From cc2160de5cc05dc3aa77f3a34358e66f6193c8c9 Mon Sep 17 00:00:00 2001 From: Michał Cichoń Date: Tue, 26 Jun 2012 20:35:30 +0200 Subject: Add support for AAC. --- faad2/src/libfaad/ssr.c | 175 ++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 175 insertions(+) create mode 100644 faad2/src/libfaad/ssr.c (limited to 'faad2/src/libfaad/ssr.c') diff --git a/faad2/src/libfaad/ssr.c b/faad2/src/libfaad/ssr.c new file mode 100644 index 0000000..c708ce2 --- /dev/null +++ b/faad2/src/libfaad/ssr.c @@ -0,0 +1,175 @@ +/* +** FAAD2 - Freeware Advanced Audio (AAC) Decoder including SBR decoding +** Copyright (C) 2003-2005 M. Bakker, Nero AG, http://www.nero.com +** +** This program is free software; you can redistribute it and/or modify +** it under the terms of the GNU General Public License as published by +** the Free Software Foundation; either version 2 of the License, or +** (at your option) any later version. +** +** This program 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 General Public License for more details. +** +** You should have received a copy of the GNU General Public License +** along with this program; if not, write to the Free Software +** Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. +** +** Any non-GPL usage of this software or parts of this software is strictly +** forbidden. +** +** The "appropriate copyright message" mentioned in section 2c of the GPLv2 +** must read: "Code from FAAD2 is copyright (c) Nero AG, www.nero.com" +** +** Commercial non-GPL licensing of this software is possible. +** For more info contact Nero AG through Mpeg4AAClicense@nero.com. +** +** $Id: ssr.c,v 1.19 2007/11/01 12:33:36 menno Exp $ +**/ + +#include "common.h" +#include "structs.h" + +#ifdef SSR_DEC + +#include "syntax.h" +#include "filtbank.h" +#include "ssr.h" +#include "ssr_fb.h" + +void ssr_decode(ssr_info *ssr, fb_info *fb, uint8_t window_sequence, + uint8_t window_shape, uint8_t window_shape_prev, + real_t *freq_in, real_t *time_out, real_t *overlap, + real_t ipqf_buffer[SSR_BANDS][96/4], + real_t *prev_fmd, uint16_t frame_len) +{ + uint8_t band; + uint16_t ssr_frame_len = frame_len/SSR_BANDS; + real_t time_tmp[2048] = {0}; + real_t output[1024] = {0}; + + for (band = 0; band < SSR_BANDS; band++) + { + int16_t j; + + /* uneven bands have inverted frequency scale */ + if (band == 1 || band == 3) + { + for (j = 0; j < ssr_frame_len/2; j++) + { + real_t tmp; + tmp = freq_in[j + ssr_frame_len*band]; + freq_in[j + ssr_frame_len*band] = + freq_in[ssr_frame_len - j - 1 + ssr_frame_len*band]; + freq_in[ssr_frame_len - j - 1 + ssr_frame_len*band] = tmp; + } + } + + /* non-overlapping inverse filterbank for SSR */ + ssr_ifilter_bank(fb, window_sequence, window_shape, window_shape_prev, + freq_in + band*ssr_frame_len, time_tmp + band*ssr_frame_len, + ssr_frame_len); + + /* gain control */ + ssr_gain_control(ssr, time_tmp, output, overlap, prev_fmd, + band, window_sequence, ssr_frame_len); + } + + /* inverse pqf to bring subbands together again */ + ssr_ipqf(ssr, output, time_out, ipqf_buffer, frame_len, SSR_BANDS); +} + +static void ssr_gain_control(ssr_info *ssr, real_t *data, real_t *output, + real_t *overlap, real_t *prev_fmd, uint8_t band, + uint8_t window_sequence, uint16_t frame_len) +{ + uint16_t i; + real_t gc_function[2*1024/SSR_BANDS]; + + if (window_sequence != EIGHT_SHORT_SEQUENCE) + { + ssr_gc_function(ssr, &prev_fmd[band * frame_len*2], + gc_function, window_sequence, band, frame_len); + + for (i = 0; i < frame_len*2; i++) + data[band * frame_len*2 + i] *= gc_function[i]; + for (i = 0; i < frame_len; i++) + { + output[band*frame_len + i] = overlap[band*frame_len + i] + + data[band*frame_len*2 + i]; + } + for (i = 0; i < frame_len; i++) + { + overlap[band*frame_len + i] = + data[band*frame_len*2 + frame_len + i]; + } + } else { + uint8_t w; + for (w = 0; w < 8; w++) + { + uint16_t frame_len8 = frame_len/8; + uint16_t frame_len16 = frame_len/16; + + ssr_gc_function(ssr, &prev_fmd[band*frame_len*2 + w*frame_len*2/8], + gc_function, window_sequence, frame_len); + + for (i = 0; i < frame_len8*2; i++) + data[band*frame_len*2 + w*frame_len8*2+i] *= gc_function[i]; + for (i = 0; i < frame_len8; i++) + { + overlap[band*frame_len + i + 7*frame_len16 + w*frame_len8] += + data[band*frame_len*2 + 2*w*frame_len8 + i]; + } + for (i = 0; i < frame_len8; i++) + { + overlap[band*frame_len + i + 7*frame_len16 + (w+1)*frame_len8] = + data[band*frame_len*2 + 2*w*frame_len8 + frame_len8 + i]; + } + } + for (i = 0; i < frame_len; i++) + output[band*frame_len + i] = overlap[band*frame_len + i]; + for (i = 0; i < frame_len; i++) + overlap[band*frame_len + i] = overlap[band*frame_len + i + frame_len]; + } +} + +static void ssr_gc_function(ssr_info *ssr, real_t *prev_fmd, + real_t *gc_function, uint8_t window_sequence, + uint8_t band, uint16_t frame_len) +{ + uint16_t i; + uint16_t len_area1, len_area2; + int32_t aloc[10]; + real_t alev[10]; + + switch (window_sequence) + { + case ONLY_LONG_SEQUENCE: + len_area1 = frame_len/SSR_BANDS; + len_area2 = 0; + break; + case LONG_START_SEQUENCE: + len_area1 = (frame_len/SSR_BANDS)*7/32; + len_area2 = (frame_len/SSR_BANDS)/16; + break; + case EIGHT_SHORT_SEQUENCE: + len_area1 = (frame_len/8)/SSR_BANDS; + len_area2 = 0; + break; + case LONG_STOP_SEQUENCE: + len_area1 = (frame_len/SSR_BANDS); + len_area2 = 0; + break; + } + + /* decode bitstream information */ + + /* build array M */ + + + for (i = 0; i < frame_len*2; i++) + gc_function[i] = 1; +} + +#endif -- cgit v1.2.3