arm_cmplx_mag_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_cmplx_mag_q31.c    
00009 *    
00010 * Description:  Q31 complex magnitude    
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  * @ingroup groupCmplxMath        
00045  */
00046 
00047 /**        
00048  * @addtogroup cmplx_mag        
00049  * @{        
00050  */
00051 
00052 /**        
00053  * @brief  Q31 complex magnitude        
00054  * @param  *pSrc points to the complex input vector        
00055  * @param  *pDst points to the real output vector        
00056  * @param  numSamples number of complex samples in the input vector        
00057  * @return none.        
00058  *        
00059  * <b>Scaling and Overflow Behavior:</b>        
00060  * \par        
00061  * The function implements 1.31 by 1.31 multiplications and finally output is converted into 2.30 format.        
00062  * Input down scaling is not required.        
00063  */
00064 
00065 void arm_cmplx_mag_q31(
00066   q31_t * pSrc,
00067   q31_t * pDst,
00068   uint32_t numSamples)
00069 {
00070   q31_t real, imag;                              /* Temporary variables to hold input values */
00071   q31_t acc0, acc1;                              /* Accumulators */
00072   uint32_t blkCnt;                               /* loop counter */
00073 
00074 #ifndef ARM_MATH_CM0_FAMILY
00075 
00076   /* Run the below code for Cortex-M4 and Cortex-M3 */
00077   q31_t real1, real2, imag1, imag2;              /* Temporary variables to hold input values */
00078   q31_t out1, out2, out3, out4;                  /* Accumulators */
00079   q63_t mul1, mul2, mul3, mul4;                  /* Temporary variables */
00080 
00081 
00082   /*loop Unrolling */
00083   blkCnt = numSamples >> 2u;
00084 
00085   /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.        
00086    ** a second loop below computes the remaining 1 to 3 samples. */
00087   while(blkCnt > 0u)
00088   {
00089     /* read complex input from source buffer */
00090     real1 = pSrc[0];
00091     imag1 = pSrc[1];
00092     real2 = pSrc[2];
00093     imag2 = pSrc[3];
00094 
00095     /* calculate power of input values */
00096     mul1 = (q63_t) real1 *real1;
00097     mul2 = (q63_t) imag1 *imag1;
00098     mul3 = (q63_t) real2 *real2;
00099     mul4 = (q63_t) imag2 *imag2;
00100 
00101     /* get the result to 3.29 format */
00102     out1 = (q31_t) (mul1 >> 33);
00103     out2 = (q31_t) (mul2 >> 33);
00104     out3 = (q31_t) (mul3 >> 33);
00105     out4 = (q31_t) (mul4 >> 33);
00106 
00107     /* add real and imaginary accumulators */
00108     out1 = out1 + out2;
00109     out3 = out3 + out4;
00110 
00111     /* read complex input from source buffer */
00112     real1 = pSrc[4];
00113     imag1 = pSrc[5];
00114     real2 = pSrc[6];
00115     imag2 = pSrc[7];
00116 
00117     /* calculate square root */
00118     arm_sqrt_q31(out1, &pDst[0]);
00119 
00120     /* calculate power of input values */
00121     mul1 = (q63_t) real1 *real1;
00122 
00123     /* calculate square root */
00124     arm_sqrt_q31(out3, &pDst[1]);
00125 
00126     /* calculate power of input values */
00127     mul2 = (q63_t) imag1 *imag1;
00128     mul3 = (q63_t) real2 *real2;
00129     mul4 = (q63_t) imag2 *imag2;
00130 
00131     /* get the result to 3.29 format */
00132     out1 = (q31_t) (mul1 >> 33);
00133     out2 = (q31_t) (mul2 >> 33);
00134     out3 = (q31_t) (mul3 >> 33);
00135     out4 = (q31_t) (mul4 >> 33);
00136 
00137     /* add real and imaginary accumulators */
00138     out1 = out1 + out2;
00139     out3 = out3 + out4;
00140 
00141     /* calculate square root */
00142     arm_sqrt_q31(out1, &pDst[2]);
00143 
00144     /* increment destination by 8 to process next samples */
00145     pSrc += 8u;
00146 
00147     /* calculate square root */
00148     arm_sqrt_q31(out3, &pDst[3]);
00149 
00150     /* increment destination by 4 to process next samples */
00151     pDst += 4u;
00152 
00153     /* Decrement the loop counter */
00154     blkCnt--;
00155   }
00156 
00157   /* If the numSamples is not a multiple of 4, compute any remaining output samples here.        
00158    ** No loop unrolling is used. */
00159   blkCnt = numSamples % 0x4u;
00160 
00161 #else
00162 
00163   /* Run the below code for Cortex-M0 */
00164   blkCnt = numSamples;
00165 
00166 #endif /* #ifndef ARM_MATH_CM0_FAMILY */
00167 
00168   while(blkCnt > 0u)
00169   {
00170     /* C[0] = sqrt(A[0] * A[0] + A[1] * A[1]) */
00171     real = *pSrc++;
00172     imag = *pSrc++;
00173     acc0 = (q31_t) (((q63_t) real * real) >> 33);
00174     acc1 = (q31_t) (((q63_t) imag * imag) >> 33);
00175     /* store the result in 2.30 format in the destination buffer. */
00176     arm_sqrt_q31(acc0 + acc1, pDst++);
00177 
00178     /* Decrement the loop counter */
00179     blkCnt--;
00180   }
00181 }
00182 
00183 /**        
00184  * @} end of cmplx_mag group        
00185  */