2 * The simplest AC-3 encoder
3 * Copyright (c) 2000 Fabrice Bellard
4 * Copyright (c) 2006-2010 Justin Ruggles <justin.ruggles@gmail.com>
5 * Copyright (c) 2006-2010 Prakash Punnoor <prakash@punnoor.de>
7 * This file is part of Libav.
9 * Libav is free software; you can redistribute it and/or
10 * modify it under the terms of the GNU Lesser General Public
11 * License as published by the Free Software Foundation; either
12 * version 2.1 of the License, or (at your option) any later version.
14 * Libav is distributed in the hope that it will be useful,
15 * but WITHOUT ANY WARRANTY; without even the implied warranty of
16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
17 * Lesser General Public License for more details.
19 * You should have received a copy of the GNU Lesser General Public
20 * License along with Libav; if not, write to the Free Software
21 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
26 * fixed-point AC-3 encoder.
29 #undef CONFIG_AC3ENC_FLOAT
33 /** Scale a float value by 2^15, convert to an integer, and clip to range -32767..32767. */
34 #define FIX15(a) av_clip(SCALE_FLOAT(a, 15), -32767, 32767)
38 * Finalize MDCT and free allocated memory.
40 static av_cold void mdct_end(AC3MDCTContext *mdct)
43 av_freep(&mdct->costab);
44 av_freep(&mdct->sintab);
45 av_freep(&mdct->xcos1);
46 av_freep(&mdct->xsin1);
47 av_freep(&mdct->rot_tmp);
48 av_freep(&mdct->cplx_tmp);
53 * Initialize FFT tables.
54 * @param ln log2(FFT size)
56 static av_cold int fft_init(AVCodecContext *avctx, AC3MDCTContext *mdct, int ln)
64 FF_ALLOC_OR_GOTO(avctx, mdct->costab, n2 * sizeof(*mdct->costab), fft_alloc_fail);
65 FF_ALLOC_OR_GOTO(avctx, mdct->sintab, n2 * sizeof(*mdct->sintab), fft_alloc_fail);
67 for (i = 0; i < n2; i++) {
68 alpha = 2.0 * M_PI * i / n;
69 mdct->costab[i] = FIX15(cos(alpha));
70 mdct->sintab[i] = FIX15(sin(alpha));
76 return AVERROR(ENOMEM);
81 * Initialize MDCT tables.
82 * @param nbits log2(MDCT size)
84 static av_cold int mdct_init(AVCodecContext *avctx, AC3MDCTContext *mdct,
94 ret = fft_init(avctx, mdct, nbits - 2);
98 mdct->window = ff_ac3_window;
100 FF_ALLOC_OR_GOTO(avctx, mdct->xcos1, n4 * sizeof(*mdct->xcos1), mdct_alloc_fail);
101 FF_ALLOC_OR_GOTO(avctx, mdct->xsin1, n4 * sizeof(*mdct->xsin1), mdct_alloc_fail);
102 FF_ALLOC_OR_GOTO(avctx, mdct->rot_tmp, n * sizeof(*mdct->rot_tmp), mdct_alloc_fail);
103 FF_ALLOC_OR_GOTO(avctx, mdct->cplx_tmp, n4 * sizeof(*mdct->cplx_tmp), mdct_alloc_fail);
105 for (i = 0; i < n4; i++) {
106 float alpha = 2.0 * M_PI * (i + 1.0 / 8.0) / n;
107 mdct->xcos1[i] = FIX15(-cos(alpha));
108 mdct->xsin1[i] = FIX15(-sin(alpha));
114 return AVERROR(ENOMEM);
119 #define BF(pre, pim, qre, qim, pre1, pim1, qre1, qim1) \
121 int ax, ay, bx, by; \
126 pre = (bx + ax) >> 1; \
127 pim = (by + ay) >> 1; \
128 qre = (bx - ax) >> 1; \
129 qim = (by - ay) >> 1; \
133 /** Complex multiply */
134 #define CMUL(pre, pim, are, aim, bre, bim, rshift) \
136 pre = (MUL16(are, bre) - MUL16(aim, bim)) >> rshift; \
137 pim = (MUL16(are, bim) + MUL16(bre, aim)) >> rshift; \
142 * Calculate a 2^n point complex FFT on 2^ln points.
143 * @param z complex input/output samples
144 * @param ln log2(FFT size)
146 static void fft(AC3MDCTContext *mdct, IComplex *z, int ln)
150 register IComplex *p,*q;
156 for (j = 0; j < np; j++) {
157 int k = av_reverse[j] >> (8 - ln);
159 FFSWAP(IComplex, z[k], z[j]);
167 BF(p[0].re, p[0].im, p[1].re, p[1].im,
168 p[0].re, p[0].im, p[1].re, p[1].im);
177 BF(p[0].re, p[0].im, p[2].re, p[2].im,
178 p[0].re, p[0].im, p[2].re, p[2].im);
179 BF(p[1].re, p[1].im, p[3].re, p[3].im,
180 p[1].re, p[1].im, p[3].im, -p[3].re);
192 for (j = 0; j < nblocks; j++) {
193 BF(p->re, p->im, q->re, q->im,
194 p->re, p->im, q->re, q->im);
197 for(l = nblocks; l < np2; l += nblocks) {
198 CMUL(tmp_re, tmp_im, mdct->costab[l], -mdct->sintab[l], q->re, q->im, 15);
199 BF(p->re, p->im, q->re, q->im,
200 p->re, p->im, tmp_re, tmp_im);
207 nblocks = nblocks >> 1;
208 nloops = nloops << 1;
214 * Calculate a 512-point MDCT
215 * @param out 256 output frequency coefficients
216 * @param in 512 windowed input audio samples
218 static void mdct512(AC3MDCTContext *mdct, int32_t *out, int16_t *in)
220 int i, re, im, n, n2, n4;
221 int16_t *rot = mdct->rot_tmp;
222 IComplex *x = mdct->cplx_tmp;
224 n = 1 << mdct->nbits;
228 /* shift to simplify computations */
229 for (i = 0; i <n4; i++)
230 rot[i] = -in[i + 3*n4];
231 memcpy(&rot[n4], &in[0], 3*n4*sizeof(*in));
234 for (i = 0; i < n4; i++) {
235 re = ((int)rot[ 2*i] - (int)rot[ n-1-2*i]) >> 1;
236 im = -((int)rot[n2+2*i] - (int)rot[n2-1-2*i]) >> 1;
237 CMUL(x[i].re, x[i].im, re, im, -mdct->xcos1[i], mdct->xsin1[i], 15);
240 fft(mdct, x, mdct->nbits - 2);
243 for (i = 0; i < n4; i++) {
246 CMUL(out[n2-1-2*i], out[2*i], re, im, mdct->xsin1[i], mdct->xcos1[i], 0);
252 * Apply KBD window to input samples prior to MDCT.
