swresample_internal.h 14 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227
  1. /*
  2. * Copyright (C) 2011-2013 Michael Niedermayer (michaelni@gmx.at)
  3. *
  4. * This file is part of libswresample
  5. *
  6. * libswresample is free software; you can redistribute it and/or
  7. * modify it under the terms of the GNU Lesser General Public
  8. * License as published by the Free Software Foundation; either
  9. * version 2.1 of the License, or (at your option) any later version.
  10. *
  11. * libswresample is distributed in the hope that it will be useful,
  12. * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13. * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
  14. * Lesser General Public License for more details.
  15. *
  16. * You should have received a copy of the GNU Lesser General Public
  17. * License along with libswresample; if not, write to the Free Software
  18. * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  19. */
  20. #ifndef SWRESAMPLE_SWRESAMPLE_INTERNAL_H
  21. #define SWRESAMPLE_SWRESAMPLE_INTERNAL_H
  22. #include "swresample.h"
  23. #include "libavutil/channel_layout.h"
  24. #include "config.h"
  25. #define SWR_CH_MAX 64
  26. #define SQRT3_2 1.22474487139158904909 /* sqrt(3/2) */
  27. #define NS_TAPS 20
  28. #if ARCH_X86_64
  29. typedef int64_t integer;
  30. #else
  31. typedef int integer;
  32. #endif
  33. typedef void (mix_1_1_func_type)(void *out, const void *in, void *coeffp, integer index, integer len);
  34. typedef void (mix_2_1_func_type)(void *out, const void *in1, const void *in2, void *coeffp, integer index1, integer index2, integer len);
  35. typedef void (mix_any_func_type)(uint8_t **out, const uint8_t **in1, void *coeffp, integer len);
  36. typedef struct AudioData{
  37. uint8_t *ch[SWR_CH_MAX]; ///< samples buffer per channel
  38. uint8_t *data; ///< samples buffer
  39. int ch_count; ///< number of channels
  40. int bps; ///< bytes per sample
  41. int count; ///< number of samples
  42. int planar; ///< 1 if planar audio, 0 otherwise
  43. enum AVSampleFormat fmt; ///< sample format
  44. } AudioData;
  45. struct DitherContext {
  46. int method;
  47. int noise_pos;
  48. float scale;
  49. float noise_scale; ///< Noise scale
  50. int ns_taps; ///< Noise shaping dither taps
  51. float ns_scale; ///< Noise shaping dither scale
  52. float ns_scale_1; ///< Noise shaping dither scale^-1
  53. int ns_pos; ///< Noise shaping dither position
  54. float ns_coeffs[NS_TAPS]; ///< Noise shaping filter coefficients
  55. float ns_errors[SWR_CH_MAX][2*NS_TAPS];
  56. AudioData noise; ///< noise used for dithering
  57. AudioData temp; ///< temporary storage when writing into the input buffer isn't possible
  58. int output_sample_bits; ///< the number of used output bits, needed to scale dither correctly
  59. };
  60. typedef struct ResampleContext * (* resample_init_func)(struct ResampleContext *c, int out_rate, int in_rate, int filter_size, int phase_shift, int linear,
  61. double cutoff, enum AVSampleFormat format, enum SwrFilterType filter_type, double kaiser_beta, double precision, int cheby, int exact_rational);
  62. typedef void (* resample_free_func)(struct ResampleContext **c);
  63. typedef int (* multiple_resample_func)(struct ResampleContext *c, AudioData *dst, int dst_size, AudioData *src, int src_size, int *consumed);
  64. typedef int (* resample_flush_func)(struct SwrContext *c);
  65. typedef int (* set_compensation_func)(struct ResampleContext *c, int sample_delta, int compensation_distance);
  66. typedef int64_t (* get_delay_func)(struct SwrContext *s, int64_t base);
  67. typedef int (* invert_initial_buffer_func)(struct ResampleContext *c, AudioData *dst, const AudioData *src, int src_size, int *dst_idx, int *dst_count);
  68. typedef int64_t (* get_out_samples_func)(struct SwrContext *s, int in_samples);
  69. struct Resampler {
  70. resample_init_func init;
  71. resample_free_func free;
  72. multiple_resample_func multiple_resample;
  73. resample_flush_func flush;
  74. set_compensation_func set_compensation;
  75. get_delay_func get_delay;
  76. invert_initial_buffer_func invert_initial_buffer;
  77. get_out_samples_func get_out_samples;
  78. };
  79. extern struct Resampler const swri_resampler;
  80. extern struct Resampler const swri_soxr_resampler;
  81. struct SwrContext {
  82. const AVClass *av_class; ///< AVClass used for AVOption and av_log()
  83. int log_level_offset; ///< logging level offset
  84. void *log_ctx; ///< parent logging context
  85. enum AVSampleFormat in_sample_fmt; ///< input sample format
