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 21 #ifndef SWRESAMPLE_SWRESAMPLE_INTERNAL_H 22 #define SWRESAMPLE_SWRESAMPLE_INTERNAL_H 23 24 #include "swresample.h" 25 #include "libavutil/channel_layout.h" 26 #include "config.h" 27 28 #define SWR_CH_MAX 64 29 30 #define SQRT3_2 1.22474487139158904909 /* sqrt(3/2) */ 31 32 #define NS_TAPS 20 33 34 #if ARCH_X86_64 35 typedef int64_t integer; 36 #else 37 typedef int integer; 38 #endif 39 40 typedef void (mix_1_1_func_type)(void *out, const void *in, void *coeffp, integer index, integer len); 41 typedef void (mix_2_1_func_type)(void *out, const void *in1, const void *in2, void *coeffp, integer index1, integer index2, integer len); 42 43 typedef void (mix_any_func_type)(uint8_t **out, const uint8_t **in1, void *coeffp, integer len); 44 45 typedef struct AudioData{ 46 uint8_t *ch[SWR_CH_MAX]; ///< samples buffer per channel 47 uint8_t *data; ///< samples buffer 48 int ch_count; ///< number of channels 49 int bps; ///< bytes per sample 50 int count; ///< number of samples 51 int planar; ///< 1 if planar audio, 0 otherwise 52 enum AVSampleFormat fmt; ///< sample format 53 } AudioData; 54 55 struct DitherContext { 56 int method; 57 int noise_pos; 58 float scale; 59 float noise_scale; ///< Noise scale 60 int ns_taps; ///< Noise shaping dither taps 61 float ns_scale; ///< Noise shaping dither scale 62 float ns_scale_1; ///< Noise shaping dither scale^-1 63 int ns_pos; ///< Noise shaping dither position 64 float ns_coeffs[NS_TAPS]; ///< Noise shaping filter coefficients 65 float ns_errors[SWR_CH_MAX][2*NS_TAPS]; 66 AudioData noise; ///< noise used for dithering 67 AudioData temp; ///< temporary storage when writing into the input buffer isn't possible 68 int output_sample_bits; ///< the number of used output bits, needed to scale dither correctly 69 }; 70 71 typedef struct ResampleContext * (* resample_init_func)(struct ResampleContext *c, int out_rate, int in_rate, int filter_size, int phase_shift, int linear, 72 double cutoff, enum AVSampleFormat format, enum SwrFilterType filter_type, double kaiser_beta, double precision, int cheby, int exact_rational); 73 typedef void (* resample_free_func)(struct ResampleContext **c); 74 typedef int (* multiple_resample_func)(struct ResampleContext *c, AudioData *dst, int dst_size, AudioData *src, int src_size, int *consumed); 75 typedef int (* resample_flush_func)(struct SwrContext *c); 76 typedef int (* set_compensation_func)(struct ResampleContext *c, int sample_delta, int compensation_distance); 77 typedef int64_t (* get_delay_func)(struct SwrContext *s, int64_t base); 78 typedef int (* invert_initial_buffer_func)(struct ResampleContext *c, AudioData *dst, const AudioData *src, int src_size, int *dst_idx, int *dst_count); 79 typedef int64_t (* get_out_samples_func)(struct SwrContext *s, int in_samples); 80 81 struct Resampler { 82 resample_init_func init; 83 resample_free_func free; 84 multiple_resample_func multiple_resample; 85 resample_flush_func flush; 86 set_compensation_func set_compensation; 87 get_delay_func get_delay; 88 invert_initial_buffer_func invert_initial_buffer; 89 get_out_samples_func get_out_samples; 90 }; 91 92 extern struct Resampler const swri_resampler; 93 extern struct Resampler const swri_soxr_resampler; 94 95 struct SwrContext { 96 const AVClass *av_class; ///< AVClass used for AVOption and av_log() 97 int log_level_offset; ///< logging level offset 98 void *log_ctx; ///< parent logging context 99 enum AVSampleFormat in_sample_fmt; ///< input sample format 100 enum AVSampleFormat int_sample_fmt; ///< internal