audioconvert.c 11 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247
  1. /*
  2. * audio conversion
  3. * Copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at>
  4. *
  5. * This file is part of FFmpeg.
  6. *
  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.
  11. *
  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.
  16. *
  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
  20. */
  21. /**
  22. * @file
  23. * audio conversion
  24. * @author Michael Niedermayer <michaelni@gmx.at>
  25. */
  26. #include "libavutil/avstring.h"
  27. #include "libavutil/avassert.h"
  28. #include "libavutil/libm.h"
  29. #include "libavutil/samplefmt.h"
  30. #include "audioconvert.h"
  31. #define CONV_FUNC_NAME(dst_fmt, src_fmt) conv_ ## src_fmt ## _to_ ## dst_fmt
  32. //FIXME rounding ?
  33. #define CONV_FUNC(ofmt, otype, ifmt, expr)\
  34. static void CONV_FUNC_NAME(ofmt, ifmt)(uint8_t *po, const uint8_t *pi, int is, int os, uint8_t *end)\
  35. {\
  36. uint8_t *end2 = end - 3*os;\
  37. while(po < end2){\
  38. *(otype*)po = expr; pi += is; po += os;\
  39. *(otype*)po = expr; pi += is; po += os;\
  40. *(otype*)po = expr; pi += is; po += os;\
  41. *(otype*)po = expr; pi += is; po += os;\
  42. }\
  43. while(po < end){\
  44. *(otype*)po = expr; pi += is; po += os;\
  45. }\
  46. }
  47. //FIXME put things below under ifdefs so we do not waste space for cases no codec will need
  48. CONV_FUNC(AV_SAMPLE_FMT_U8 , uint8_t, AV_SAMPLE_FMT_U8 , *(const uint8_t*)pi)
  49. CONV_FUNC(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_U8 , (*(const uint8_t*)pi - 0x80U)<<8)
  50. CONV_FUNC(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_U8 , (*(const uint8_t*)pi - 0x80U)<<24)
  51. CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_U8 , (uint64_t)((*(const uint8_t*)pi - 0x80U))<<56)
  52. CONV_FUNC(AV_SAMPLE_FMT_FLT, float , AV_SAMPLE_FMT_U8 , (*(const uint8_t*)pi - 0x80)*(1.0f/ (1<<7)))
  53. CONV_FUNC(AV_SAMPLE_FMT_DBL, double , AV_SAMPLE_FMT_U8 , (*(const uint8_t*)pi - 0x80)*(1.0 / (1<<7)))
  54. CONV_FUNC(AV_SAMPLE_FMT_U8 , uint8_t, AV_SAMPLE_FMT_S16, (*(const int16_t*)pi>>8) + 0x80)
  55. CONV_FUNC(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_S16, *(const int16_t*)pi)
  56. CONV_FUNC(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_S16, *(const int16_t*)pi<<16)
  57. CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_S16, (uint64_t)(*(const int16_t*)pi)<<48)
  58. CONV_FUNC(AV_SAMPLE_FMT_FLT, float , AV_SAMPLE_FMT_S16, *(const int16_t*)pi*(1.0f/ (1<<15)))
  59. CONV_FUNC(AV_SAMPLE_FMT_DBL, double , AV_SAMPLE_FMT_S16, *(const int16_t*)pi*(1.0 / (1<<15)))
  60. CONV_FUNC(AV_SAMPLE_FMT_U8 , uint8_t, AV_SAMPLE_FMT_S32, (*(const int32_t*)pi>>24) + 0x80)
  61. CONV_FUNC(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_S32, *(const int32_t*)pi>>16)
