arm_fir_lattice_f32.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_fir_lattice_f32.c    
00009 *    
00010 * Description:  Processing function for the floating-point FIR Lattice filter.    
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 groupFilters    
00045  */
00046 
00047 /**    
00048  * @defgroup FIR_Lattice Finite Impulse Response (FIR) Lattice Filters    
00049  *    
00050  * This set of functions implements Finite Impulse Response (FIR) lattice filters    
00051  * for Q15, Q31 and floating-point data types.  Lattice filters are used in a     
00052  * variety of adaptive filter applications.  The filter structure is feedforward and    
00053  * the net impulse response is finite length.    
00054  * The functions operate on blocks    
00055  * of input and output data and each call to the function processes    
00056  * <code>blockSize</code> samples through the filter.  <code>pSrc</code> and    
00057  * <code>pDst</code> point to input and output arrays containing <code>blockSize</code> values.    
00058  *    
00059  * \par Algorithm:    
00060  * \image html FIRLattice.gif "Finite Impulse Response Lattice filter"    
00061  * The following difference equation is implemented:    
00062  * <pre>    
00063  *    f0[n] = g0[n] = x[n]    
00064  *    fm[n] = fm-1[n] + km * gm-1[n-1] for m = 1, 2, ...M    
00065  *    gm[n] = km * fm-1[n] + gm-1[n-1] for m = 1, 2, ...M    
00066  *    y[n] = fM[n]    
00067  * </pre>    
00068  * \par    
00069  * <code>pCoeffs</code> points to tha array of reflection coefficients of size <code>numStages</code>.    
00070  * Reflection Coefficients are stored in the following order.    
00071  * \par    
00072  * <pre>    
00073  *    {k1, k2, ..., kM}    
00074  * </pre>    
00075  * where M is number of stages    
00076  * \par    
00077  * <code>pState</code> points to a state array of size <code>numStages</code>.    
00078  * The state variables (g values) hold previous inputs and are stored in the following order.    
00079  * <pre>    
00080  *    {g0[n], g1[n], g2[n] ...gM-1[n]}    
00081  * </pre>    
00082  * The state variables are updated after each block of data is processed; the coefficients are untouched.    
00083  * \par Instance Structure    
00084  * The coefficients and state variables for a filter are stored together in an instance data structure.    
00085  * A separate instance structure must be defined for each filter.    
00086  * Coefficient arrays may be shared among several instances while state variable arrays cannot be shared.    
00087  * There are separate instance structure declarations for each of the 3 supported data types.    
00088  *    
00089  * \par Initialization Functions    
00090  * There is also an associated initialization function for each data type.    
00091  * The initialization function performs the following operations:    
00092  * - Sets the values of the internal structure fields.    
00093  * - Zeros out the values in the state buffer.    
00094  * To do this manually without calling the init function, assign the follow subfields of the instance structure:
00095  * numStages, pCoeffs, pState. Also set all of the values in pState to zero. 
00096  *    
00097  * \par    
00098  * Use of the initialization function is optional.    
00099  * However, if the initialization function is used, then the instance structure cannot be placed into a const data section.    
00100  * To place an instance structure into a const data section, the instance structure must be manually initialized.    
00101  * Set the values in the state buffer to zeros and then manually initialize the instance structure as follows:    
00102  * <pre>    
00103  *arm_fir_lattice_instance_f32 S = {numStages, pState, pCoeffs};    
00104  *arm_fir_lattice_instance_q31 S = {numStages, pState, pCoeffs};    
00105  *arm_fir_lattice_instance_q15 S = {numStages, pState, pCoeffs};    
00106  * </pre>    
00107  * \par    
00108  * where <code>numStages</code> is the number of stages in the filter; <code>pState</code> is the address of the state buffer;    
00109  * <code>pCoeffs</code> is the address of the coefficient buffer.    
00110  * \par Fixed-Point Behavior    
00111  * Care must be taken when using the fixed-point versions of the FIR Lattice filter functions.    
00112  * In particular, the overflow and saturation behavior of the accumulator used in each function must be considered.    
00113  * Refer to the function specific documentation below for usage guidelines.    
