yading@10: /*
yading@10:  * This file is part of FFmpeg.
yading@10:  *
yading@10:  * FFmpeg is free software; you can redistribute it and/or
yading@10:  * modify it under the terms of the GNU Lesser General Public
yading@10:  * License as published by the Free Software Foundation; either
yading@10:  * version 2.1 of the License, or (at your option) any later version.
yading@10:  *
yading@10:  * FFmpeg is distributed in the hope that it will be useful,
yading@10:  * but WITHOUT ANY WARRANTY; without even the implied warranty of
yading@10:  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
yading@10:  * Lesser General Public License for more details.
yading@10:  *
yading@10:  * You should have received a copy of the GNU Lesser General Public
yading@10:  * License along with FFmpeg; if not, write to the Free Software
yading@10:  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
yading@10:  */
yading@10: 
yading@10: #include <stdint.h>
yading@10: #include "libavutil/avutil.h"
yading@10: #include "mathops.h"
yading@10: 
yading@10: #undef FUNC
yading@10: #undef sum_type
yading@10: #undef MUL
yading@10: #undef CLIP
yading@10: #undef FSUF
yading@10: 
yading@10: #define FUNC(n) AV_JOIN(n ## _, SAMPLE_SIZE)
yading@10: 
yading@10: #if SAMPLE_SIZE == 32
yading@10: #   define sum_type  int64_t
yading@10: #   define MUL(a, b) MUL64(a, b)
yading@10: #   define CLIP(x) av_clipl_int32(x)
yading@10: #else
yading@10: #   define sum_type  int32_t
yading@10: #   define MUL(a, b) ((a) * (b))
yading@10: #   define CLIP(x) (x)
yading@10: #endif
yading@10: 
yading@10: #define LPC1(x) {           \
yading@10:     int c = coefs[(x)-1];   \
yading@10:     p0   += MUL(c, s);      \
yading@10:     s     = smp[i-(x)+1];   \
yading@10:     p1   += MUL(c, s);      \
yading@10: }
yading@10: 
yading@10: static av_always_inline void FUNC(lpc_encode_unrolled)(int32_t *res,
yading@10:                                   const int32_t *smp, int len, int order,
yading@10:                                   const int32_t *coefs, int shift, int big)
yading@10: {
yading@10:     int i;
yading@10:     for (i = order; i < len; i += 2) {
yading@10:         int s  = smp[i-order];
yading@10:         sum_type p0 = 0, p1 = 0;
yading@10:         if (big) {
yading@10:             switch (order) {
yading@10:             case 32: LPC1(32)
yading@10:             case 31: LPC1(31)
yading@10:             case 30: LPC1(30)
yading@10:             case 29: LPC1(29)
yading@10:             case 28: LPC1(28)
yading@10:             case 27: LPC1(27)
yading@10:             case 26: LPC1(26)
yading@10:             case 25: LPC1(25)
yading@10:             case 24: LPC1(24)
yading@10:             case 23: LPC1(23)
yading@10:             case 22: LPC1(22)
yading@10:             case 21: LPC1(21)
yading@10:             case 20: LPC1(20)
yading@10:             case 19: LPC1(19)
yading@10:             case 18: LPC1(18)
yading@10:             case 17: LPC1(17)
yading@10:             case 16: LPC1(16)
yading@10:             case 15: LPC1(15)
yading@10:             case 14: LPC1(14)
yading@10:             case 13: LPC1(13)
yading@10:             case 12: LPC1(12)
yading@10:             case 11: LPC1(11)
yading@10:             case 10: LPC1(10)
yading@10:             case  9: LPC1( 9)
yading@10:                      LPC1( 8)
yading@10:                      LPC1( 7)
yading@10:                      LPC1( 6)
yading@10:                      LPC1( 5)
yading@10:                      LPC1( 4)
yading@10:                      LPC1( 3)
yading@10:                      LPC1( 2)
yading@10:                      LPC1( 1)
yading@10:             }
yading@10:         } else {
yading@10:             switch (order) {
yading@10:             case  8: LPC1( 8)
yading@10:             case  7: LPC1( 7)
yading@10:             case  6: LPC1( 6)
yading@10:             case  5: LPC1( 5)
yading@10:             case  4: LPC1( 4)
yading@10:             case  3: LPC1( 3)
yading@10:             case  2: LPC1( 2)
yading@10:             case  1: LPC1( 1)
yading@10:             }
yading@10:         }
yading@10:         res[i  ] = smp[i  ] - CLIP(p0 >> shift);
yading@10:         res[i+1] = smp[i+1] - CLIP(p1 >> shift);
yading@10:     }
yading@10: }
yading@10: 
yading@10: static void FUNC(flac_lpc_encode_c)(int32_t *res, const int32_t *smp, int len,
yading@10:                                     int order, const int32_t *coefs, int shift)
yading@10: {
yading@10:     int i;
yading@10:     for (i = 0; i < order; i++)
yading@10:         res[i] = smp[i];
yading@10: #if CONFIG_SMALL
yading@10:     for (i = order; i < len; i += 2) {
yading@10:         int j;
yading@10:         int s  = smp[i];
yading@10:         sum_type p0 = 0, p1 = 0;
yading@10:         for (j = 0; j < order; j++) {
yading@10:             int c = coefs[j];
yading@10:             p1   += MUL(c, s);
yading@10:             s     = smp[i-j-1];
yading@10:             p0   += MUL(c, s);
yading@10:         }
yading@10:         res[i  ] = smp[i  ] - CLIP(p0 >> shift);
yading@10:         res[i+1] = smp[i+1] - CLIP(p1 >> shift);
yading@10:     }
yading@10: #else
yading@10:     switch (order) {
yading@10:     case  1: FUNC(lpc_encode_unrolled)(res, smp, len,     1, coefs, shift, 0); break;
yading@10:     case  2: FUNC(lpc_encode_unrolled)(res, smp, len,     2, coefs, shift, 0); break;
yading@10:     case  3: FUNC(lpc_encode_unrolled)(res, smp, len,     3, coefs, shift, 0); break;
yading@10:     case  4: FUNC(lpc_encode_unrolled)(res, smp, len,     4, coefs, shift, 0); break;
yading@10:     case  5: FUNC(lpc_encode_unrolled)(res, smp, len,     5, coefs, shift, 0); break;
yading@10:     case  6: FUNC(lpc_encode_unrolled)(res, smp, len,     6, coefs, shift, 0); break;
yading@10:     case  7: FUNC(lpc_encode_unrolled)(res, smp, len,     7, coefs, shift, 0); break;
yading@10:     case  8: FUNC(lpc_encode_unrolled)(res, smp, len,     8, coefs, shift, 0); break;
yading@10:     default: FUNC(lpc_encode_unrolled)(res, smp, len, order, coefs, shift, 1); break;
yading@10:     }
yading@10: #endif
yading@10: }