3 * Copyright (c) 2004-2012 Michael Niedermayer <michaelni@gmx.at>
5 * This file is part of FFmpeg.
7 * FFmpeg is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU Lesser General Public
9 * License as published by the Free Software Foundation; either
10 * version 2.1 of the License, or (at your option) any later version.
12 * FFmpeg is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 * Lesser General Public License for more details.
17 * You should have received a copy of the GNU Lesser General Public
18 * License along with FFmpeg; if not, write to the Free Software
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
25 * @author Michael Niedermayer <michaelni@gmx.at>
28 #if defined(TEMPLATE_RESAMPLE_DBL) \
29 || defined(TEMPLATE_RESAMPLE_DBL_SSE2)
31 # define FILTER_SHIFT 0
34 # define FELEM2 double
35 # define FELEML double
36 # define OUT(d, v) d = v
38 # if defined(TEMPLATE_RESAMPLE_DBL)
39 # define RENAME(N) N ## _double
40 # elif defined(TEMPLATE_RESAMPLE_DBL_SSE2)
41 # define COMMON_CORE COMMON_CORE_DBL_SSE2
42 # define LINEAR_CORE LINEAR_CORE_DBL_SSE2
43 # define RENAME(N) N ## _double_sse2
46 #elif defined(TEMPLATE_RESAMPLE_FLT) \
47 || defined(TEMPLATE_RESAMPLE_FLT_SSE)
49 # define FILTER_SHIFT 0
54 # define OUT(d, v) d = v
56 # if defined(TEMPLATE_RESAMPLE_FLT)
57 # define RENAME(N) N ## _float
58 # elif defined(TEMPLATE_RESAMPLE_FLT_SSE)
59 # define COMMON_CORE COMMON_CORE_FLT_SSE
60 # define LINEAR_CORE LINEAR_CORE_FLT_SSE
61 # define RENAME(N) N ## _float_sse
64 #elif defined(TEMPLATE_RESAMPLE_S32)
65 # define RENAME(N) N ## _int32
66 # define FILTER_SHIFT 30
67 # define DELEM int32_t
68 # define FELEM int32_t
69 # define FELEM2 int64_t
70 # define FELEML int64_t
71 # define FELEM_MAX INT32_MAX
72 # define FELEM_MIN INT32_MIN
73 # define OUT(d, v) v = (v + (1<<(FILTER_SHIFT-1)))>>FILTER_SHIFT;\
74 d = (uint64_t)(v + 0x80000000) > 0xFFFFFFFF ? (v>>63) ^ 0x7FFFFFFF : v
76 #elif defined(TEMPLATE_RESAMPLE_S16) \
77 || defined(TEMPLATE_RESAMPLE_S16_MMX2) \
78 || defined(TEMPLATE_RESAMPLE_S16_SSE2)
80 # define FILTER_SHIFT 15
81 # define DELEM int16_t
82 # define FELEM int16_t
83 # define FELEM2 int32_t
84 # define FELEML int64_t
85 # define FELEM_MAX INT16_MAX
86 # define FELEM_MIN INT16_MIN
87 # define OUT(d, v) v = (v + (1<<(FILTER_SHIFT-1)))>>FILTER_SHIFT;\
88 d = (unsigned)(v + 32768) > 65535 ? (v>>31) ^ 32767 : v
90 # if defined(TEMPLATE_RESAMPLE_S16)
91 # define RENAME(N) N ## _int16
92 # elif defined(TEMPLATE_RESAMPLE_S16_MMX2)
93 # define COMMON_CORE COMMON_CORE_INT16_MMX2
94 # define LINEAR_CORE LINEAR_CORE_INT16_MMX2
95 # define RENAME(N) N ## _int16_mmx2
96 # elif defined(TEMPLATE_RESAMPLE_S16_SSE2)
97 # define COMMON_CORE COMMON_CORE_INT16_SSE2
98 # define LINEAR_CORE LINEAR_CORE_INT16_SSE2
99 # define RENAME(N) N ## _int16_sse2
104 int RENAME(swri_resample)(ResampleContext *c, DELEM *dst, const DELEM *src, int *consumed, int src_size, int dst_size, int update_ctx){
