1 /* ----------------------------------------------------------------------
2 * Project: CMSIS DSP Library
3 * Title: arm_rms_f32.c
4 * Description: Root mean square value of the elements of a floating-point vector
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/statistics_functions.h"
30
31 /**
32 @ingroup groupStats
33 */
34
35 /**
36 @defgroup RMS Root mean square (RMS)
37
38 Calculates the Root Mean Square of the elements in the input vector.
39 The underlying algorithm is used:
40
41 <pre>
42 Result = sqrt(((pSrc[0] * pSrc[0] + pSrc[1] * pSrc[1] + ... + pSrc[blockSize-1] * pSrc[blockSize-1]) / blockSize));
43 </pre>
44
45 There are separate functions for floating point, Q31, and Q15 data types.
46 */
47
48 /**
49 @addtogroup RMS
50 @{
51 */
52
53 /**
54 @brief Root Mean Square of the elements of a floating-point vector.
55 @param[in] pSrc points to the input vector
56 @param[in] blockSize number of samples in input vector
57 @param[out] pResult root mean square value returned here
58 */
59
60 #if defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE)
arm_rms_f32(const float32_t * pSrc,uint32_t blockSize,float32_t * pResult)61 ARM_DSP_ATTRIBUTE void arm_rms_f32(
62 const float32_t * pSrc,
63 uint32_t blockSize,
64 float32_t * pResult)
65 {
66 float32_t pow = 0.0f;
67
68 arm_power_f32(pSrc, blockSize, &pow);
69
70 /* Compute Rms and store the result in the destination */
71 arm_sqrt_f32(pow / (float32_t) blockSize, pResult);
72 }
73 #else
74 #if defined(ARM_MATH_NEON) && !defined(ARM_MATH_AUTOVECTORIZE)
arm_rms_f32(const float32_t * pSrc,uint32_t blockSize,float32_t * pResult)75 ARM_DSP_ATTRIBUTE void arm_rms_f32(
76 const float32_t * pSrc,
77 uint32_t blockSize,
78 float32_t * pResult)
79 {
80 float32_t sum = 0.0f; /* accumulator */
81 float32_t in; /* Temporary variable to store input value */
82 uint32_t blkCnt; /* loop counter */
83
84 float32x4_t sumV = vdupq_n_f32(0.0f); /* Temporary result storage */
85 float32x2_t sumV2;
86 float32x4_t inV;
87
88 blkCnt = blockSize >> 2U;
89
90 /* Compute 4 outputs at a time.
91 ** a second loop below computes the remaining 1 to 3 samples. */
92 while (blkCnt > 0U)
93 {
94 /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */
95 /* Compute Power and then store the result in a temporary variable, sum. */
96 inV = vld1q_f32(pSrc);
97 sumV = vmlaq_f32(sumV, inV, inV);
98 pSrc += 4;
99
100 /* Decrement the loop counter */
101 blkCnt--;
102 }
103
104 sumV2 = vpadd_f32(vget_low_f32(sumV),vget_high_f32(sumV));
105 sum = vget_lane_f32(sumV2, 0) + vget_lane_f32(sumV2, 1);
106
107 /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
108 ** No loop unrolling is used. */
109 blkCnt = blockSize % 0x4U;
110
111 while (blkCnt > 0U)
112 {
113 /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */
114 /* compute power and then store the result in a temporary variable, sum. */
115 in = *pSrc++;
116 sum += in * in;
117
118 /* Decrement the loop counter */
119 blkCnt--;
120 }
121
122 /* Compute Rms and store the result in the destination */
123 arm_sqrt_f32(sum / (float32_t) blockSize, pResult);
124 }
125 #else
arm_rms_f32(const float32_t * pSrc,uint32_t blockSize,float32_t * pResult)126 ARM_DSP_ATTRIBUTE void arm_rms_f32(
127 const float32_t * pSrc,
128 uint32_t blockSize,
129 float32_t * pResult)
130 {
131 uint32_t blkCnt; /* Loop counter */
132 float32_t sum = 0.0f; /* Temporary result storage */
133 float32_t in; /* Temporary variable to store input value */
134
135 #if defined (ARM_MATH_LOOPUNROLL) && !defined(ARM_MATH_AUTOVECTORIZE)
136
137 /* Loop unrolling: Compute 4 outputs at a time */
138 blkCnt = blockSize >> 2U;
139
140 while (blkCnt > 0U)
141 {
142 /* C = A[0] * A[0] + A[1] * A[1] + ... + A[blockSize-1] * A[blockSize-1] */
143
144 in = *pSrc++;
145 /* Compute sum of squares and store result in a temporary variable, sum. */
146 sum += in * in;
147
148 in = *pSrc++;
149 sum += in * in;
150
151 in = *pSrc++;
152 sum += in * in;
153
154 in = *pSrc++;
155 sum += in * in;
156
157 /* Decrement loop counter */
158 blkCnt--;
159 }
160
161 /* Loop unrolling: Compute remaining outputs */
162 blkCnt = blockSize % 0x4U;
163
164 #else
165
166 /* Initialize blkCnt with number of samples */
167 blkCnt = blockSize;
168
169 #endif /* #if defined (ARM_MATH_LOOPUNROLL) */
170
171 while (blkCnt > 0U)
172 {
173 /* C = A[0] * A[0] + A[1] * A[1] + ... + A[blockSize-1] * A[blockSize-1] */
174
175 in = *pSrc++;
176 /* Compute sum of squares and store result in a temporary variable. */
177 sum += ( in * in);
178
179 /* Decrement loop counter */
180 blkCnt--;
181 }
182
183 /* Compute Rms and store result in destination */
184 arm_sqrt_f32(sum / (float32_t) blockSize, pResult);
185 }
186 #endif /* #if defined(ARM_MATH_NEON) */
187 #endif /* defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE) */
188
189 /**
190 @} end of RMS group
191 */
192
