X-Git-Url: https://git.sesse.net/?a=blobdiff_plain;f=libavcodec%2Fpsymodel.c;h=5179ede0836901f8b7692f41a86f3151f84d8de9;hb=081961f819c0b16c7a860d7da7d39f1fd91bd2f0;hp=a2e469c5d743529f14366843cf4e9d0bf4a3eb94;hpb=0e107f7890338aad1190f23997562a6de3ca29e9;p=ffmpeg diff --git a/libavcodec/psymodel.c b/libavcodec/psymodel.c index a2e469c5d74..5179ede0836 100644 --- a/libavcodec/psymodel.c +++ b/libavcodec/psymodel.c @@ -2,41 +2,65 @@ * audio encoder psychoacoustic model * Copyright (C) 2008 Konstantin Shishkov * - * This file is part of FFmpeg. + * This file is part of Libav. * - * FFmpeg is free software; you can redistribute it and/or + * Libav 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, + * Libav 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 + * License along with Libav; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ +#include + #include "avcodec.h" #include "psymodel.h" #include "iirfilter.h" +#include "libavutil/mem.h" extern const FFPsyModel ff_aac_psy_model; -av_cold int ff_psy_init(FFPsyContext *ctx, AVCodecContext *avctx, - int num_lens, - const uint8_t **bands, const int* num_bands) +av_cold int ff_psy_init(FFPsyContext *ctx, AVCodecContext *avctx, int num_lens, + const uint8_t **bands, const int* num_bands, + int num_groups, const uint8_t *group_map) { + int i, j, k = 0; + ctx->avctx = avctx; - ctx->psy_bands = av_mallocz(sizeof(FFPsyBand) * PSY_MAX_BANDS * avctx->channels); + ctx->ch = av_mallocz(sizeof(ctx->ch[0]) * avctx->channels * 2); + ctx->group = av_mallocz(sizeof(ctx->group[0]) * num_groups); ctx->bands = av_malloc (sizeof(ctx->bands[0]) * num_lens); ctx->num_bands = av_malloc (sizeof(ctx->num_bands[0]) * num_lens); + + if (!ctx->ch || !ctx->group || !ctx->bands || !ctx->num_bands) { + ff_psy_end(ctx); + return AVERROR(ENOMEM); + } + memcpy(ctx->bands, bands, sizeof(ctx->bands[0]) * num_lens); memcpy(ctx->num_bands, num_bands, sizeof(ctx->num_bands[0]) * num_lens); + + /* assign channels to groups (with virtual channels for coupling) */ + for (i = 0; i < num_groups; i++) { + /* NOTE: Add 1 to handle the AAC chan_config without modification. + * This has the side effect of allowing an array of 0s to map + * to one channel per group. + */ + ctx->group[i].num_ch = group_map[i] + 1; + for (j = 0; j < ctx->group[i].num_ch * 2; j++) + ctx->group[i].ch[j] = &ctx->ch[k++]; + } + switch (ctx->avctx->codec_id) { - case CODEC_ID_AAC: + case AV_CODEC_ID_AAC: ctx->model = &ff_aac_psy_model; break; } @@ -45,17 +69,14 @@ av_cold int ff_psy_init(FFPsyContext *ctx, AVCodecContext *avctx, return 0; } -FFPsyWindowInfo ff_psy_suggest_window(FFPsyContext *ctx, - const int16_t *audio, const int16_t *la, - int channel, int prev_type) +FFPsyChannelGroup *ff_psy_find_group(FFPsyContext *ctx, int channel) { - return ctx->model->window(ctx, audio, la, channel, prev_type); -} + int i = 0, ch = 0; -void ff_psy_set_band_info(FFPsyContext *ctx, int channel, - const float *coeffs, const FFPsyWindowInfo *wi) -{ - ctx->model->analyze(ctx, channel, coeffs, wi); + while (ch <= channel) + ch += ctx->group[i++].num_ch; + + return &ctx->group[i-1]; } av_cold void ff_psy_end(FFPsyContext *ctx) @@ -64,7 +85,8 @@ av_cold void ff_psy_end(FFPsyContext *ctx) ctx->model->end(ctx); av_freep(&ctx->bands); av_freep(&ctx->num_bands); - av_freep(&ctx->psy_bands); + av_freep(&ctx->group); + av_freep(&ctx->ch); } typedef struct FFPsyPreprocessContext{ @@ -82,36 +104,38 @@ av_cold struct FFPsyPreprocessContext* ff_psy_preprocess_init(AVCodecContext *av int i; float cutoff_coeff = 0; ctx = av_mallocz(sizeof(FFPsyPreprocessContext)); + if (!ctx) + return NULL; ctx->avctx = avctx; if (avctx->cutoff > 0) cutoff_coeff = 2.0 * avctx->cutoff / avctx->sample_rate; if (cutoff_coeff) - ctx->fcoeffs = ff_iir_filter_init_coeffs(FF_FILTER_TYPE_BUTTERWORTH, FF_FILTER_MODE_LOWPASS, - FILT_ORDER, cutoff_coeff, 0.0, 0.0); + ctx->fcoeffs = ff_iir_filter_init_coeffs(avctx, FF_FILTER_TYPE_BUTTERWORTH, + FF_FILTER_MODE_LOWPASS, FILT_ORDER, + cutoff_coeff, 0.0, 0.0); if (ctx->fcoeffs) { ctx->fstate = av_mallocz(sizeof(ctx->fstate[0]) * avctx->channels); + if (!ctx->fstate) { + av_free(ctx); + return NULL; + } for (i = 0; i < avctx->channels; i++) ctx->fstate[i] = ff_iir_filter_init_state(FILT_ORDER); } return ctx; } -void ff_psy_preprocess(struct FFPsyPreprocessContext *ctx, - const int16_t *audio, int16_t *dest, - int tag, int channels) +void ff_psy_preprocess(struct FFPsyPreprocessContext *ctx, float **audio, int channels) { - int ch, i; + int ch; + int frame_size = ctx->avctx->frame_size; + if (ctx->fstate) { for (ch = 0; ch < channels; ch++) - ff_iir_filter(ctx->fcoeffs, ctx->fstate[tag+ch], ctx->avctx->frame_size, - audio + ch, ctx->avctx->channels, - dest + ch, ctx->avctx->channels); - } else { - for (ch = 0; ch < channels; ch++) - for (i = 0; i < ctx->avctx->frame_size; i++) - dest[i*ctx->avctx->channels + ch] = audio[i*ctx->avctx->channels + ch]; + ff_iir_filter_flt(ctx->fcoeffs, ctx->fstate[ch], frame_size, + &audio[ch][frame_size], 1, &audio[ch][frame_size], 1); } } @@ -125,4 +149,3 @@ av_cold void ff_psy_preprocess_end(struct FFPsyPreprocessContext *ctx) av_freep(&ctx->fstate); av_free(ctx); } -