0f };
78 uint32_t blockSize = numSamples * CMPLX_DIM; /* loop counters */
79 uint32_t blkCnt;
80 f32x4_t vecSrc;
81 f32x4_t vecSign;
82
83 /*
84 * load sign vector
85 */
86 vecSign = *(f32x4_t *) cmplx_conj_sign;
87
88 /* Compute 4 real samples at a time */
89 blkCnt = blockSize >> 2U;
90
91 while (blkCnt > 0U)
92 {
93 vecSrc = vld1q(pSrc);
94 vst1q(pDst,vmulq(vecSrc, vecSign));
95 /*
96 * Decrement the blkCnt loop counter
97 * Advance vector source and destination pointers
98 */
99 pSrc += 4;
100 pDst += 4;
101 blkCnt--;
102 }
103
104 /* Tail */
105 blkCnt = (blockSize & 0x3) >> 1;
106
107 while (blkCnt > 0U)
108 {
109 /* C[0] + jC[1] = A[0]+ j(-1)A[1] */
110
111 /* Calculate Complex Conjugate and store result in destination buffer. */
112 *pDst++ = *pSrc++;
113 *pDst++ = -*pSrc++;
114
115 /* Decrement loop counter */
116 blkCnt--;
117 }
118
119 }
120
121 #else
arm_cmplx_conj_f32(const float32_t * pSrc,float32_t * pDst,uint32_t numSamples)122 void arm_cmplx_conj_f32(
123 const float32_t * pSrc,
124 float32_t * pDst,
125 uint32_t numSamples)
126 {
127 uint32_t blkCnt; /* Loop counter */
128
129 #if defined(ARM_MATH_NEON) && !defined(ARM_MATH_AUTOVECTORIZE)
130 float32x4_t zero;
131 float32x4x2_t vec;
132
133 zero = vdupq_n_f32(0.0f);
134
135 /* Compute 4 outputs at a time */
136 blkCnt = numSamples >> 2U;
137
138 while (blkCnt > 0U)
139 {
140 /* C[0]+jC[1] = A[0]+(-1)*jA[1] */
141 /* Calculate Complex Conjugate and then store the results in the destination buffer. */
142 vec = vld2q_f32(pSrc);
143 vec.val[1] = vsubq_f32(zero,vec.val[1]);
144 vst2q_f32(pDst,vec);
145
146 /* Increment pointers */
147 pSrc += 8;
148 pDst += 8;
149
150 /* Decrement the loop counter */
151 blkCnt--;
152 }
153
154 /* Tail */
155 blkCnt = numSamples & 0x3;
156
157 #else
158 #if defined (ARM_MATH_LOOPUNROLL) && !defined(ARM_MATH_AUTOVECTORIZE)
159
160 /* Loop unrolling: Compute 4 outputs at a time */
161 blkCnt = numSamples >> 2U;
162
163 while (blkCnt > 0U)
164 {
165 /* C[0] + jC[1] = A[0]+ j(-1)A[1] */
166
167 /* Calculate Complex Conjugate and store result in destination buffer. */
168 *pDst++ = *pSrc++;
169 *pDst++ = -*pSrc++;
170
171 *pDst++ = *pSrc++;
172 *pDst++ = -*pSrc++;
173
174 *pDst++ = *pSrc++;
175 *pDst++ = -*pSrc++;
176
177 *pDst++ = *pSrc++;
178 *pDst++ = -*pSrc++;
179
180 /* Decrement loop counter */
181 blkCnt--;
182 }
183
184 /* Loop unrolling: Compute remaining outputs */
185 blkCnt = numSamples % 0x4U;
186
187 #else
188
189 /* Initialize blkCnt with number of samples */
190 blkCnt = numSamples;
191
192 #endif /* #if defined (ARM_MATH_LOOPUNROLL) */
193 #endif /* #if defined (ARM_MATH_NEON) */
194
195 while (blkCnt > 0U)
196 {
197 /* C[0] + jC[1] = A[0]+ j(-1)A[1] */
198
199 /* Calculate Complex Conjugate and store result in destination buffer. */
200 *pDst++ = *pSrc++;
201 *pDst++ = -*pSrc++;
202
203 /* Decrement loop counter */
204 blkCnt--;
205 }
206
207 }
208 #endif /* defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE) */
209
210 /**
211 @} end of cmplx_conj group
212 */
213