  62. CONV_FUNC(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_S32, *(const int32_t*)pi)
  63. CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_S32, (uint64_t)(*(const int32_t*)pi)<<32)
  64. CONV_FUNC(AV_SAMPLE_FMT_FLT, float , AV_SAMPLE_FMT_S32, *(const int32_t*)pi*(1.0f/ (1U<<31)))
  65. CONV_FUNC(AV_SAMPLE_FMT_DBL, double , AV_SAMPLE_FMT_S32, *(const int32_t*)pi*(1.0 / (1U<<31)))
  66. CONV_FUNC(AV_SAMPLE_FMT_U8 , uint8_t, AV_SAMPLE_FMT_S64, (*(const int64_t*)pi>>56) + 0x80)
  67. CONV_FUNC(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_S64, *(const int64_t*)pi>>48)
  68. CONV_FUNC(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_S64, *(const int64_t*)pi>>32)
  69. CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_S64, *(const int64_t*)pi)
  70. CONV_FUNC(AV_SAMPLE_FMT_FLT, float , AV_SAMPLE_FMT_S64, *(const int64_t*)pi*(1.0f/ (INT64_C(1)<<63)))
  71. CONV_FUNC(AV_SAMPLE_FMT_DBL, double , AV_SAMPLE_FMT_S64, *(const int64_t*)pi*(1.0 / (INT64_C(1)<<63)))
  72. CONV_FUNC(AV_SAMPLE_FMT_U8 , uint8_t, AV_SAMPLE_FMT_FLT, av_clip_uint8( lrintf(*(const float*)pi * (1<<7)) + 0x80))
  73. CONV_FUNC(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_FLT, av_clip_int16( lrintf(*(const float*)pi * (1<<15))))
  74. CONV_FUNC(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_FLT, av_clipl_int32(llrintf(*(const float*)pi * (1U<<31))))
  75. CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_FLT, llrintf(*(const float*)pi * (INT64_C(1)<<63)))
  76. CONV_FUNC(AV_SAMPLE_FMT_FLT, float , AV_SAMPLE_FMT_FLT, *(const float*)pi)
  77. CONV_FUNC(AV_SAMPLE_FMT_DBL, double , AV_SAMPLE_FMT_FLT, *(const float*)pi)
  78. CONV_FUNC(AV_SAMPLE_FMT_U8 , uint8_t, AV_SAMPLE_FMT_DBL, av_clip_uint8( lrint(*(const double*)pi * (1<<7)) + 0x80))
  79. CONV_FUNC(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_DBL, av_clip_int16( lrint(*(const double*)pi * (1<<15))))
  80. CONV_FUNC(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_DBL, av_clipl_int32(llrint(*(const double*)pi * (1U<<31))))
  81. CONV_FUNC(AV_SAMPLE_FMT_S64, int64_t, AV_SAMPLE_FMT_DBL, llrint(*(const double*)pi * (INT64_C(1)<<63)))
  82. CONV_FUNC(AV_SAMPLE_FMT_FLT, float , AV_SAMPLE_FMT_DBL, *(const double*)pi)
  83. CONV_FUNC(AV_SAMPLE_FMT_DBL, double , AV_SAMPLE_FMT_DBL, *(const double*)pi)
  84. #define FMT_PAIR_FUNC(out, in) [(out) + AV_SAMPLE_FMT_NB*(in)] = CONV_FUNC_NAME(out, in)
  85. static conv_func_type * const fmt_pair_to_conv_functions[AV_SAMPLE_FMT_NB*AV_SAMPLE_FMT_NB] = {
  86. FMT_PAIR_FUNC(AV_SAMPLE_FMT_U8 , AV_SAMPLE_FMT_U8 ),
  87. FMT_PAIR_FUNC(AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_U8 ),
  88. FMT_PAIR_FUNC(AV_SAMPLE_FMT_S32, AV_SAMPLE_FMT_U8 ),
  89. FMT_PAIR_FUNC(AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_U8 ),
  90. FMT_PAIR_FUNC(AV_SAMPLE_FMT_DBL, AV_SAMPLE_FMT_U8 ),
  91. FMT_PAIR_FUNC(AV_SAMPLE_FMT_S64, AV_SAMPLE_FMT_U8 ),
