X-Git-Url: https://git.sesse.net/?a=blobdiff_plain;f=libavcodec%2Faacdec.c;h=5afc9b820e0d94d50fc913013c22b3fe7c588183;hb=033a86f9bb6fd59ca71d4951b8e2e27cdc1b29d9;hp=186de96856ecfcb0511d91d77952b96a9a01fe85;hpb=5e239c7f9e57d09c6a4c1d5762f441950f8d979c;p=ffmpeg diff --git a/libavcodec/aacdec.c b/libavcodec/aacdec.c index 186de96856e..5afc9b820e0 100644 --- a/libavcodec/aacdec.c +++ b/libavcodec/aacdec.c @@ -5,7 +5,7 @@ * * AAC LATM decoder * Copyright (c) 2008-2010 Paul Kendall - * Copyright (c) 2010 Janne Grunau + * Copyright (c) 2010 Janne Grunau * * This file is part of Libav. * @@ -79,7 +79,7 @@ Parametric Stereo. */ - +#include "libavutil/float_dsp.h" #include "avcodec.h" #include "internal.h" #include "get_bits.h" @@ -114,55 +114,6 @@ static VLC vlc_spectral[11]; static const char overread_err[] = "Input buffer exhausted before END element found\n"; -static ChannelElement *get_che(AACContext *ac, int type, int elem_id) -{ - // For PCE based channel configurations map the channels solely based on tags. - if (!ac->m4ac.chan_config) { - return ac->tag_che_map[type][elem_id]; - } - // For indexed channel configurations map the channels solely based on position. - switch (ac->m4ac.chan_config) { - case 7: - if (ac->tags_mapped == 3 && type == TYPE_CPE) { - ac->tags_mapped++; - return ac->tag_che_map[TYPE_CPE][elem_id] = ac->che[TYPE_CPE][2]; - } - case 6: - /* Some streams incorrectly code 5.1 audio as SCE[0] CPE[0] CPE[1] SCE[1] - instead of SCE[0] CPE[0] CPE[1] LFE[0]. If we seem to have - encountered such a stream, transfer the LFE[0] element to the SCE[1]'s mapping */ - if (ac->tags_mapped == tags_per_config[ac->m4ac.chan_config] - 1 && (type == TYPE_LFE || type == TYPE_SCE)) { - ac->tags_mapped++; - return ac->tag_che_map[type][elem_id] = ac->che[TYPE_LFE][0]; - } - case 5: - if (ac->tags_mapped == 2 && type == TYPE_CPE) { - ac->tags_mapped++; - return ac->tag_che_map[TYPE_CPE][elem_id] = ac->che[TYPE_CPE][1]; - } - case 4: - if (ac->tags_mapped == 2 && ac->m4ac.chan_config == 4 && type == TYPE_SCE) { - ac->tags_mapped++; - return ac->tag_che_map[TYPE_SCE][elem_id] = ac->che[TYPE_SCE][1]; - } - case 3: - case 2: - if (ac->tags_mapped == (ac->m4ac.chan_config != 2) && type == TYPE_CPE) { - ac->tags_mapped++; - return ac->tag_che_map[TYPE_CPE][elem_id] = ac->che[TYPE_CPE][0]; - } else if (ac->m4ac.chan_config == 2) { - return NULL; - } - case 1: - if (!ac->tags_mapped && type == TYPE_SCE) { - ac->tags_mapped++; - return ac->tag_che_map[TYPE_SCE][elem_id] = ac->che[TYPE_SCE][0]; - } - default: - return NULL; - } -} - static int count_channels(uint8_t (*layout)[3], int tags) { int i, sum = 0; @@ -198,10 +149,10 @@ static av_cold int che_configure(AACContext *ac, ff_aac_sbr_ctx_init(ac, &ac->che[type][id]->sbr); } if (type != TYPE_CCE) { - ac->output_data[(*channels)++] = ac->che[type][id]->ch[0].ret; + ac->output_element[(*channels)++] = &ac->che[type][id]->ch[0]; if (type == TYPE_CPE || - (type == TYPE_SCE && ac->m4ac.ps == 1)) { - ac->output_data[(*channels)++] = ac->che[type][id]->ch[1].ret; + (type == TYPE_SCE && ac->oc[1].m4ac.ps == 1)) { + ac->output_element[(*channels)++] = &ac->che[type][id]->ch[1]; } } } else { @@ -212,6 +163,38 @@ static av_cold int che_configure(AACContext *ac, return 0; } +static int frame_configure_elements(AVCodecContext *avctx) +{ + AACContext *ac = avctx->priv_data; + int type, id, ch, ret; + + /* set channel pointers to internal buffers by default */ + for (type = 0; type < 4; type++) { + for (id = 0; id < MAX_ELEM_ID; id++) { + ChannelElement *che = ac->che[type][id]; + if (che) { + che->ch[0].ret = che->ch[0].ret_buf; + che->ch[1].ret = che->ch[1].ret_buf; + } + } + } + + /* get output buffer */ + ac->frame.nb_samples = 2048; + if ((ret = ff_get_buffer(avctx, &ac->frame)) < 0) { + av_log(avctx, AV_LOG_ERROR, "get_buffer() failed\n"); + return ret; + } + + /* map output channel pointers to AVFrame data */ + for (ch = 0; ch < avctx->channels; ch++) { + if (ac->output_element[ch]) + ac->output_element[ch]->ret = (float *)ac->frame.extended_data[ch]; + } + + return 0; +} + struct elem_to_channel { uint64_t av_position; uint8_t syn_ele; @@ -220,7 +203,7 @@ struct elem_to_channel { }; static int assign_pair(struct elem_to_channel e2c_vec[MAX_ELEM_ID], - uint8_t (*layout_map)[3], int offset, int tags, uint64_t left, + uint8_t (*layout_map)[3], int offset, uint64_t left, uint64_t right, int pos) { if (layout_map[offset][0] == TYPE_CPE) { @@ -249,7 +232,7 @@ static int count_paired_channels(uint8_t (*layout_map)[3], int tags, int pos, in break; if (layout_map[i][0] == TYPE_CPE) { if (sce_parity) { - if (pos == AAC_CHANNEL_FRONT || !first_cpe) { + if (pos == AAC_CHANNEL_FRONT && !first_cpe) { sce_parity = 0; } else { return -1; @@ -272,10 +255,13 @@ static int count_paired_channels(uint8_t (*layout_map)[3], int tags, int pos, in static uint64_t sniff_channel_order(uint8_t (*layout_map)[3], int tags) { int i, n, total_non_cc_elements; - struct elem_to_channel e2c_vec[MAX_ELEM_ID] = {{ 0 }}; + struct elem_to_channel e2c_vec[4*MAX_ELEM_ID] = {{ 0 }}; int num_front_channels, num_side_channels, num_back_channels; uint64_t layout; + if (FF_ARRAY_ELEMS(e2c_vec) < tags) + return 0; + i = 0; num_front_channels = count_paired_channels(layout_map, tags, AAC_CHANNEL_FRONT, &i); @@ -299,21 +285,21 @@ static uint64_t sniff_channel_order(uint8_t (*layout_map)[3], int tags) num_front_channels--; } if (num_front_channels >= 4) { - i += assign_pair(e2c_vec, layout_map, i, tags, + i += assign_pair(e2c_vec, layout_map, i, AV_CH_FRONT_LEFT_OF_CENTER, AV_CH_FRONT_RIGHT_OF_CENTER, AAC_CHANNEL_FRONT); num_front_channels -= 2; } if (num_front_channels >= 2) { - i += assign_pair(e2c_vec, layout_map, i, tags, + i += assign_pair(e2c_vec, layout_map, i, AV_CH_FRONT_LEFT, AV_CH_FRONT_RIGHT, AAC_CHANNEL_FRONT); num_front_channels -= 2; } while (num_front_channels >= 2) { - i += assign_pair(e2c_vec, layout_map, i, tags, + i += assign_pair(e2c_vec, layout_map, i, UINT64_MAX, UINT64_MAX, AAC_CHANNEL_FRONT); @@ -321,14 +307,14 @@ static uint64_t sniff_channel_order(uint8_t (*layout_map)[3], int tags) } if (num_side_channels >= 2) { - i += assign_pair(e2c_vec, layout_map, i, tags, + i += assign_pair(e2c_vec, layout_map, i, AV_CH_SIDE_LEFT, AV_CH_SIDE_RIGHT, AAC_CHANNEL_FRONT); num_side_channels -= 2; } while (num_side_channels >= 2) { - i += assign_pair(e2c_vec, layout_map, i, tags, + i += assign_pair(e2c_vec, layout_map, i, UINT64_MAX, UINT64_MAX, AAC_CHANNEL_SIDE); @@ -336,14 +322,14 @@ static uint64_t sniff_channel_order(uint8_t (*layout_map)[3], int tags) } while (num_back_channels >= 4) { - i += assign_pair(e2c_vec, layout_map, i, tags, + i += assign_pair(e2c_vec, layout_map, i, UINT64_MAX, UINT64_MAX, AAC_CHANNEL_BACK); num_back_channels -= 2; } if (num_back_channels >= 2) { - i += assign_pair(e2c_vec, layout_map, i, tags, + i += assign_pair(e2c_vec, layout_map, i, AV_CH_BACK_LEFT, AV_CH_BACK_RIGHT, AAC_CHANNEL_BACK); @@ -396,24 +382,44 @@ static uint64_t sniff_channel_order(uint8_t (*layout_map)[3], int tags) return layout; } +/** + * Save current output configuration if and only if it has been locked. + */ +static void push_output_configuration(AACContext *ac) { + if (ac->oc[1].status == OC_LOCKED) { + ac->oc[0] = ac->oc[1]; + } + ac->oc[1].status = OC_NONE; +} + +/** + * Restore the previous output configuration if and only if the current + * configuration is unlocked. + */ +static void pop_output_configuration(AACContext *ac) { + if (ac->oc[1].status != OC_LOCKED && ac->oc[0].status != OC_NONE) { + ac->oc[1] = ac->oc[0]; + ac->avctx->channels = ac->oc[1].channels; + ac->avctx->channel_layout = ac->oc[1].channel_layout; + } +} + /** * Configure output channel order based on the current program configuration element. * - * @param che_pos current channel position configuration - * * @return Returns error status. 0 - OK, !0 - error */ -static av_cold int output_configure(AACContext *ac, - uint8_t layout_map[MAX_ELEM_ID*4][3], int tags, - int channel_config, enum OCStatus oc_type) +static int output_configure(AACContext *ac, + uint8_t layout_map[MAX_ELEM_ID*4][3], int tags, + enum OCStatus oc_type, int get_new_frame) { AVCodecContext *avctx = ac->avctx; int i, channels = 0, ret; uint64_t layout = 0; - if (ac->layout_map != layout_map) { - memcpy(ac->layout_map, layout_map, tags * sizeof(layout_map[0])); - ac->layout_map_tags = tags; + if (ac->oc[1].layout_map != layout_map) { + memcpy(ac->oc[1].layout_map, layout_map, tags * sizeof(layout_map[0])); + ac->oc[1].layout_map_tags = tags; } // Try to sniff a reasonable channel order, otherwise output the @@ -430,20 +436,133 @@ static av_cold int output_configure(AACContext *ac, if (ret < 0) return ret; } + if (ac->oc[1].m4ac.ps == 1 && channels == 2) { + if (layout == AV_CH_FRONT_CENTER) { + layout = AV_CH_FRONT_LEFT|AV_CH_FRONT_RIGHT; + } else { + layout = 0; + } + } memcpy(ac->tag_che_map, ac->che, 4 * MAX_ELEM_ID * sizeof(ac->che[0][0])); - avctx->channel_layout = layout; - avctx->channels = channels; - ac->output_configured = oc_type; + avctx->channel_layout = ac->oc[1].channel_layout = layout; + avctx->channels = ac->oc[1].channels = channels; + ac->oc[1].status = oc_type; + + if (get_new_frame) { + if ((ret = frame_configure_elements(ac->avctx)) < 0) + return ret; + } return 0; } +/** + * Set up channel positions based on a default channel configuration + * as specified in table 1.17. + * + * @return Returns error status. 