00114  */
00115 
00116 /**    
00117  * @addtogroup FIR_Lattice    
00118  * @{    
00119  */
00120 
00121 
00122   /**    
00123    * @brief Processing function for the floating-point FIR lattice filter.    
00124    * @param[in]  *S        points to an instance of the floating-point FIR lattice structure.    
00125    * @param[in]  *pSrc     points to the block of input data.    
00126    * @param[out] *pDst     points to the block of output data    
00127    * @param[in]  blockSize number of samples to process.    
00128    * @return none.    
00129    */
00130 
00131 void arm_fir_lattice_f32(
00132   const arm_fir_lattice_instance_f32 * S,
00133   float32_t * pSrc,
00134   float32_t * pDst,
00135   uint32_t blockSize)
00136 {
00137   float32_t *pState;                             /* State pointer */
00138   float32_t *pCoeffs = S->pCoeffs;               /* Coefficient pointer */
00139   float32_t *px;                                 /* temporary state pointer */
00140   float32_t *pk;                                 /* temporary coefficient pointer */
00141 
00142 
00143 #ifndef ARM_MATH_CM0_FAMILY
00144 
00145   /* Run the below code for Cortex-M4 and Cortex-M3 */
00146 
00147   float32_t fcurr1, fnext1, gcurr1, gnext1;      /* temporary variables for first sample in loop unrolling */
00148   float32_t fcurr2, fnext2, gnext2;              /* temporary variables for second sample in loop unrolling */
00149   float32_t fcurr3, fnext3, gnext3;              /* temporary variables for third sample in loop unrolling */
00150   float32_t fcurr4, fnext4, gnext4;              /* temporary variables for fourth sample in loop unrolling */
00151   uint32_t numStages = S->numStages;             /* Number of stages in the filter */
00152   uint32_t blkCnt, stageCnt;                     /* temporary variables for counts */
00153 
00154   gcurr1 = 0.0f;
00155   pState = &S->pState[0];
00156 
00157   blkCnt = blockSize >> 2;
00158 
00159   /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.    
00160      a second loop below computes the remaining 1 to 3 samples. */
00161   while(blkCnt > 0u)
00162   {
00163 
00164     /* Read two samples from input buffer */
00165     /* f0(n) = x(n) */
00166     fcurr1 = *pSrc++;
00167     fcurr2 = *pSrc++;
00168 
00169     /* Initialize coeff pointer */
00170     pk = (pCoeffs);
00171 
00172     /* Initialize state pointer */
00173     px = pState;
00174 
00175     /* Read g0(n-1) from state */
00176     gcurr1 = *px;
00177 
00178     /* Process first sample for first tap */
00179     /* f1(n) = f0(n) +  K1 * g0(n-1) */
00180     fnext1 = fcurr1 + ((*pk) * gcurr1);
00181     /* g1(n) = f0(n) * K1  +  g0(n-1) */
00182     gnext1 = (fcurr1 * (*pk)) + gcurr1;
00183 
00184     /* Process second sample for first tap */
00185     /* for sample 2 processing */
00186     fnext2 = fcurr2 + ((*pk) * fcurr1);
00187     gnext2 = (fcurr2 * (*pk)) + fcurr1;
00188 
00189     /* Read next two samples from input buffer */
00190     /* f0(n+2) = x(n+2) */
00191     fcurr3 = *pSrc++;
00192     fcurr4 = *pSrc++;
00193 
00194     /* Copy only last input samples into the state buffer    
00195        which will be used for next four samples processing */
00196     *px++ = fcurr4;
00197 
00198     /* Process third sample for first tap */
00199     fnext3 = fcurr3 + ((*pk) * fcurr2);
00200     gnext3 = (fcurr3 * (*pk)) + fcurr2;
00201 
00202     /* Process fourth sample for first tap */
00203     fnext4 = fcurr4 + ((*pk) * fcurr3);
00204     gnext4 = (fcurr4 * (*pk++)) + fcurr3;
00205 
00206     /* Update of f values for next coefficient set processing */
00207     fcurr1 = fnext1;
00208     fcurr2 = fnext2;
00209     fcurr3 = fnext3;
00210     fcurr4 = fnext4;
00211 
00212     /* Loop unrolling.  Process 4 taps at a time . */
00213     stageCnt = (numStages - 1u) >> 2u;
00214 
00215     /* Loop over the number of taps.  Unroll by a factor of 4.    
