tmk_core/tool/mbed/mbed-sdk/libraries/dsp/cmsis_dsp/StatisticsFunctions/arm_std_q31.c

   1 /* ----------------------------------------------------------------------
   2 * Copyright (C) 2010-2013 ARM Limited. All rights reserved.
   3 *
   4 * $Date:        17. January 2013
   5 * $Revision:    V1.4.1
   6 *
   7 * Project:          CMSIS DSP Library
   8 * Title:                arm_std_q31.c
   9 *
  10 * Description:  Standard deviation of an array of Q31 type.
  11 *
  12 * Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
  13 *
  14 * Redistribution and use in source and binary forms, with or without
  15 * modification, are permitted provided that the following conditions
  16 * are met:
  17 *   - Redistributions of source code must retain the above copyright
  18 *     notice, this list of conditions and the following disclaimer.
  19 *   - Redistributions in binary form must reproduce the above copyright
  20 *     notice, this list of conditions and the following disclaimer in
  21 *     the documentation and/or other materials provided with the
  22 *     distribution.
  23 *   - Neither the name of ARM LIMITED nor the names of its contributors
  24 *     may be used to endorse or promote products derived from this
  25 *     software without specific prior written permission.
  26 *
  27 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
  28 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
  29 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
  30 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
  31 * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
  32 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
  33 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
  34 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
  35 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  36 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
  37 * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
  38 * POSSIBILITY OF SUCH DAMAGE.
  39 * -------------------------------------------------------------------- */
  40
  41 #include "arm_math.h"
  42
  43 /**
  44  * @ingroup groupStats
  45  */
  46
  47 /**
  48  * @addtogroup STD
  49  * @{
  50  */
  51
  52
  53 /**
  54  * @brief Standard deviation of the elements of a Q31 vector.
  55  * @param[in]       *pSrc points to the input vector
  56  * @param[in]       blockSize length of the input vector
  57  * @param[out]      *pResult standard deviation value returned here
  58  * @return none.
  59  * @details
  60  * <b>Scaling and Overflow Behavior:</b>
  61  *
  62  *\par
  63  * The function is implemented using an internal 64-bit accumulator.
  64  * The input is represented in 1.31 format, and intermediate multiplication
  65  * yields a 2.62 format.
  66  * The accumulator maintains full precision of the intermediate multiplication results,
  67  * but provides only a single guard bit.
  68  * There is no saturation on intermediate additions.
  69  * If the accumulator overflows it wraps around and distorts the result.
  70  * In order to avoid overflows completely the input signal must be scaled down by
  71  * log2(blockSize) bits, as a total of blockSize additions are performed internally.
  72  * Finally, the 2.62 accumulator is right shifted by 31 bits to yield a 1.31 format value.
  73  *
  74  */
  75
  76
  77 void arm_std_q31(
  78   q31_t * pSrc,
  79   uint32_t blockSize,
  80   q31_t * pResult)
  81 {
  82   q63_t sum = 0;                                 /* Accumulator */
  83   q31_t meanOfSquares, squareOfMean;             /* square of mean and mean of square */
  84   q31_t mean;                                    /* mean */
  85   q31_t in;                                      /* input value */
  86   q31_t t;                                       /* Temporary variable */
  87   uint32_t blkCnt;                               /* loop counter */
  88   q63_t sumOfSquares = 0;                        /* Accumulator */
  89
  90 #ifndef ARM_MATH_CM0_FAMILY
  91
  92   /* Run the below code for Cortex-M4 and Cortex-M3 */
  93
  94   /*loop Unrolling */
  95   blkCnt = blockSize >> 2u;
  96
  97   /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.
  98    ** a second loop below computes the remaining 1 to 3 samples. */
  99   while(blkCnt > 0u)
 100   {
 101     /* C = (A[0] * A[0] + A[1] * A[1] + ... + A[blockSize-1] * A[blockSize-1])  */
 102     /* Compute Sum of squares of the input samples
 103      * and then store the result in a temporary variable, sum. */
 104     in = *pSrc++;
 105     sum += in;
 106     sumOfSquares += ((q63_t) (in) * (in));
 107     in = *pSrc++;
 108     sum += in;
 109     sumOfSquares += ((q63_t) (in) * (in));
 110     in = *pSrc++;
 111     sum += in;
 112     sumOfSquares += ((q63_t) (in) * (in));
 113     in = *pSrc++;
 114     sum += in;
 115     sumOfSquares += ((q63_t) (in) * (in));
 116
 117     /* Decrement the loop counter */
 118     blkCnt--;
 119   }
 120
 121   /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
 122    ** No loop unrolling is used. */
 123   blkCnt = blockSize % 0x4u;
 124
 125   while(blkCnt > 0u)
 126   {
 127     /* C = (A[0] * A[0] + A[1] * A[1] + ... + A[blockSize-1] * A[blockSize-1]) */
 128     /* Compute Sum of squares of the input samples
 129      * and then store the result in a temporary variable, sum. */
 130     in = *pSrc++;
 131     sum += in;
 132     sumOfSquares += ((q63_t) (in) * (in));
 133
 134     /* Decrement the loop counter */
 135     blkCnt--;
 136   }
 137
 138   t = (q31_t) ((1.0f / (float32_t) (blockSize - 1u)) * 1073741824.0f);
 139
 140   /* Compute Mean of squares of the input samples
 141    * and then store the result in a temporary variable, meanOfSquares. */
 142   sumOfSquares = (sumOfSquares >> 31);
 143   meanOfSquares = (q31_t) ((sumOfSquares * t) >> 30);
 144
 145 #else
 146
 147   /* Run the below code for Cortex-M0 */
 148
 149   /* Loop over blockSize number of values */
 150   blkCnt = blockSize;
 151
 152   while(blkCnt > 0u)
 153   {
 154     /* C = (A[0] * A[0] + A[1] * A[1] + ... + A[blockSize-1] * A[blockSize-1]) */
 155     /* Compute Sum of squares of the input samples
 156      * and then store the result in a temporary variable, sumOfSquares. */
 157     in = *pSrc++;
 158     sumOfSquares += ((q63_t) (in) * (in));
 159
 160     /* C = (A[0] + A[1] + A[2] + ... + A[blockSize-1]) */
 161     /* Compute sum of all input values and then store the result in a temporary variable, sum. */
 162     sum += in;
 163
 164     /* Decrement the loop counter */
 165     blkCnt--;
 166   }
 167
 168   /* Compute Mean of squares of the input samples
 169    * and then store the result in a temporary variable, meanOfSquares. */
 170   t = (q31_t) ((1.0f / (float32_t) (blockSize - 1u)) * 1073741824.0f);
 171   sumOfSquares = (sumOfSquares >> 31);
 172   meanOfSquares = (q31_t) ((sumOfSquares * t) >> 30);
 173
 174 #endif /* #ifndef ARM_MATH_CM0_FAMILY */
 175
 176   /* Compute mean of all input values */
 177   t = (q31_t) ((1.0f / (blockSize * (blockSize - 1u))) * 2147483648.0f);
 178   mean = (q31_t) (sum);
 179
 180   /* Compute square of mean */
 181   squareOfMean = (q31_t) (((q63_t) mean * mean) >> 31);
 182   squareOfMean = (q31_t) (((q63_t) squareOfMean * t) >> 31);
 183
 184
 185   /* Compute standard deviation and then store the result to the destination */
 186   arm_sqrt_q31(meanOfSquares - squareOfMean, pResult);
 187
 188 }
 189
 190 /**
 191  * @} end of STD group
 192  */