  92. FMT_PAIR_FUNC(AV_SAMPLE_FMT_U8 , AV_SAMPLE_FMT_S16),
  93. FMT_PAIR_FUNC(AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_S16),
  94. FMT_PAIR_FUNC(AV_SAMPLE_FMT_S32, AV_SAMPLE_FMT_S16),
  95. FMT_PAIR_FUNC(AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_S16),
  96. FMT_PAIR_FUNC(AV_SAMPLE_FMT_DBL, AV_SAMPLE_FMT_S16),
  97. FMT_PAIR_FUNC(AV_SAMPLE_FMT_S64, AV_SAMPLE_FMT_S16),
  98. FMT_PAIR_FUNC(AV_SAMPLE_FMT_U8 , AV_SAMPLE_FMT_S32),
  99. FMT_PAIR_FUNC(AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_S32),
  100. FMT_PAIR_FUNC(AV_SAMPLE_FMT_S32, AV_SAMPLE_FMT_S32),
  101. FMT_PAIR_FUNC(AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_S32),
  102. FMT_PAIR_FUNC(AV_SAMPLE_FMT_DBL, AV_SAMPLE_FMT_S32),
  103. FMT_PAIR_FUNC(AV_SAMPLE_FMT_S64, AV_SAMPLE_FMT_S32),
  104. FMT_PAIR_FUNC(AV_SAMPLE_FMT_U8 , AV_SAMPLE_FMT_FLT),
  105. FMT_PAIR_FUNC(AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_FLT),
  106. FMT_PAIR_FUNC(AV_SAMPLE_FMT_S32, AV_SAMPLE_FMT_FLT),
  107. FMT_PAIR_FUNC(AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_FLT),
  108. FMT_PAIR_FUNC(AV_SAMPLE_FMT_DBL, AV_SAMPLE_FMT_FLT),
  109. FMT_PAIR_FUNC(AV_SAMPLE_FMT_S64, AV_SAMPLE_FMT_FLT),
  110. FMT_PAIR_FUNC(AV_SAMPLE_FMT_U8 , AV_SAMPLE_FMT_DBL),
  111. FMT_PAIR_FUNC(AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_DBL),
  112. FMT_PAIR_FUNC(AV_SAMPLE_FMT_S32, AV_SAMPLE_FMT_DBL),
  113. FMT_PAIR_FUNC(AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_DBL),
  114. FMT_PAIR_FUNC(AV_SAMPLE_FMT_DBL, AV_SAMPLE_FMT_DBL),
  115. FMT_PAIR_FUNC(AV_SAMPLE_FMT_S64, AV_SAMPLE_FMT_DBL),
  116. FMT_PAIR_FUNC(AV_SAMPLE_FMT_U8 , AV_SAMPLE_FMT_S64),
  117. FMT_PAIR_FUNC(AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_S64),
  118. FMT_PAIR_FUNC(AV_SAMPLE_FMT_S32, AV_SAMPLE_FMT_S64),
  119. FMT_PAIR_FUNC(AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_S64),
  120. FMT_PAIR_FUNC(AV_SAMPLE_FMT_DBL, AV_SAMPLE_FMT_S64),
  121. FMT_PAIR_FUNC(AV_SAMPLE_FMT_S64, AV_SAMPLE_FMT_S64),
  122. };
  123. static void cpy1(uint8_t **dst, const uint8_t **src, int len){
  124. memcpy(*dst, *src, len);
  125. }
  126. static void cpy2(uint8_t **dst, const uint8_t **src, int len){
  127. memcpy(*dst, *src, 2*len);
  128. }
  129. static void cpy4(uint8_t **dst, const uint8_t **src, int len){
  130. memcpy(*dst, *src, 4*len);
  131. }
  132. static void cpy8(uint8_t **dst, const uint8_t **src, int len){
  133. memcpy(*dst, *src, 8*len);
  134. }
  135. AudioConvert *swri_audio_convert_alloc(enum AVSampleFormat out_fmt,
  136. enum AVSampleFormat in_fmt,
  137. int channels, const int *ch_map,
  138. int flags)
  139. {
  140. AudioConvert *ctx;
  141. conv_func_type *f = fmt_pair_to_conv_functions[av_get_packed_sample_fmt(out_fmt) + AV_SAMPLE_FMT_NB*av_get_packed_sample_fmt(in_fmt)];
  142. if (!f)
  143. return NULL;
  144. ctx = av_mallocz(sizeof(*ctx));