00216      ** Repeat until we've computed numStages-3 coefficients. */
00217 
00218     /* Process 2nd, 3rd, 4th and 5th taps ... here */
00219     while(stageCnt > 0u)
00220     {
00221       /* Read g1(n-1), g3(n-1) .... from state */
00222       gcurr1 = *px;
00223 
00224       /* save g1(n) in state buffer */
00225       *px++ = gnext4;
00226 
00227       /* Process first sample for 2nd, 6th .. tap */
00228       /* Sample processing for K2, K6.... */
00229       /* f2(n) = f1(n) +  K2 * g1(n-1) */
00230       fnext1 = fcurr1 + ((*pk) * gcurr1);
00231       /* Process second sample for 2nd, 6th .. tap */
00232       /* for sample 2 processing */
00233       fnext2 = fcurr2 + ((*pk) * gnext1);
00234       /* Process third sample for 2nd, 6th .. tap */
00235       fnext3 = fcurr3 + ((*pk) * gnext2);
00236       /* Process fourth sample for 2nd, 6th .. tap */
00237       fnext4 = fcurr4 + ((*pk) * gnext3);
00238 
00239       /* g2(n) = f1(n) * K2  +  g1(n-1) */
00240       /* Calculation of state values for next stage */
00241       gnext4 = (fcurr4 * (*pk)) + gnext3;
00242       gnext3 = (fcurr3 * (*pk)) + gnext2;
00243       gnext2 = (fcurr2 * (*pk)) + gnext1;
00244       gnext1 = (fcurr1 * (*pk++)) + gcurr1;
00245 
00246 
00247       /* Read g2(n-1), g4(n-1) .... from state */
00248       gcurr1 = *px;
00249 
00250       /* save g2(n) in state buffer */
00251       *px++ = gnext4;
00252 
00253       /* Sample processing for K3, K7.... */
00254       /* Process first sample for 3rd, 7th .. tap */
00255       /* f3(n) = f2(n) +  K3 * g2(n-1) */
00256       fcurr1 = fnext1 + ((*pk) * gcurr1);
00257       /* Process second sample for 3rd, 7th .. tap */
00258       fcurr2 = fnext2 + ((*pk) * gnext1);
00259       /* Process third sample for 3rd, 7th .. tap */
00260       fcurr3 = fnext3 + ((*pk) * gnext2);
00261       /* Process fourth sample for 3rd, 7th .. tap */
00262       fcurr4 = fnext4 + ((*pk) * gnext3);
00263 
00264       /* Calculation of state values for next stage */
00265       /* g3(n) = f2(n) * K3  +  g2(n-1) */
00266       gnext4 = (fnext4 * (*pk)) + gnext3;
00267       gnext3 = (fnext3 * (*pk)) + gnext2;
00268       gnext2 = (fnext2 * (*pk)) + gnext1;
00269       gnext1 = (fnext1 * (*pk++)) + gcurr1;
00270 
00271 
00272       /* Read g1(n-1), g3(n-1) .... from state */
00273       gcurr1 = *px;
00274 
00275       /* save g3(n) in state buffer */
00276       *px++ = gnext4;
00277 
00278       /* Sample processing for K4, K8.... */
00279       /* Process first sample for 4th, 8th .. tap */
00280       /* f4(n) = f3(n) +  K4 * g3(n-1) */
00281       fnext1 = fcurr1 + ((*pk) * gcurr1);
00282       /* Process second sample for 4th, 8th .. tap */
00283       /* for sample 2 processing */
00284       fnext2 = fcurr2 + ((*pk) * gnext1);
00285       /* Process third sample for 4th, 8th .. tap */
00286       fnext3 = fcurr3 + ((*pk) * gnext2);
