arm_rms_q31.c

00001 /* ----------------------------------------------------------------------    
00002 * Copyright (C) 2010-2013 ARM Limited. All rights reserved.    
00003 *    
00004 * $Date:        17. January 2013
00005 * $Revision:    V1.4.1  
00006 *    
00007 * Project:      CMSIS DSP Library    
00008 * Title:        arm_rms_q31.c    
00009 *    
00010 * Description:  Root Mean Square of the elements of a Q31 vector.    
00011 *    
00012 * Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
00013 *  
00014 * Redistribution and use in source and binary forms, with or without 
00015 * modification, are permitted provided that the following conditions
00016 * are met:
00017 *   - Redistributions of source code must retain the above copyright
00018 *     notice, this list of conditions and the following disclaimer.
00019 *   - Redistributions in binary form must reproduce the above copyright
00020 *     notice, this list of conditions and the following disclaimer in
00021 *     the documentation and/or other materials provided with the 
00022 *     distribution.
00023 *   - Neither the name of ARM LIMITED nor the names of its contributors
00024 *     may be used to endorse or promote products derived from this
00025 *     software without specific prior written permission.
00026 *
00027 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
00028 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
00029 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
00030 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE 
00031 * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
00032 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
00033 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
00034 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
00035 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
00036 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
00037 * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
00038 * POSSIBILITY OF SUCH DAMAGE.    
00039 * ---------------------------------------------------------------------------- */
00040 
00041 #include "arm_math.h"
00042 
00043 /**        
00044  * @addtogroup RMS        
00045  * @{        
00046  */
00047 
00048 
00049 /**        
00050  * @brief Root Mean Square of the elements of a Q31 vector.        
00051  * @param[in]       *pSrc points to the input vector        
00052  * @param[in]       blockSize length of the input vector        
00053  * @param[out]      *pResult rms value returned here        
00054  * @return none.        
00055  *        
00056  * @details        
00057  * <b>Scaling and Overflow Behavior:</b>        
00058  *        
00059  *\par        
00060  * The function is implemented using an internal 64-bit accumulator.        
00061  * The input is represented in 1.31 format, and intermediate multiplication        
00062  * yields a 2.62 format.        
00063  * The accumulator maintains full precision of the intermediate multiplication results,         
00064  * but provides only a single guard bit.        
00065  * There is no saturation on intermediate additions.        
00066  * If the accumulator overflows, it wraps around and distorts the result.         
00067  * In order to avoid overflows completely, the input signal must be scaled down by         
00068  * log2(blockSize) bits, as a total of blockSize additions are performed internally.         
00069  * Finally, the 2.62 accumulator is right shifted by 31 bits to yield a 1.31 format value.        
00070  *        
00071  */
00072 
00073 void arm_rms_q31(
00074   q31_t * pSrc,
00075   uint32_t blockSize,
00076   q31_t * pResult)
00077 {
00078   q63_t sum = 0;                                 /* accumulator */
00079   q31_t in;                                      /* Temporary variable to store the input */
00080   uint32_t blkCnt;                               /* loop counter */
00081 
00082 #ifndef ARM_MATH_CM0_FAMILY
00083 
00084   /* Run the below code for Cortex-M4 and Cortex-M3 */
00085 
00086   q31_t in1, in2, in3, in4;                      /* Temporary input variables */
00087 
00088   /*loop Unrolling */
00089   blkCnt = blockSize >> 2u;
00090 
00091   /* First part of the processing with loop unrolling.  Compute 8 outputs at a time.        
00092    ** a second loop below computes the remaining 1 to 7 samples. */
00093   while(blkCnt > 0u)
00094   {
00095     /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */
00096     /* Compute sum of the squares and then store the result in a temporary variable, sum */
00097     /* read two samples from source buffer */
00098     in1 = pSrc[0];
00099     in2 = pSrc[1];
00100 
00101     /* calculate power and accumulate to accumulator */
00102     sum += (q63_t) in1 *in1;
00103     sum += (q63_t) in2 *in2;
00104 
00105     /* read two samples from source buffer */
00106     in3 = pSrc[2];
00107     in4 = pSrc[3];
00108 
00109     /* calculate power and accumulate to accumulator */
00110     sum += (q63_t) in3 *in3;
00111     sum += (q63_t) in4 *in4;
00112 
00113 
00114     /* update source buffer to process next samples */
00115     pSrc += 4u;
00116 
00117     /* Decrement the loop counter */
00118     blkCnt--;
00119   }
00120 
00121   /* If the blockSize is not a multiple of 8, compute any remaining output samples here.        
00122    ** No loop unrolling is used. */
00123   blkCnt = blockSize % 0x4u;
00124 
00125 #else
00126 
00127   /* Run the below code for Cortex-M0 */
00128   blkCnt = blockSize;
00129 
00130 #endif /* #ifndef ARM_MATH_CM0_FAMILY */
00131 
00132   while(blkCnt > 0u)
00133   {
00134     /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */
00135     /* Compute sum of the squares and then store the results in a temporary variable, sum */
00136     in = *pSrc++;
00137     sum += (q63_t) in *in;
00138 
00139     /* Decrement the loop counter */
00140     blkCnt--;
00141   }
00142 
00143   /* Convert data in 2.62 to 1.31 by 31 right shifts and saturate */
00144 
00145   sum = __SSAT(sum >> 31, 31);
00146 
00147 
00148   /* Compute Rms and store the result in the destination vector */
00149   arm_sqrt_q31((q31_t) ((q31_t) sum / (int32_t) blockSize), pResult);
00150 }
00151 
00152 /**        
00153  * @} end of RMS group        
00154  */