| 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 SWR_INTERNAL_H |
| 22 | #define SWR_INTERNAL_H |
| 23 | |
| 24 | #include "swresample.h" |
| 25 | #include "libavutil/channel_layout.h" |
| 26 | #include "config.h" |
| 27 | |
| 28 | #define SWR_CH_MAX 32 |
| 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 | enum SwrDitherType 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 | struct SwrContext { |
| 72 | const AVClass *av_class; ///< AVClass used for AVOption and av_log() |
| 73 | int log_level_offset; ///< logging level offset |
| 74 | void *log_ctx; ///< parent logging context |
| 75 | enum AVSampleFormat in_sample_fmt; ///< input sample format |
| 76 | enum AVSampleFormat int_sample_fmt; ///< internal sample format (AV_SAMPLE_FMT_FLTP or AV_SAMPLE_FMT_S16P) |
| 77 | enum AVSampleFormat out_sample_fmt; ///< output sample format |
| 78 | int64_t in_ch_layout; ///< input channel layout |
| 79 | int64_t out_ch_layout; ///< output channel layout |
| 80 | int in_sample_rate; ///< input sample rate |
| 81 | int out_sample_rate; ///< output sample rate |
| 82 | int flags; ///< miscellaneous flags such as SWR_FLAG_RESAMPLE |
| 83 | float slev; ///< surround mixing level |
| 84 | float clev; ///< center mixing level |
| 85 | float lfe_mix_level; ///< LFE mixing level |
| 86 | float rematrix_volume; ///< rematrixing volume coefficient |
| 87 | float rematrix_maxval; ///< maximum value for rematrixing output |
| 88 | enum AVMatrixEncoding matrix_encoding; /**< matrixed stereo encoding */ |
| 89 | const int *channel_map; ///< channel index (or -1 if muted channel) map |
| 90 | int used_ch_count; ///< number of used input channels (mapped channel count if channel_map, otherwise in.ch_count) |
| 91 | enum SwrEngine engine; |
| 92 | |
| 93 | struct DitherContext dither; |
| 94 | |
| 95 | int filter_size; /**< length of each FIR filter in the resampling filterbank relative to the cutoff frequency */ |
| 96 | int phase_shift; /**< log2 of the number of entries in the resampling polyphase filterbank */ |
| 97 | int linear_interp; /**< if 1 then the resampling FIR filter will be linearly interpolated */ |
| 98 | double cutoff; /**< resampling cutoff frequency (swr: 6dB point; soxr: 0dB point). 1.0 corresponds to half the output sample rate */ |
| 99 | enum SwrFilterType filter_type; /**< swr resampling filter type */ |
| 100 | int kaiser_beta; /**< swr beta value for Kaiser window (only applicable if filter_type == AV_FILTER_TYPE_KAISER) */ |
| 101 | double precision; /**< soxr resampling precision (in bits) */ |
| 102 | int cheby; /**< soxr: if 1 then passband rolloff will be none (Chebyshev) & irrational ratio approximation precision will be higher */ |
| 103 | |
| 104 | float min_compensation; ///< swr minimum below which no compensation will happen |
| 105 | float min_hard_compensation; ///< swr minimum below which no silence inject / sample drop will happen |
| 106 | float soft_compensation_duration; ///< swr duration over which soft compensation is applied |
| 107 | float max_soft_compensation; ///< swr maximum soft compensation in seconds over soft_compensation_duration |
| 108 | float async; ///< swr simple 1 parameter async, similar to ffmpegs -async |
| 109 | int64_t firstpts_in_samples; ///< swr first pts in samples |
| 110 | |
| 111 | int resample_first; ///< 1 if resampling must come first, 0 if rematrixing |
| 112 | int rematrix; ///< flag to indicate if rematrixing is needed (basically if input and output layouts mismatch) |
| 113 | int rematrix_custom; ///< flag to indicate that a custom matrix has been defined |
| 114 | |
| 115 | AudioData in; ///< input audio data |
| 116 | AudioData postin; ///< post-input audio data: used for rematrix/resample |
| 117 | AudioData midbuf; ///< intermediate audio data (postin/preout) |
| 118 | AudioData preout; ///< pre-output audio data: used for rematrix/resample |
| 119 | AudioData out; ///< converted output audio data |
| 120 | AudioData in_buffer; ///< cached audio data (convert and resample purpose) |
| 121 | AudioData silence; ///< temporary with silence |
| 122 | AudioData drop_temp; ///< temporary used to discard output |
| 123 | int in_buffer_index; ///< cached buffer position |
| 124 | int in_buffer_count; ///< cached buffer length |
| 125 | int resample_in_constraint; ///< 1 if the input end was reach before the output end, 0 otherwise |
| 126 | int flushed; ///< 1 if data is to be flushed and no further input is expected |