00287       /* Process fourth sample for 4th, 8th .. tap */
00288       fnext4 = fcurr4 + ((*pk) * gnext3);
00289 
00290       /* g4(n) = f3(n) * K4  +  g3(n-1) */
00291       /* Calculation of state values for next stage */
00292       gnext4 = (fcurr4 * (*pk)) + gnext3;
00293       gnext3 = (fcurr3 * (*pk)) + gnext2;
00294       gnext2 = (fcurr2 * (*pk)) + gnext1;
00295       gnext1 = (fcurr1 * (*pk++)) + gcurr1;
00296 
00297       /* Read g2(n-1), g4(n-1) .... from state */
00298       gcurr1 = *px;
00299 
00300       /* save g4(n) in state buffer */
00301       *px++ = gnext4;
00302 
00303       /* Sample processing for K5, K9.... */
00304       /* Process first sample for 5th, 9th .. tap */
00305       /* f5(n) = f4(n) +  K5 * g4(n-1) */
00306       fcurr1 = fnext1 + ((*pk) * gcurr1);
00307       /* Process second sample for 5th, 9th .. tap */
00308       fcurr2 = fnext2 + ((*pk) * gnext1);
00309       /* Process third sample for 5th, 9th .. tap */
00310       fcurr3 = fnext3 + ((*pk) * gnext2);
00311       /* Process fourth sample for 5th, 9th .. tap */
00312       fcurr4 = fnext4 + ((*pk) * gnext3);
00313 
00314       /* Calculation of state values for next stage */
00315       /* g5(n) = f4(n) * K5  +  g4(n-1) */
00316       gnext4 = (fnext4 * (*pk)) + gnext3;
00317       gnext3 = (fnext3 * (*pk)) + gnext2;
00318       gnext2 = (fnext2 * (*pk)) + gnext1;
00319       gnext1 = (fnext1 * (*pk++)) + gcurr1;
00320 
00321       stageCnt--;
00322     }
00323 
00324     /* If the (filter length -1) is not a multiple of 4, compute the remaining filter taps */
00325     stageCnt = (numStages - 1u) % 0x4u;
00326 
00327     while(stageCnt > 0u)
00328     {
00329       gcurr1 = *px;
00330 
00331       /* save g value in state buffer */
00332       *px++ = gnext4;
00333 
00334       /* Process four samples for last three taps here */
00335       fnext1 = fcurr1 + ((*pk) * gcurr1);
00336       fnext2 = fcurr2 + ((*pk) * gnext1);
00337       fnext3 = fcurr3 + ((*pk) * gnext2);
00338       fnext4 = fcurr4 + ((*pk) * gnext3);
00339 
00340       /* g1(n) = f0(n) * K1  +  g0(n-1) */
00341       gnext4 = (fcurr4 * (*pk)) + gnext3;
00342       gnext3 = (fcurr3 * (*pk)) + gnext2;
00343       gnext2 = (fcurr2 * (*pk)) + gnext1;
00344       gnext1 = (fcurr1 * (*pk++)) + gcurr1;
00345 
00346       /* Update of f values for next coefficient set processing */
00347       fcurr1 = fnext1;
00348       fcurr2 = fnext2;
00349       fcurr3 = fnext3;
00350       fcurr4 = fnext4;
00351 
00352       stageCnt--;
00353 
00354     }
00355 
00356     /* The results in the 4 accumulators, store in the destination buffer. */
00357     /* y(n) = fN(n) */
00358     *pDst++ = fcurr1;
00359     *pDst++ = fcurr2;
00360     *pDst++ = fcurr3;
00361     *pDst++ = fcurr4;
00362 
00363     blkCnt--;
00364   }
00365 
00366   /* If the blockSize is not a multiple of 4, compute any remaining output samples here.    
