aacenc_utils.h 8.46 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
/*
 * AAC encoder utilities
 * Copyright (C) 2015 Rostislav Pehlivanov
 *
 * This file is part of FFmpeg.
 *
 * FFmpeg is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * FFmpeg is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with FFmpeg; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

/**
 * @file
 * AAC encoder utilities
 * @author Rostislav Pehlivanov ( atomnuker gmail com )
 */

#ifndef AVCODEC_AACENC_UTILS_H
#define AVCODEC_AACENC_UTILS_H

31
#include "libavutil/ffmath.h"
32 33
#include "aac.h"
#include "aacenctab.h"
34
#include "aactab.h"
35 36 37 38 39 40 41 42 43 44 45 46 47 48

#define ROUND_STANDARD 0.4054f
#define ROUND_TO_ZERO 0.1054f
#define C_QUANT 0.4054f

static inline void abs_pow34_v(float *out, const float *in, const int size)
{
    int i;
    for (i = 0; i < size; i++) {
        float a = fabsf(in[i]);
        out[i] = sqrtf(a * sqrtf(a));
    }
}

49 50 51 52 53
static inline float pos_pow34(float a)
{
    return sqrtf(a * sqrtf(a));
}

54 55 56 57 58 59 60 61 62 63 64 65
/**
 * Quantize one coefficient.
 * @return absolute value of the quantized coefficient
 * @see 3GPP TS26.403 5.6.2 "Scalefactor determination"
 */
static inline int quant(float coef, const float Q, const float rounding)
{
    float a = coef * Q;
    return sqrtf(a * sqrtf(a)) + rounding;
}

static inline void quantize_bands(int *out, const float *in, const float *scaled,
66
                                  int size, int is_signed, int maxval, const float Q34,
67 68 69 70
                                  const float rounding)
{
    int i;
    for (i = 0; i < size; i++) {
71
        float qc = scaled[i] * Q34;
72
        int tmp = (int)FFMIN(qc + rounding, (float)maxval);
73
        if (is_signed && in[i] < 0.0f) {
74
            tmp = -tmp;
75
        }
76
        out[i] = tmp;
77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93
    }
}

static inline float find_max_val(int group_len, int swb_size, const float *scaled)
{
    float maxval = 0.0f;
    int w2, i;
    for (w2 = 0; w2 < group_len; w2++) {
        for (i = 0; i < swb_size; i++) {
            maxval = FFMAX(maxval, scaled[w2*128+i]);
        }
    }
    return maxval;
}

static inline int find_min_book(float maxval, int sf)
{
94
    float Q34 = ff_aac_pow34sf_tab[POW_SF2_ZERO - sf + SCALE_ONE_POS - SCALE_DIV_512];
95 96
    int qmaxval, cb;
    qmaxval = maxval * Q34 + C_QUANT;
97 98 99 100
    if (qmaxval >= (FF_ARRAY_ELEMS(aac_maxval_cb)))
        cb = 11;
    else
        cb = aac_maxval_cb[qmaxval];
101 102 103
    return cb;
}

104 105
static inline float find_form_factor(int group_len, int swb_size, float thresh,
                                     const float *scaled, float nzslope) {
106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125
    const float iswb_size = 1.0f / swb_size;
    const float iswb_sizem1 = 1.0f / (swb_size - 1);
    const float ethresh = thresh;
    float form = 0.0f, weight = 0.0f;
    int w2, i;
    for (w2 = 0; w2 < group_len; w2++) {
        float e = 0.0f, e2 = 0.0f, var = 0.0f, maxval = 0.0f;
        float nzl = 0;
        for (i = 0; i < swb_size; i++) {
            float s = fabsf(scaled[w2*128+i]);
            maxval = FFMAX(maxval, s);
            e += s;
            e2 += s *= s;
            /* We really don't want a hard non-zero-line count, since
             * even below-threshold lines do add up towards band spectral power.
             * So, fall steeply towards zero, but smoothly
             */
            if (s >= ethresh) {
                nzl += 1.0f;
            } else {
126 127 128 129
                if (nzslope == 2.f)
                    nzl += (s / ethresh) * (s / ethresh);
                else
                    nzl += ff_fast_powf(s / ethresh, nzslope);
130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155
            }
        }
        if (e2 > thresh) {
            float frm;
            e *= iswb_size;

            /** compute variance */
            for (i = 0; i < swb_size; i++) {
                float d = fabsf(scaled[w2*128+i]) - e;
                var += d*d;
            }
            var = sqrtf(var * iswb_sizem1);

            e2 *= iswb_size;
            frm = e / FFMIN(e+4*var,maxval);
            form += e2 * sqrtf(frm) / FFMAX(0.5f,nzl);
            weight += e2;
        }
    }
    if (weight > 0) {
        return form / weight;
    } else {
        return 1.0f;
    }
}