  86. enum AVSampleFormat int_sample_fmt; ///< internal sample format (AV_SAMPLE_FMT_FLTP or AV_SAMPLE_FMT_S16P)
  87. enum AVSampleFormat out_sample_fmt; ///< output sample format
  88. AVChannelLayout used_ch_layout; ///< number of used input channels (mapped channel count if channel_map, otherwise in.ch_count)
  89. AVChannelLayout in_ch_layout; ///< input channel layout
  90. AVChannelLayout out_ch_layout; ///< output channel layout
  91. int in_sample_rate; ///< input sample rate
  92. int out_sample_rate; ///< output sample rate
  93. int flags; ///< miscellaneous flags such as SWR_FLAG_RESAMPLE
  94. float slev; ///< surround mixing level
  95. float clev; ///< center mixing level
  96. float lfe_mix_level; ///< LFE mixing level
  97. float rematrix_volume; ///< rematrixing volume coefficient
  98. float rematrix_maxval; ///< maximum value for rematrixing output
  99. int matrix_encoding; /**< matrixed stereo encoding */
  100. const int *channel_map; ///< channel index (or -1 if muted channel) map
  101. int engine;
  102. #if FF_API_OLD_CHANNEL_LAYOUT
  103. int user_used_ch_count; ///< User set used channel count
  104. int user_in_ch_count; ///< User set input channel count
  105. int user_out_ch_count; ///< User set output channel count
  106. int64_t user_in_ch_layout; ///< User set input channel layout
  107. int64_t user_out_ch_layout; ///< User set output channel layout
  108. #endif
  109. AVChannelLayout user_used_chlayout; ///< User set used channel layout
  110. AVChannelLayout user_in_chlayout; ///< User set input channel layout
  111. AVChannelLayout user_out_chlayout; ///< User set output channel layout
  112. enum AVSampleFormat user_int_sample_fmt; ///< User set internal sample format
  113. int user_dither_method; ///< User set dither method
  114. struct DitherContext dither;
  115. int filter_size; /**< length of each FIR filter in the resampling filterbank relative to the cutoff frequency */
  116. int phase_shift; /**< log2 of the number of entries in the resampling polyphase filterbank */
  117. int linear_interp; /**< if 1 then the resampling FIR filter will be linearly interpolated */
  118. int exact_rational; /**< if 1 then enable non power of 2 phase_count */
  119. double cutoff; /**< resampling cutoff frequency (swr: 6dB point; soxr: 0dB point). 1.0 corresponds to half the output sample rate */
  120. int filter_type; /**< swr resampling filter type */
  121. double kaiser_beta; /**< swr beta value for Kaiser window (only applicable if filter_type == AV_FILTER_TYPE_KAISER) */
  122. double precision; /**< soxr resampling precision (in bits) */
  123. int cheby; /**< soxr: if 1 then passband rolloff will be none (Chebyshev) & irrational ratio approximation precision will be higher */
  124. float min_compensation; ///< swr minimum below which no compensation will happen
  125. float min_hard_compensation; ///< swr minimum below which no silence inject / sample drop will happen
  126. float soft_compensation_duration; ///< swr duration over which soft compensation is applied
  127. float max_soft_compensation; ///< swr maximum soft compensation in seconds over soft_compensation_duration
  128. float async; ///< swr simple 1 parameter async, similar to ffmpegs -async
  129. int64_t firstpts_in_samples; ///< swr first pts in samples
  130. int resample_first; ///< 1 if resampling must come first, 0 if rematrixing
  131. int rematrix; ///< flag to indicate if rematrixing is needed (basically if input and output layouts mismatch)
  132. int rematrix_custom; ///< flag to indicate that a custom matrix has been defined
  133. AudioData in; ///< input audio data
  134. AudioData postin; ///< post-input audio data: used for rematrix/resample
  135. AudioData midbuf; ///< intermediate audio data (postin/preout)
  136. AudioData preout; ///< pre-output audio data: used for rematrix/resample
  137. AudioData out; ///< converted output audio data
  138. AudioData in_buffer; ///< cached audio data (convert and resample purpose)
  139. AudioData silence; ///< temporary with silence
  140. AudioData drop_temp; ///< temporary used to discard output
  141. int in_buffer_index; ///< cached buffer position
  142. int in_buffer_count; ///< cached buffer length
  143. int resample_in_constraint; ///< 1 if the input end was reach before the output end, 0 otherwise
  144. int flushed; ///< 1 if data is to be flushed and no further input is expected
  145. int64_t outpts; ///< output PTS
  146. int64_t firstpts; ///< first PTS
  147. int drop_output; ///< number of output samples to drop
  148. double delayed_samples_fixup; ///< soxr 0.1.1: needed to fixup delayed_samples after flush has been called.