sample format (AV_SAMPLE_FMT_FLTP or AV_SAMPLE_FMT_S16P) 101 enum AVSampleFormat out_sample_fmt; ///< output sample format 102 AVChannelLayout in_ch_layout; ///< input channel layout 103 AVChannelLayout out_ch_layout; ///< output channel layout 104 int in_sample_rate; ///< input sample rate 105 int out_sample_rate; ///< output sample rate 106 int flags; ///< miscellaneous flags such as SWR_FLAG_RESAMPLE 107 float slev; ///< surround mixing level 108 float clev; ///< center mixing level 109 float lfe_mix_level; ///< LFE mixing level 110 float rematrix_volume; ///< rematrixing volume coefficient 111 float rematrix_maxval; ///< maximum value for rematrixing output 112 int matrix_encoding; /**< matrixed stereo encoding */ 113 const int *channel_map; ///< channel index (or -1 if muted channel) map 114 int used_ch_count; ///< number of used input channels (mapped channel count if channel_map, otherwise in.ch_count) 115 int engine; 116 117 int user_used_ch_count; ///< User set used channel count 118 #if FF_API_OLD_CHANNEL_LAYOUT 119 int user_in_ch_count; ///< User set input channel count 120 int user_out_ch_count; ///< User set output channel count 121 int64_t user_in_ch_layout; ///< User set input channel layout 122 int64_t user_out_ch_layout; ///< User set output channel layout 123 #endif 124 AVChannelLayout user_in_chlayout; ///< User set input channel layout 125 AVChannelLayout user_out_chlayout; ///< User set output channel layout 126 enum AVSampleFormat user_int_sample_fmt; ///< User set internal sample format 127 int user_dither_method; ///< User set dither method 128 129 struct DitherContext dither; 130 131 int filter_size; /**< length of each FIR filter in the resampling filterbank relative to the cutoff frequency */ 132 int phase_shift; /**< log2 of the number of entries in the resampling polyphase filterbank */ 133 int linear_interp; /**< if 1 then the resampling FIR filter will be linearly interpolated */ 134 int exact_rational; /**< if 1 then enable non power of 2 phase_count */ 135 double cutoff; /**< resampling cutoff frequency (swr: 6dB point; soxr: 0dB point). 1.0 corresponds to half the output sample rate */ 136 int filter_type; /**< swr resampling filter type */ 137 double kaiser_beta; /**< swr beta value for Kaiser window (only applicable if filter_type == AV_FILTER_TYPE_KAISER) */ 138 double precision; /**< soxr resampling precision (in bits) */ 139 int cheby; /**< soxr: if 1 then passband rolloff will be none (Chebyshev) & irrational ratio approximation precision will be higher */ 140 141 float min_compensation; ///< swr minimum below which no compensation will happen 142 float min_hard_compensation; ///< swr minimum below which no silence inject / sample drop will happen 143 float soft_compensation_duration; ///< swr duration over which soft compensation is applied 144 float max_soft_compensation; ///< swr maximum soft compensation in seconds over soft_compensation_duration 145 float async; ///< swr simple 1 parameter async, similar to ffmpegs -async 146 int64_t firstpts_in_samples; ///< swr first pts in samples 147 148 int resample_first; ///< 1 if resampling must come first, 0 if rematrixing 149 int rematrix; ///< flag to indicate if rematrixing is needed (basically if input and output layouts mismatch) 150 int rematrix_custom; ///< flag to indicate that a custom matrix has been defined 151 152 AudioData in; ///< input audio data 153 AudioData postin; ///< post-input audio data: used for rematrix/resample 154 AudioData midbuf; ///< intermediate audio data (postin/preout) 155 AudioData preout; ///< pre-output