| 127 | int64_t outpts; ///< output PTS |
| 128 | int64_t firstpts; ///< first PTS |
| 129 | int drop_output; ///< number of output samples to drop |
| 130 | |
| 131 | struct AudioConvert *in_convert; ///< input conversion context |
| 132 | struct AudioConvert *out_convert; ///< output conversion context |
| 133 | struct AudioConvert *full_convert; ///< full conversion context (single conversion for input and output) |
| 134 | struct ResampleContext *resample; ///< resampling context |
| 135 | struct Resampler const *resampler; ///< resampler virtual function table |
| 136 | |
| 137 | float matrix[SWR_CH_MAX][SWR_CH_MAX]; ///< floating point rematrixing coefficients |
| 138 | uint8_t *native_matrix; |
| 139 | uint8_t *native_one; |
| 140 | uint8_t *native_simd_one; |
| 141 | uint8_t *native_simd_matrix; |
| 142 | int32_t matrix32[SWR_CH_MAX][SWR_CH_MAX]; ///< 17.15 fixed point rematrixing coefficients |
| 143 | uint8_t matrix_ch[SWR_CH_MAX][SWR_CH_MAX+1]; ///< Lists of input channels per output channel that have non zero rematrixing coefficients |
| 144 | mix_1_1_func_type *mix_1_1_f; |
| 145 | mix_1_1_func_type *mix_1_1_simd; |
| 146 | |
| 147 | mix_2_1_func_type *mix_2_1_f; |
| 148 | mix_2_1_func_type *mix_2_1_simd; |
| 149 | |
| 150 | mix_any_func_type *mix_any_f; |
| 151 | |
| 152 | /* TODO: callbacks for ASM optimizations */ |
| 153 | }; |
| 154 | |
| 155 | typedef struct ResampleContext * (* resample_init_func)(struct ResampleContext *c, int out_rate, int in_rate, int filter_size, int phase_shift, int linear, |
| 156 | double cutoff, enum AVSampleFormat format, enum SwrFilterType filter_type, int kaiser_beta, double precision, int cheby); |
| 157 | typedef void (* resample_free_func)(struct ResampleContext **c); |
| 158 | typedef int (* multiple_resample_func)(struct ResampleContext *c, AudioData *dst, int dst_size, AudioData *src, int src_size, int *consumed); |
| 159 | typedef int (* resample_flush_func)(struct SwrContext *c); |
| 160 | typedef int (* set_compensation_func)(struct ResampleContext *c, int sample_delta, int compensation_distance); |
| 161 | typedef int64_t (* get_delay_func)(struct SwrContext *s, int64_t base); |
| 162 | typedef int (* invert_initial_buffer_func)(struct ResampleContext *c, AudioData *dst, const AudioData *src, int src_size, int *dst_idx, int *dst_count); |
| 163 | |
| 164 | struct Resampler { |
| 165 | resample_init_func init; |
| 166 | resample_free_func free; |
| 167 | multiple_resample_func multiple_resample; |
| 168 | resample_flush_func flush; |
| 169 | set_compensation_func set_compensation; |
| 170 | get_delay_func get_delay; |
| 171 | invert_initial_buffer_func invert_initial_buffer; |
| 172 | }; |
| 173 | |
| 174 | extern struct Resampler const swri_resampler; |
| 175 | |
| 176 | int swri_realloc_audio(AudioData *a, int count); |
| 177 | |
| 178 | void swri_noise_shaping_int16 (SwrContext *s, AudioData *dsts, const AudioData *srcs, const AudioData *noises, int count); |
| 179 | void swri_noise_shaping_int32 (SwrContext *s, AudioData *dsts, const AudioData *srcs, const AudioData *noises, int count); |
| 180 | void swri_noise_shaping_float (SwrContext *s, AudioData *dsts, const AudioData *srcs, const AudioData *noises, int count); |
| 181 | void swri_noise_shaping_double(SwrContext *s, AudioData *dsts, const AudioData *srcs, const AudioData *noises, int count); |
| 182 | |
| 183 | int swri_rematrix_init(SwrContext *s); |
| 184 | void swri_rematrix_free(SwrContext *s); |
| 185 | int swri_rematrix(SwrContext *s, AudioData *out, AudioData *in, int len, int mustcopy); |
| 186 | void swri_rematrix_init_x86(struct SwrContext *s); |
| 187 | |
| 188 | void swri_get_dither(SwrContext *s, void *dst, int len, unsigned seed, enum AVSampleFormat noise_fmt); |
| 189 | int swri_dither_init(SwrContext *s, enum AVSampleFormat out_fmt, enum AVSampleFormat in_fmt); |
| 190 | |
| 191 | void swri_audio_convert_init_aarch64(struct AudioConvert *ac, |
| 192 | enum AVSampleFormat out_fmt, |
| 193 | enum AVSampleFormat in_fmt, |
| 194 | int channels); |
| 195 | void swri_audio_convert_init_arm(struct AudioConvert *ac, |
| 196 | enum AVSampleFormat out_fmt, |
| 197 | enum AVSampleFormat in_fmt, |
| 198 | int channels); |
| 199 | void swri_audio_convert_init_x86(struct AudioConvert *ac, |
| 200 | enum AVSampleFormat out_fmt, |
| 201 | enum AVSampleFormat in_fmt, |
| 202 | int channels); |
| 203 | #endif |