  149. struct AudioConvert *in_convert; ///< input conversion context
  150. struct AudioConvert *out_convert; ///< output conversion context
  151. struct AudioConvert *full_convert; ///< full conversion context (single conversion for input and output)
  152. struct ResampleContext *resample; ///< resampling context
  153. struct Resampler const *resampler; ///< resampler virtual function table
  154. double matrix[SWR_CH_MAX][SWR_CH_MAX]; ///< floating point rematrixing coefficients
  155. float matrix_flt[SWR_CH_MAX][SWR_CH_MAX]; ///< single precision floating point rematrixing coefficients
  156. uint8_t *native_matrix;
  157. uint8_t *native_one;
  158. uint8_t *native_simd_one;
  159. uint8_t *native_simd_matrix;
  160. int32_t matrix32[SWR_CH_MAX][SWR_CH_MAX]; ///< 17.15 fixed point rematrixing coefficients
  161. uint8_t matrix_ch[SWR_CH_MAX][SWR_CH_MAX+1]; ///< Lists of input channels per output channel that have non zero rematrixing coefficients
  162. mix_1_1_func_type *mix_1_1_f;
  163. mix_1_1_func_type *mix_1_1_simd;
  164. mix_2_1_func_type *mix_2_1_f;
  165. mix_2_1_func_type *mix_2_1_simd;
  166. mix_any_func_type *mix_any_f;
  167. /* TODO: callbacks for ASM optimizations */
  168. };
  169. av_warn_unused_result
  170. int swri_realloc_audio(AudioData *a, int count);
  171. void swri_noise_shaping_int16 (SwrContext *s, AudioData *dsts, const AudioData *srcs, const AudioData *noises, int count);
  172. void swri_noise_shaping_int32 (SwrContext *s, AudioData *dsts, const AudioData *srcs, const AudioData *noises, int count);
  173. void swri_noise_shaping_float (SwrContext *s, AudioData *dsts, const AudioData *srcs, const AudioData *noises, int count);
  174. void swri_noise_shaping_double(SwrContext *s, AudioData *dsts, const AudioData *srcs, const AudioData *noises, int count);
  175. av_warn_unused_result
  176. int swri_rematrix_init(SwrContext *s);
  177. void swri_rematrix_free(SwrContext *s);
  178. int swri_rematrix(SwrContext *s, AudioData *out, AudioData *in, int len, int mustcopy);
  179. int swri_rematrix_init_x86(struct SwrContext *s);
  180. av_warn_unused_result
  181. int swri_get_dither(SwrContext *s, void *dst, int len, unsigned seed, enum AVSampleFormat noise_fmt);
  182. av_warn_unused_result
  183. int swri_dither_init(SwrContext *s, enum AVSampleFormat out_fmt, enum AVSampleFormat in_fmt);
  184. void swri_audio_convert_init_aarch64(struct AudioConvert *ac,
  185. enum AVSampleFormat out_fmt,
  186. enum AVSampleFormat in_fmt,
  187. int channels);
  188. void swri_audio_convert_init_arm(struct AudioConvert *ac,
  189. enum AVSampleFormat out_fmt,
  190. enum AVSampleFormat in_fmt,
  191. int channels);
  192. void swri_audio_convert_init_x86(struct AudioConvert *ac,
  193. enum AVSampleFormat out_fmt,
  194. enum AVSampleFormat in_fmt,
  195. int channels);
  196. #endif