  145. if (!ctx)
  146. return NULL;
  147. if(channels == 1){
  148. in_fmt = av_get_planar_sample_fmt( in_fmt);
  149. out_fmt = av_get_planar_sample_fmt(out_fmt);
  150. }
  151. ctx->channels = channels;
  152. ctx->conv_f = f;
  153. ctx->ch_map = ch_map;
  154. if (in_fmt == AV_SAMPLE_FMT_U8 || in_fmt == AV_SAMPLE_FMT_U8P)
  155. memset(ctx->silence, 0x80, sizeof(ctx->silence));
  156. if(out_fmt == in_fmt && !ch_map) {
  157. switch(av_get_bytes_per_sample(in_fmt)){
  158. case 1:ctx->simd_f = cpy1; break;
  159. case 2:ctx->simd_f = cpy2; break;
  160. case 4:ctx->simd_f = cpy4; break;
  161. case 8:ctx->simd_f = cpy8; break;
  162. }
  163. }
  164. if(HAVE_X86ASM && HAVE_MMX) swri_audio_convert_init_x86(ctx, out_fmt, in_fmt, channels);
  165. if(ARCH_ARM) swri_audio_convert_init_arm(ctx, out_fmt, in_fmt, channels);
  166. if(ARCH_AARCH64) swri_audio_convert_init_aarch64(ctx, out_fmt, in_fmt, channels);
  167. return ctx;
  168. }
  169. void swri_audio_convert_free(AudioConvert **ctx)
  170. {
  171. av_freep(ctx);
  172. }
  173. int swri_audio_convert(AudioConvert *ctx, AudioData *out, AudioData *in, int len)
  174. {
  175. int ch;
  176. int off=0;
  177. const int os= (out->planar ? 1 :out->ch_count) *out->bps;
  178. unsigned misaligned = 0;
  179. av_assert0(ctx->channels == out->ch_count);
  180. if (ctx->in_simd_align_mask) {
  181. int planes = in->planar ? in->ch_count : 1;
  182. unsigned m = 0;
  183. for (ch = 0; ch < planes; ch++)
  184. m |= (intptr_t)in->ch[ch];
  185. misaligned |= m & ctx->in_simd_align_mask;
  186. }
  187. if (ctx->out_simd_align_mask) {
  188. int planes = out->planar ? out->ch_count : 1;
  189. unsigned m = 0;
  190. for (ch = 0; ch < planes; ch++)
  191. m |= (intptr_t)out->ch[ch];
  192. misaligned |= m & ctx->out_simd_align_mask;
  193. }
  194. //FIXME optimize common cases
  195. if(ctx->simd_f && !ctx->ch_map && !misaligned){
  196. off = len&~15;
  197. av_assert1(off>=0);
  198. av_assert1(off<=len);
  199. av_assert2(ctx->channels == SWR_CH_MAX || !in->ch[ctx->channels]);
  200. if(off>0){
  201. if(out->planar == in->planar){
  202. int planes = out->planar ? out->ch_count : 1;
  203. for(ch=0; ch<planes; ch++){
  204. ctx->simd_f(out->ch+ch, (const uint8_t **)in->ch+ch, off * (out->planar ? 1 :out->ch_count));
  205. }
  206. }else{
  207. ctx->simd_f(out->ch, (const uint8_t **)in->ch, off);
  208. }
  209. }
  210. if(off == len)
  211. return 0;
  212. }
  213. for(ch=0; ch<ctx->channels; ch++){
  214. const int ich= ctx->ch_map ? ctx->ch_map[ch] : ch;
  215. const int is= ich < 0 ? 0 : (in->planar ? 1 : in->ch_count) * in->bps;
  216. const uint8_t *pi= ich < 0 ? ctx->silence : in->ch[ich];
  217. uint8_t *po= out->ch[ch];
  218. uint8_t *end= po + os*len;
  219. if(!po)
  220. continue;
  221. ctx->conv_f(po+off*os, pi+off*is, is, os, end);
  222. }
  223. return 0;
  224. }