arm_std_f32.c

00001 /* ----------------------------------------------------------------------  
00002 * Copyright (C) 2010 ARM Limited. All rights reserved.  
00003 *  
00004 * $Date:        29. November 2010  
00005 * $Revision:    V1.0.3  
00006 *  
00007 * Project:      CMSIS DSP Library  
00008 * Title:        arm_std_f32.c  
00009 *  
00010 * Description:  Standard deviation of an array of F32 type.  
00011 *  
00012 * Target Processor: Cortex-M4/Cortex-M3
00013 *  
00014 * Version 1.0.3 2010/11/29 
00015 *    Re-organized the CMSIS folders and updated documentation.  
00016 *   
00017 * Version 1.0.2 2010/11/11  
00018 *    Documentation updated.   
00019 *  
00020 * Version 1.0.1 2010/10/05   
00021 *    Production release and review comments incorporated.  
00022 *  
00023 * Version 1.0.0 2010/09/20   
00024 *    Production release and review comments incorporated.  
00025 * ---------------------------------------------------------------------------- */ 
00026  
00027 #include "arm_math.h" 
00028  
00029 /**  
00030  * @ingroup groupStats  
00031  */ 
00032  
00033 /**  
00034  * @defgroup STD Standard deviation  
00035  *  
00036  * Calculates the standard deviation of the elements in the input vector.   
00037  * The underlying algorithm is used:  
00038  * 
00039  * <pre>  
00040  *  Result = sqrt((sumOfSquares - sum<sup>2</sup> / blockSize) / (blockSize - 1)) 
00041  * 
00042  *     where, sumOfSquares = pSrc[0] * pSrc[0] + pSrc[1] * pSrc[1] + ... + pSrc[blockSize-1] * pSrc[blockSize-1] 
00043  * 
00044  *                     sum = pSrc[0] + pSrc[1] + pSrc[2] + ... + pSrc[blockSize-1] 
00045  * </pre> 
00046  *  
00047  * There are separate functions for floating point, Q31, and Q15 data types.  
00048  */ 
00049  
00050 /**  
00051  * @addtogroup STD  
00052  * @{  
00053  */ 
00054  
00055  
00056 /**  
00057  * @brief Standard deviation of the elements of a floating-point vector.  
00058  * @param[in]       *pSrc points to the input vector  
00059  * @param[in]       blockSize length of the input vector  
00060  * @param[out]      *pResult standard deviation value returned here  
00061  * @return none.  
00062  *  
00063  */ 
00064  
00065  
00066 void arm_std_f32( 
00067   float32_t * pSrc, 
00068   uint32_t blockSize, 
00069   float32_t * pResult) 
00070 { 
00071   float32_t sum = 0.0f;                            /* Temporary result storage */ 
00072   float32_t meanOfSquares, mean, in, squareOfMean; 
00073   uint32_t blkCnt;                                 /* loop counter */ 
00074   float32_t *pIn;                                  /* Temporary pointer */ 
00075  
00076   pIn = pSrc; 
00077  
00078  
00079   /*loop Unrolling */ 
00080   blkCnt = blockSize >> 2u; 
00081  
00082   /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.  
00083    ** a second loop below computes the remaining 1 to 3 samples. */ 
00084   while(blkCnt > 0u) 
00085   { 
00086     /* C = (A[0] * A[0] + A[1] * A[1] + ... + A[blockSize-1] * A[blockSize-1])  */ 
00087     /* Compute Sum of squares of the input samples  
00088      * and then store the result in a temporary variable, sum. */ 
00089     in = *pSrc++; 
00090     sum += in * in; 
00091     in = *pSrc++; 
00092     sum += in * in; 
00093     in = *pSrc++; 
00094     sum += in * in; 
00095     in = *pSrc++; 
00096     sum += in * in; 
00097  
00098     /* Decrement the loop counter */ 
00099     blkCnt--; 
00100   } 
00101  
00102   /* If the blockSize is not a multiple of 4, compute any remaining output samples here.  
00103    ** No loop unrolling is used. */ 
00104   blkCnt = blockSize % 0x4u; 
00105  
00106   while(blkCnt > 0u) 
00107   { 
00108     /* C = (A[0] * A[0] + A[1] * A[1] + ... + A[blockSize-1] * A[blockSize-1]) */ 
00109     /* Compute Sum of squares of the input samples  
00110      * and then store the result in a temporary variable, sum. */ 
00111     in = *pSrc++; 
00112     sum += in * in; 
00113  
00114     /* Decrement the loop counter */ 
00115     blkCnt--; 
00116   } 
00117  
00118   /* Compute Mean of squares of the input samples  
00119    * and then store the result in a temporary variable, meanOfSquares. */ 
00120   meanOfSquares = sum / ((float32_t) blockSize - 1.0f); 
00121  
00122   /* Reset the accumulator */ 
00123   sum = 0.0f; 
00124  
00125   /*loop Unrolling */ 
00126   blkCnt = blockSize >> 2u; 
00127  
00128   /* Reset the input working pointer */ 
00129   pSrc = pIn; 
00130  
00131   /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.  
00132    ** a second loop below computes the remaining 1 to 3 samples. */ 
00133   while(blkCnt > 0u) 
00134   { 
00135     /* C = (A[0] + A[1] + A[2] + ... + A[blockSize-1]) */ 
00136     /* Compute sum of all input values and then store the result in a temporary variable, sum. */ 
00137     sum += *pSrc++; 
00138     sum += *pSrc++; 
00139     sum += *pSrc++; 
00140     sum += *pSrc++; 
00141  
00142     /* Decrement the loop counter */ 
00143     blkCnt--; 
00144   } 
00145  
00146   /* If the blockSize is not a multiple of 4, compute any remaining output samples here.  
00147    ** No loop unrolling is used. */ 
00148   blkCnt = blockSize % 0x4u; 
00149  
00150   while(blkCnt > 0u) 
00151   { 
00152     /* C = (A[0] + A[1] + A[2] + ... + A[blockSize-1]) */ 
00153     /* Compute sum of all input values and then store the result in a temporary variable, sum. */ 
00154     sum += *pSrc++; 
00155  
00156     /* Decrement the loop counter */ 
00157     blkCnt--; 
00158   } 
00159   /* Compute mean of all input values */ 
00160   mean = sum / (float32_t) blockSize; 
00161  
00162   /* Compute square of mean */ 
00163   squareOfMean = (mean * mean) * (((float32_t) blockSize) / 
00164                                   ((float32_t) blockSize - 1.0f)); 
00165  
00166   /* Compute standard deviation and then store the result to the destination */ 
00167   arm_sqrt_f32((meanOfSquares - squareOfMean), pResult); 
00168  
00169 } 
00170  
00171 /**  
00172  * @} end of STD group  
00173  */