0 - OK, !0 - error + */ +static int set_default_channel_config(AVCodecContext *avctx, + uint8_t (*layout_map)[3], + int *tags, + int channel_config) +{ + if (channel_config < 1 || channel_config > 7) { + av_log(avctx, AV_LOG_ERROR, "invalid default channel configuration (%d)\n", + channel_config); + return -1; + } + *tags = tags_per_config[channel_config]; + memcpy(layout_map, aac_channel_layout_map[channel_config-1], *tags * sizeof(*layout_map)); + return 0; +} + +static ChannelElement *get_che(AACContext *ac, int type, int elem_id) +{ + // For PCE based channel configurations map the channels solely based on tags. + if (!ac->oc[1].m4ac.chan_config) { + return ac->tag_che_map[type][elem_id]; + } + // Allow single CPE stereo files to be signalled with mono configuration. + if (!ac->tags_mapped && type == TYPE_CPE && ac->oc[1].m4ac.chan_config == 1) { + uint8_t layout_map[MAX_ELEM_ID*4][3]; + int layout_map_tags; + push_output_configuration(ac); + + if (set_default_channel_config(ac->avctx, layout_map, &layout_map_tags, + 2) < 0) + return NULL; + if (output_configure(ac, layout_map, layout_map_tags, + OC_TRIAL_FRAME, 1) < 0) + return NULL; + + ac->oc[1].m4ac.chan_config = 2; + ac->oc[1].m4ac.ps = 0; + } + // And vice-versa + if (!ac->tags_mapped && type == TYPE_SCE && ac->oc[1].m4ac.chan_config == 2) { + uint8_t layout_map[MAX_ELEM_ID*4][3]; + int layout_map_tags; + push_output_configuration(ac); + + if (set_default_channel_config(ac->avctx, layout_map, &layout_map_tags, + 1) < 0) + return NULL; + if (output_configure(ac, layout_map, layout_map_tags, + OC_TRIAL_FRAME, 1) < 0) + return NULL; + + ac->oc[1].m4ac.chan_config = 1; + if (ac->oc[1].m4ac.sbr) + ac->oc[1].m4ac.ps = -1; + } + // For indexed channel configurations map the channels solely based on position. + switch (ac->oc[1].m4ac.chan_config) { + case 7: + if (ac->tags_mapped == 3 && type == TYPE_CPE) { + ac->tags_mapped++; + return ac->tag_che_map[TYPE_CPE][elem_id] = ac->che[TYPE_CPE][2]; + } + case 6: + /* Some streams incorrectly code 5.1 audio as SCE[0] CPE[0] CPE[1] SCE[1] + instead of SCE[0] CPE[0] CPE[1] LFE[0]. If we seem to have + encountered such a stream, transfer the LFE[0] element to the SCE[1]'s mapping */ + if (ac->tags_mapped == tags_per_config[ac->oc[1].m4ac.chan_config] - 1 && (type == TYPE_LFE || type == TYPE_SCE)) { + ac->tags_mapped++; + return ac->tag_che_map[type][elem_id] = ac->che[TYPE_LFE][0]; + } + case 5: + if (ac->tags_mapped == 2 && type == TYPE_CPE) { + ac->tags_mapped++; + return ac->tag_che_map[TYPE_CPE][elem_id] = ac->che[TYPE_CPE][1]; + } + case 4: + if (ac->tags_mapped == 2 && ac->oc[1].m4ac.chan_config == 4 && type == TYPE_SCE) { + ac->tags_mapped++; + return ac->tag_che_map[TYPE_SCE][elem_id] = ac->che[TYPE_SCE][1]; + } + case 3: + case 2: + if (ac->tags_mapped == (ac->oc[1].m4ac.chan_config != 2) && type == TYPE_CPE) { + ac->tags_mapped++; + return ac->tag_che_map[TYPE_CPE][elem_id] = ac->che[TYPE_CPE][0]; + } else if (ac->oc[1].m4ac.chan_config == 2) { + return NULL; + } + case 1: + if (!ac->tags_mapped && type == TYPE_SCE) { + ac->tags_mapped++; + return ac->tag_che_map[TYPE_SCE][elem_id] = ac->che[TYPE_SCE][0]; + } + default: + return