254 static void apply_window(DSPContext *dsp, int16_t *output, const int16_t *input,
255 const int16_t *window, unsigned int len)
257 dsp->apply_window_int16(output, input, window, len);
262 * Calculate the log2() of the maximum absolute value in an array.
263 * @param tab input array
264 * @param n number of values in the array
265 * @return log2(max(abs(tab[])))
267 static int log2_tab(AC3EncodeContext *s, int16_t *src, int len)
269 int v = s->ac3dsp.ac3_max_msb_abs_int16(src, len);
275 * Normalize the input samples to use the maximum available precision.
276 * This assumes signed 16-bit input samples.
278 * @return exponent shift
280 static int normalize_samples(AC3EncodeContext *s)
282 int v = 14 - log2_tab(s, s->windowed_samples, AC3_WINDOW_SIZE);
284 s->ac3dsp.ac3_lshift_int16(s->windowed_samples, AC3_WINDOW_SIZE, v);
285 /* +6 to right-shift from 31-bit to 25-bit */
291 * Scale MDCT coefficients to 25-bit signed fixed-point.
293 static void scale_coefficients(AC3EncodeContext *s)
297 for (blk = 0; blk < AC3_MAX_BLOCKS; blk++) {
298 AC3Block *block = &s->blocks[blk];
299 for (ch = 0; ch < s->channels; ch++) {
300 s->ac3dsp.ac3_rshift_int32(block->mdct_coef[ch], AC3_MAX_COEFS,
301 block->coeff_shift[ch]);
308 /*************************************************************************/
311 #include "libavutil/lfg.h"
314 #define MDCT_SAMPLES (1 << MDCT_NBITS)
315 #define FN (MDCT_SAMPLES/4)
318 static void fft_test(AC3MDCTContext *mdct, AVLFG *lfg)
320 IComplex in[FN], in1[FN];
322 float sum_re, sum_im, a;
324 for (i = 0; i < FN; i++) {
325 in[i].re = av_lfg_get(lfg) % 65535 - 32767;
326 in[i].im = av_lfg_get(lfg) % 65535 - 32767;
332 for (k = 0; k < FN; k++) {
335 for (n = 0; n < FN; n++) {
336 a = -2 * M_PI * (n * k) / FN;
337 sum_re += in1[n].re * cos(a) - in1[n].im * sin(a);
338 sum_im += in1[n].re * sin(a) + in1[n].im * cos(a);
340 av_log(NULL, AV_LOG_DEBUG, "%3d: %6d,%6d %6.0f,%6.0f\n",
341 k, in[k].re, in[k].im, sum_re / FN, sum_im / FN);
346 static void mdct_test(AC3MDCTContext *mdct, AVLFG *lfg)
348 int16_t input[MDCT_SAMPLES];
349 int32_t output[AC3_MAX_COEFS];
350 float input1[MDCT_SAMPLES];
351 float output1[AC3_MAX_COEFS];
352 float s, a, err, e, emax;
355 for (i = 0; i < MDCT_SAMPLES; i++) {
356 input[i] = (av_lfg_get(lfg) % 65535 - 32767) * 9 / 10;
357 input1[i] = input[i];
360 mdct512(mdct, output, input);
363 for (k = 0; k < AC3_MAX_COEFS; k++) {
365 for (n = 0; n < MDCT_SAMPLES; n++) {
366 a = (2*M_PI*(2*n+1+MDCT_SAMPLES/2)*(2*k+1) / (4 * MDCT_SAMPLES));
367 s += input1[n] * cos(a);
369 output1[k] = -2 * s / MDCT_SAMPLES;
374 for (i = 0; i < AC3_MAX_COEFS; i++) {
375 av_log(NULL, AV_LOG_DEBUG, "%3d: %7d %7.0f\n", i, output[i], output1[i]);
376 e = output[i] - output1[i];
381 av_log(NULL, AV_LOG_DEBUG, "err2=%f emax=%f\n", err / AC3_MAX_COEFS, emax);
391 av_log_set_level(AV_LOG_DEBUG);
394 fft_test(&mdct, &lfg);
395 mdct_test(&mdct, &lfg);
402 AVCodec ff_ac3_fixed_encoder = {
406 sizeof(AC3EncodeContext),
411 .sample_fmts = (const enum AVSampleFormat[]){AV_SAMPLE_FMT_S16,AV_SAMPLE_FMT_NONE},
412 .long_name = NULL_IF_CONFIG_SMALL("ATSC A/52A (AC-3)"),
413 .channel_layouts = ac3_channel_layouts,