audio data: used for rematrix/resample 156 AudioData out; ///< converted output audio data 157 AudioData in_buffer; ///< cached audio data (convert and resample purpose) 158 AudioData silence; ///< temporary with silence 159 AudioData drop_temp; ///< temporary used to discard output 160 int in_buffer_index; ///< cached buffer position 161 int in_buffer_count; ///< cached buffer length 162 int resample_in_constraint; ///< 1 if the input end was reach before the output end, 0 otherwise 163 int flushed; ///< 1 if data is to be flushed and no further input is expected 164 int64_t outpts; ///< output PTS 165 int64_t firstpts; ///< first PTS 166 int drop_output; ///< number of output samples to drop 167 double delayed_samples_fixup; ///< soxr 0.1.1: needed to fixup delayed_samples after flush has been called. 168 169 struct AudioConvert *in_convert; ///< input conversion context 170 struct AudioConvert *out_convert; ///< output conversion context 171 struct AudioConvert *full_convert; ///< full conversion context (single conversion for input and output) 172 struct ResampleContext *resample; ///< resampling context 173 struct Resampler const *resampler; ///< resampler virtual function table 174 175 double matrix[SWR_CH_MAX][SWR_CH_MAX]; ///< floating point rematrixing coefficients 176 float matrix_flt[SWR_CH_MAX][SWR_CH_MAX]; ///< single precision floating point rematrixing coefficients 177 uint8_t *native_matrix; 178 uint8_t *native_one; 179 uint8_t *native_simd_one; 180 uint8_t *native_simd_matrix; 181 int32_t matrix32[SWR_CH_MAX][SWR_CH_MAX]; ///< 17.15 fixed point rematrixing coefficients 182 uint8_t matrix_ch[SWR_CH_MAX][SWR_CH_MAX+1]; ///< Lists of input channels per output channel that have non zero rematrixing coefficients 183 mix_1_1_func_type *mix_1_1_f; 184 mix_1_1_func_type *mix_1_1_simd; 185 186 mix_2_1_func_type *mix_2_1_f; 187 mix_2_1_func_type *mix_2_1_simd; 188 189 mix_any_func_type *mix_any_f; 190 191 /* TODO: callbacks for ASM optimizations */ 192 }; 193 194 av_warn_unused_result 195 int swri_realloc_audio(AudioData *a, int count); 196 197 void swri_noise_shaping_int16 (SwrContext *s, AudioData *dsts, const AudioData *srcs, const AudioData *noises, int count); 198 void swri_noise_shaping_int32 (SwrContext *s, AudioData *dsts, const AudioData *srcs, const AudioData *noises, int count); 199 void swri_noise_shaping_float (SwrContext *s, AudioData *dsts, const AudioData *srcs, const AudioData *noises, int count); 200 void swri_noise_shaping_double(SwrContext *s, AudioData *dsts, const AudioData *srcs, const AudioData *noises, int count); 201 202 av_warn_unused_result 203 int swri_rematrix_init(SwrContext *s); 204 void swri_rematrix_free(SwrContext *s); 205 int swri_rematrix(SwrContext *s, AudioData *out, AudioData *in, int len, int mustcopy); 206 int swri_rematrix_init_x86(struct SwrContext *s); 207 208 av_warn_unused_result 209 int swri_get_dither(SwrContext *s, void *dst, int len, unsigned seed, enum AVSampleFormat noise_fmt); 210 av_warn_unused_result 211 int swri_dither_init(SwrContext *s, enum AVSampleFormat out_fmt, enum AVSampleFormat in_fmt); 212 213 void swri_audio_convert_init_aarch64(struct AudioConvert *ac, 214 enum AVSampleFormat out_fmt, 215 enum AVSampleFormat in_fmt, 216 int channels); 217 void swri_audio_convert_init_arm(struct AudioConvert *ac, 218 enum AVSampleFormat out_fmt, 219 enum AVSampleFormat in_fmt, 220 int channels); 221 void swri_audio_convert_init_x86(struct AudioConvert *ac, 222 enum AVSampleFormat out_fmt, 223 enum AVSampleFormat in_fmt, 224 int channels); 225 226 #endif 227