00367    ** No loop unrolling is used. */
00368   blkCnt = blockSize % 0x4u;
00369 
00370   while(blkCnt > 0u)
00371   {
00372     /* f0(n) = x(n) */
00373     fcurr1 = *pSrc++;
00374 
00375     /* Initialize coeff pointer */
00376     pk = (pCoeffs);
00377 
00378     /* Initialize state pointer */
00379     px = pState;
00380 
00381     /* read g2(n) from state buffer */
00382     gcurr1 = *px;
00383 
00384     /* for sample 1 processing */
00385     /* f1(n) = f0(n) +  K1 * g0(n-1) */
00386     fnext1 = fcurr1 + ((*pk) * gcurr1);
00387     /* g1(n) = f0(n) * K1  +  g0(n-1) */
00388     gnext1 = (fcurr1 * (*pk++)) + gcurr1;
00389 
00390     /* save g1(n) in state buffer */
00391     *px++ = fcurr1;
00392 
00393     /* f1(n) is saved in fcurr1    
00394        for next stage processing */
00395     fcurr1 = fnext1;
00396 
00397     stageCnt = (numStages - 1u);
00398 
00399     /* stage loop */
00400     while(stageCnt > 0u)
00401     {
00402       /* read g2(n) from state buffer */
00403       gcurr1 = *px;
00404 
00405       /* save g1(n) in state buffer */
00406       *px++ = gnext1;
00407 
00408       /* Sample processing for K2, K3.... */
00409       /* f2(n) = f1(n) +  K2 * g1(n-1) */
00410       fnext1 = fcurr1 + ((*pk) * gcurr1);
00411       /* g2(n) = f1(n) * K2  +  g1(n-1) */
00412       gnext1 = (fcurr1 * (*pk++)) + gcurr1;
00413 
00414       /* f1(n) is saved in fcurr1    
00415          for next stage processing */
00416       fcurr1 = fnext1;
00417 
00418       stageCnt--;
00419 
00420     }
00421 
00422     /* y(n) = fN(n) */
00423     *pDst++ = fcurr1;
00424 
00425     blkCnt--;
00426 
00427   }
00428 
00429 #else
00430 
00431   /* Run the below code for Cortex-M0 */
00432 
00433   float32_t fcurr, fnext, gcurr, gnext;          /* temporary variables */
00434   uint32_t numStages = S->numStages;             /* Length of the filter */
00435   uint32_t blkCnt, stageCnt;                     /* temporary variables for counts */
00436 
00437   pState = &S->pState[0];
00438 
00439   blkCnt = blockSize;
00440 
00441   while(blkCnt > 0u)
00442   {
00443     /* f0(n) = x(n) */
00444     fcurr = *pSrc++;
00445 
00446     /* Initialize coeff pointer */
00447     pk = pCoeffs;
00448 
00449     /* Initialize state pointer */
00450     px = pState;
00451 
00452     /* read g0(n-1) from state buffer */
00453     gcurr = *px;
00454 
00455     /* for sample 1 processing */
00456     /* f1(n) = f0(n) +  K1 * g0(n-1) */
00457     fnext = fcurr + ((*pk) * gcurr);
00458     /* g1(n) = f0(n) * K1  +  g0(n-1) */
00459     gnext = (fcurr * (*pk++)) + gcurr;
00460 
00461     /* save f0(n) in state buffer */
00462     *px++ = fcurr;
00463 
00464     /* f1(n) is saved in fcurr            
00465        for next stage processing */
00466     fcurr = fnext;
00467 
00468     stageCnt = (numStages - 1u);
00469 
00470     /* stage loop */
00471     while(stageCnt > 0u)
00472     {
00473       /* read g2(n) from state buffer */
00474       gcurr = *px;
00475 
00476       /* save g1(n) in state buffer */
00477       *px++ = gnext;
00478 
00479       /* Sample processing for K2, K3.... */
00480       /* f2(n) = f1(n) +  K2 * g1(n-1) */
00481       fnext = fcurr + ((*pk) * gcurr);
00482       /* g2(n) = f1(n) * K2  +  g1(n-1) */
00483       gnext = (fcurr * (*pk++)) + gcurr;
00484 
00485       /* f1(n) is saved in fcurr1            
00486          for next stage processing */
00487       fcurr = fnext;
00488 
00489       stageCnt--;
00490 
00491     }
00492 
00493     /* y(n) = fN(n) */
00494     *pDst++ = fcurr;
00495 
00496     blkCnt--;
00497 
00498   }
00499 
00500 #endif /*   #ifndef ARM_MATH_CM0_FAMILY */
00501 
00502 }
00503 
00504 /**    
00505  * @} end of FIR_Lattice group    
00506  */