156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184
/** Return the minimum scalefactor where the quantized coef does not clip. */
static inline uint8_t coef2minsf(float coef)
{
    return av_clip_uint8(log2f(coef)*4 - 69 + SCALE_ONE_POS - SCALE_DIV_512);
}

/** Return the maximum scalefactor where the quantized coef is not zero. */
static inline uint8_t coef2maxsf(float coef)
{
    return av_clip_uint8(log2f(coef)*4 +  6 + SCALE_ONE_POS - SCALE_DIV_512);
}

/*
 * Returns the closest possible index to an array of float values, given a value.
 */
static inline int quant_array_idx(const float val, const float *arr, const int num)
{
    int i, index = 0;
    float quant_min_err = INFINITY;
    for (i = 0; i < num; i++) {
        float error = (val - arr[i])*(val - arr[i]);
        if (error < quant_min_err) {
            quant_min_err = error;
            index = i;
        }
    }
    return index;
}

185 186 187 188 189 190 191 192
/**
 * approximates exp10f(-3.0f*(0.5f + 0.5f * cosf(FFMIN(b,15.5f) / 15.5f)))
 */
static av_always_inline float bval2bmax(float b)
{
    return 0.001f + 0.0035f * (b*b*b) / (15.5f*15.5f*15.5f);
}

193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254
/*
 * Compute a nextband map to be used with SF delta constraint utilities.
 * The nextband array should contain 128 elements, and positions that don't
 * map to valid, nonzero bands of the form w*16+g (with w being the initial
 * window of the window group, only) are left indetermined.
 */
static inline void ff_init_nextband_map(const SingleChannelElement *sce, uint8_t *nextband)
{
    unsigned char prevband = 0;
    int w, g;
    /** Just a safe default */
    for (g = 0; g < 128; g++)
        nextband[g] = g;

    /** Now really navigate the nonzero band chain */
    for (w = 0; w < sce->ics.num_windows; w += sce->ics.group_len[w]) {
        for (g = 0; g < sce->ics.num_swb; g++) {
            if (!sce->zeroes[w*16+g] && sce->band_type[w*16+g] < RESERVED_BT)
                prevband = nextband[prevband] = w*16+g;
        }
    }
    nextband[prevband] = prevband; /* terminate */
}

/*
 * Updates nextband to reflect a removed band (equivalent to
 * calling ff_init_nextband_map after marking a band as zero)
 */
static inline void ff_nextband_remove(uint8_t *nextband, int prevband, int band)
{
    nextband[prevband] = nextband[band];
}

/*
 * Checks whether the specified band could be removed without inducing
 * scalefactor delta that violates SF delta encoding constraints.
 * prev_sf has to be the scalefactor of the previous nonzero, nonspecial
 * band, in encoding order, or negative if there was no such band.
 */
static inline int ff_sfdelta_can_remove_band(const SingleChannelElement *sce,
    const uint8_t *nextband, int prev_sf, int band)
{
    return prev_sf >= 0
        && sce->sf_idx[nextband[band]] >= (prev_sf - SCALE_MAX_DIFF)
        && sce->sf_idx[nextband[band]] <= (prev_sf + SCALE_MAX_DIFF);
}

/*
 * Checks whether the specified band's scalefactor could be replaced
 * with another one without violating SF delta encoding constraints.
 * prev_sf has to be the scalefactor of the previous nonzero, nonsepcial
 * band, in encoding order, or negative if there was no such band.
 */
static inline int ff_sfdelta_can_replace(const SingleChannelElement *sce,
    const uint8_t *nextband, int prev_sf, int new_sf, int band)
{
    return new_sf >= (prev_sf - SCALE_MAX_DIFF)
        && new_sf <= (prev_sf + SCALE_MAX_DIFF)
        && sce->sf_idx[nextband[band]] >= (new_sf - SCALE_MAX_DIFF)
        && sce->sf_idx[nextband[band]] <= (new_sf + SCALE_MAX_DIFF);
}

255 256 257 258 259 260 261 262 263 264 265 266 267
/**
 * linear congruential pseudorandom number generator
 *
 * @param   previous_val    pointer to the current state of the generator
 *
 * @return  Returns a 32-bit pseudorandom integer
 */
static av_always_inline int lcg_random(unsigned previous_val)
{
    union { unsigned u; int s; } v = { previous_val * 1664525u + 1013904223 };
    return v.s;
}

268 269 270 271 272 273 274 275 276 277 278 279
#define ERROR_IF(cond, ...) \
    if (cond) { \
        av_log(avctx, AV_LOG_ERROR, __VA_ARGS__); \
        return AVERROR(EINVAL); \
    }

#define WARN_IF(cond, ...) \
    if (cond) { \
        av_log(avctx, AV_LOG_WARNING, __VA_ARGS__); \
    }

#endif /* AVCODEC_AACENC_UTILS_H */