108 int dst_incr_frac= c->dst_incr % c->src_incr;
109 int dst_incr= c->dst_incr / c->src_incr;
110 int compensation_distance= c->compensation_distance;
112 av_assert1(c->filter_shift == FILTER_SHIFT);
113 av_assert1(c->felem_size == sizeof(FELEM));
115 if(compensation_distance == 0 && c->filter_length == 1 && c->phase_shift==0){
116 int64_t index2= (1LL<<32)*c->frac/c->src_incr + (1LL<<32)*index;
117 int64_t incr= (1LL<<32) * c->dst_incr / c->src_incr;
118 int new_size = (src_size * (int64_t)c->src_incr - frac + c->dst_incr - 1) / c->dst_incr;
120 dst_size= FFMIN(dst_size, new_size);
122 for(dst_index=0; dst_index < dst_size; dst_index++){
123 dst[dst_index] = src[index2>>32];
126 index += dst_index * dst_incr;
127 index += (frac + dst_index * (int64_t)dst_incr_frac) / c->src_incr;
128 frac = (frac + dst_index * (int64_t)dst_incr_frac) % c->src_incr;
129 av_assert2(index >= 0);
132 }else if(compensation_distance == 0 && !c->linear && index >= 0){
133 int sample_index = 0;
134 for(dst_index=0; dst_index < dst_size; dst_index++){
136 sample_index += index >> c->phase_shift;
137 index &= c->phase_mask;
138 filter= ((FELEM*)c->filter_bank) + c->filter_alloc*index;
140 if(sample_index + c->filter_length > src_size){
147 for(i=0; i<c->filter_length; i++){
148 val += src[sample_index + i] * (FELEM2)filter[i];
150 OUT(dst[dst_index], val);
154 frac += dst_incr_frac;
156 if(frac >= c->src_incr){
161 *consumed = sample_index;
163 int sample_index = 0;
164 for(dst_index=0; dst_index < dst_size; dst_index++){
168 sample_index += index >> c->phase_shift;
169 index &= c->phase_mask;
170 filter = ((FELEM*)c->filter_bank) + c->filter_alloc*index;
172 if(sample_index + c->filter_length > src_size || -sample_index >= src_size){
174 }else if(sample_index < 0){
175 for(i=0; i<c->filter_length; i++)
176 val += src[FFABS(sample_index + i)] * (FELEM2)filter[i];
177 OUT(dst[dst_index], val);
183 for(i=0; i<c->filter_length; i++){
184 val += src[sample_index + i] * (FELEM2)filter[i];
185 v2 += src[sample_index + i] * (FELEM2)filter[i + c->filter_alloc];
188 val+=(v2-val)*(FELEML)frac / c->src_incr;
189 OUT(dst[dst_index], val);
194 for(i=0; i<c->filter_length; i++){
195 val += src[sample_index + i] * (FELEM2)filter[i];
197 OUT(dst[dst_index], val);
201 frac += dst_incr_frac;
203 if(frac >= c->src_incr){
208 if(dst_index + 1 == compensation_distance){
209 compensation_distance= 0;
210 dst_incr_frac= c->ideal_dst_incr % c->src_incr;
211 dst_incr= c->ideal_dst_incr / c->src_incr;
214 *consumed= FFMAX(sample_index, 0);
215 index += FFMIN(sample_index, 0) << c->phase_shift;
217 if(compensation_distance){
218 compensation_distance -= dst_index;
219 av_assert1(compensation_distance > 0);
226 c->dst_incr= dst_incr_frac + c->src_incr*dst_incr;
227 c->compensation_distance= compensation_distance;