NULL; + } +} + /** * Decode an array of 4 bit element IDs, optionally interleaved with a stereo/mono switching bit. * - * @param cpe_map Stereo (Channel Pair Element) map, NULL if stereo bit is not present. - * @param sce_map mono (Single Channel Element) map * @param type speaker type/position for these channels */ static void decode_channel_map(uint8_t layout_map[][3], @@ -533,27 +652,6 @@ static int decode_pce(AVCodecContext *avctx, MPEG4AudioConfig *m4ac, return tags; } -/** - * Set up channel positions based on a default channel configuration - * as specified in table 1.17. - * - * @return Returns error status. 0 - OK, !0 - error - */ -static av_cold int set_default_channel_config(AVCodecContext *avctx, - uint8_t (*layout_map)[3], - int *tags, - int channel_config) -{ - if (channel_config < 1 || channel_config > 7) { - av_log(avctx, AV_LOG_ERROR, "invalid default channel configuration (%d)\n", - channel_config); - return -1; - } - *tags = tags_per_config[channel_config]; - memcpy(layout_map, aac_channel_layout_map[channel_config-1], *tags * sizeof(*layout_map)); - return 0; -} - /** * Decode GA "General Audio" specific configuration; reference: table 4.1. * @@ -572,8 +670,8 @@ static int decode_ga_specific_config(AACContext *ac, AVCodecContext *avctx, int tags = 0; if (get_bits1(gb)) { // frameLengthFlag - av_log_missing_feature(avctx, "960/120 MDCT window is", 1); - return -1; + av_log_missing_feature(avctx, "960/120 MDCT window", 1); + return AVERROR_PATCHWELCOME; } if (get_bits1(gb)) // dependsOnCoreCoder @@ -599,8 +697,7 @@ static int decode_ga_specific_config(AACContext *ac, AVCodecContext *avctx, } else if (m4ac->sbr == 1 && m4ac->ps == -1) m4ac->ps = 1; - if (ac && (ret = output_configure(ac, layout_map, tags, - channel_config, OC_GLOBAL_HDR))) + if (ac && (ret = output_configure(ac, layout_map, tags, OC_GLOBAL_HDR, 0))) return ret; if (extension_flag) { @@ -690,7 +787,8 @@ static int decode_audio_specific_config(AACContext *ac, */ static av_always_inline int lcg_random(int previous_val) { - return previous_val * 1664525 + 1013904223; + union { unsigned u; int s; } v = { previous_val * 1664525u + 1013904223 }; + return v.s; } static av_always_inline void reset_predict_state(PredictorState *ps) @@ -742,13 +840,14 @@ static void reset_predictor_group(PredictorState *ps, int group_num) static av_cold int aac_decode_init(AVCodecContext *avctx) { AACContext *ac = avctx->priv_data; - float output_scale_factor; ac->avctx = avctx; - ac->m4ac.sample_rate = avctx->sample_rate; + ac->oc[1].m4ac.sample_rate = avctx->sample_rate; + + avctx->sample_fmt = AV_SAMPLE_FMT_FLTP; if (avctx->extradata_size > 0) { - if (decode_audio_specific_config(ac, ac->avctx, &ac->m4ac, + if (decode_audio_specific_config(ac, ac->avctx, &ac->oc[1].m4ac, avctx->extradata, avctx->extradata_size*8, 1) < 0) return -1; @@ -758,10 +857,10 @@ static av_cold int aac_decode_init(AVCodecContext *avctx) int layout_map_tags; sr = sample_rate_idx(avctx->sample_rate); - ac->m4ac.sampling_index = sr; - ac->m4ac.channels = avctx->channels; - ac->m4ac.sbr = -1; - ac->m4ac.ps = -1; + ac->oc[1].m4ac.sampling_index = sr; + ac->oc[1].m4ac.channels = avctx->channels; + ac->oc[1].m4ac.sbr = -1; + ac->oc[1].m4ac.ps = -1; for (i = 0; i < FF_ARRAY_ELEMS(ff_mpeg4audio_channels); i++) if (ff_mpeg4audio_channels[i] == avctx->channels) @@ -769,27 +868,19 @@ static av_cold int aac_decode_init(AVCodecContext *avctx) if (i == FF_ARRAY_ELEMS(ff_mpeg4audio_channels)) { i = 0; } - ac->m4ac.chan_config = i; + ac->oc[1].m4ac.chan_config = i; - if (ac->m4ac.chan_config) { + if (ac->oc[1].m4ac.chan_config) { int ret = set_default_channel_config(avctx, layout_map, - &layout_map_tags, ac->m4ac.chan_config); + &layout_map_tags, ac->oc[1].m4ac.chan_config); if (!ret) output_configure(ac, layout_map, layout_map_tags, - ac->m4ac.chan_config, OC_GLOBAL_HDR); + OC_GLOBAL_HDR, 0); else if (avctx->err_recognition & AV_EF_EXPLODE) return AVERROR_INVALIDDATA; } } - if (avctx->request_sample_fmt == AV_SAMPLE_FMT_FLT) { - avctx->sample_fmt = AV_SAMPLE_FMT_FLT; - output_scale_factor = 1.0 / 32768.0; - } else { - avctx->sample_fmt = AV_SAMPLE_FMT_S16; - output_scale_factor = 1.0; - } - AAC_INIT_VLC_STATIC( 0, 304); AAC_INIT_VLC_STATIC( 1, 270); AAC_INIT_VLC_STATIC( 2, 550); @@ -804,8 +895,8 @@ static av_cold int aac_decode_init(AVCodecContext *avctx) ff_aac_sbr_init(); - ff_dsputil_init(&ac->dsp, avctx); ff_fmt_convert_init(&ac->fmt_conv, avctx); + avpriv_float_dsp_init(&ac->fdsp, avctx->flags & CODEC_FLAG_BITEXACT); ac->random_state = 0x1f2e3d4c; @@ -816,9 +907,9 @@ static av_cold int aac_decode_init(AVCodecContext *avctx) ff_aac_scalefactor_code, sizeof(ff_aac_scalefactor_code[0]), sizeof(ff_aac_scalefactor_code[0]), 352); - ff_mdct_init(&ac->mdct, 11, 1, output_scale_factor/1024.0); - ff_mdct_init(&ac->mdct_small, 8, 1, output_scale_factor/128.0); - ff_mdct_init(&ac->mdct_ltp, 11, 0, -2.0/output_scale_factor); + ff_mdct_init(&ac->mdct, 11, 1, 1.0 / (32768.0 * 1024.0)); + ff_mdct_init(&ac->mdct_small, 8, 1, 1.0 / (32768.0 * 128.0)); + ff_mdct_init(&ac->mdct_ltp, 11, 0, -2.0 * 32768.0); // window initialization ff_kbd_window_init(ff_aac_kbd_long_1024, 4.0, 1024); ff_kbd_window_init(ff_aac_kbd_short_128, 6.0, 128); @@ -864,7 +955,7 @@ static int decode_prediction(AACContext *ac, IndividualChannelStream *ics, return -1; } } - for (sfb = 0; sfb < FFMIN(ics->max_sfb, ff_aac_pred_sfb_max[ac->m4ac.sampling_index]); sfb++) { + for (sfb = 0; sfb < FFMIN(ics->max_sfb, ff_aac_pred_sfb_max[ac->oc[1].m4ac.sampling_index]); sfb++) { ics->prediction_used[sfb] = get_bits1(gb); } return 0; @@ -873,7 +964,7 @@ static int decode_prediction(AACContext *ac, IndividualChannelStream *ics, /** * Decode Long Term Prediction data; reference: table 4.xx. */ -static void decode_ltp(AACContext *ac, LongTermPrediction *ltp, +static void decode_ltp(LongTermPrediction *ltp, GetBitContext *gb, uint8_t max_sfb) { int sfb; @@ -912,29 +1003,29 @@ static int decode_ics_info(AACContext *ac, IndividualChannelStream *ics, } } ics->num_windows = 8; - ics->swb_offset = ff_swb_offset_128[ac->m4ac.sampling_index]; - ics->num_swb = ff_aac_num_swb_128[ac->m4ac.sampling_index]; - ics->tns_max_bands = ff_tns_max_bands_128[ac->m4ac.sampling_index]; + ics->swb_offset = ff_swb_offset_128[ac->oc[1].m4ac.sampling_index]; + ics->num_swb = ff_aac_num_swb_128[ac->oc[1].m4ac.sampling_index]; + ics->tns_max_bands = ff_tns_max_bands_128[ac->oc[1].m4ac.sampling_index]; ics->predictor_present = 0; } else { ics->max_sfb = get_bits(gb, 6); ics->num_windows = 1; - ics->swb_offset = ff_swb_offset_1024[ac->m4ac.sampling_index]; - ics->num_swb = ff_aac_num_swb_1024[ac->m4ac.sampling_index]; - ics->tns_max_bands = ff_tns_max_bands_1024[ac->m4ac.sampling_index]; + ics->swb_offset = ff_swb_offset_1024[ac->oc[1].m4ac.sampling_index]; + ics->num_swb = ff_aac_num_swb_1024[ac->oc[1].m4ac.sampling_index]; + ics->tns_max_bands = ff_tns_max_bands_1024[ac->oc[1].m4ac.sampling_index]; ics->predictor_present = get_bits1(gb); ics->predictor_reset_group = 0; if (ics->predictor_present) { - if (ac->m4ac.object_type == AOT_AAC_MAIN) { + if (ac->oc[1].m4ac.object_type == AOT_AAC_MAIN) { if (decode_prediction(ac, ics, gb)) { return AVERROR_INVALIDDATA; } - } else if (ac->m4ac.object_type == AOT_AAC_LC) { + } else if (ac->oc[1].m4ac.object_type == AOT_AAC_LC) { av_log(ac->avctx, AV_LOG_ERROR, "Prediction is not allowed in AAC-LC.\n"); return AVERROR_INVALIDDATA; } else { if ((ics->ltp.present = get_bits(gb, 1))) - decode_ltp(ac, &ics->ltp, gb, ics->max_sfb); + decode_ltp(&ics->ltp, gb, ics->max_sfb); } } } @@ -1100,7 +1191,7 @@ static int decode_tns(AACContext *ac, TemporalNoiseShaping *tns, { int w, filt, i, coef_len, coef_res, coef_compress; const int is8 = ics->window_sequence[0] == EIGHT_SHORT_SEQUENCE; - const int tns_max_order = is8 ? 7 : ac->m4ac.object_type == AOT_AAC_MAIN ? 20 : 12; + const int tns_max_order = is8 ? 7 : ac->oc[1].m4ac.object_type == AOT_AAC_MAIN ? 20 : 12; for (w = 0; w < ics->num_windows; w++) { if ((tns->n_filt[w] = get_bits(gb, 2 - is8))) { coef_res = get_bits1(gb); @@ -1209,7 +1300,7 @@ static inline float *VMUL4S(float *dst, const float *v, unsigned idx, t.i = s.i ^ (sign & 1U<<31); *dst++ = v[idx>>4 & 3] * t.f; - sign <<= nz & 1; nz >>= 1; + sign <<= nz & 1; t.i = s.i ^ (sign & 1U<<31); *dst++ = v[idx>>6 & 3] * t.f; @@ -1266,9 +1357,9 @@ static int decode_spectrum_and_dequant(AACContext *ac, float coef[1024], cfo[k] = ac->random_state; } - band_energy = ac->dsp.scalarproduct_float(cfo, cfo, off_len); + band_energy = ac->fdsp.scalarproduct_float(cfo, cfo, off_len); scale = sf[idx] / sqrtf(band_energy); - ac->dsp.vector_fmul_scalar(cfo, cfo, scale, off_len); + ac->fdsp.vector_fmul_scalar(cfo, cfo, scale, off_len); } } else { const float *vq = ff_aac_codebook_vector_vals[cbt_m1]; @@ -1414,7 +1505,7 @@ static int decode_spectrum_and_dequant(AACContext *ac, float coef[1024], } } while (len -= 2); - ac->dsp.vector_fmul_scalar(cfo, cfo, sf[idx], off_len); + ac->fdsp.vector_fmul_scalar(cfo, cfo, sf[idx], off_len); } } @@ -1511,7 +1602,7 @@ static void apply_prediction(AACContext *ac, SingleChannelElement *sce) } if (sce->ics.window_sequence[0] != EIGHT_SHORT_SEQUENCE) { - for (sfb = 0; sfb < ff_aac_pred_sfb_max[ac->m4ac.sampling_index]; sfb++) { + for (sfb = 0; sfb < ff_aac_pred_sfb_max[ac->oc[1].m4ac.sampling_index]; sfb++) { for (k = sce->ics.swb_offset[sfb]; k < sce->ics.swb_offset[sfb + 1]; k++) { predict(&sce->predictor_state[k], &sce->coeffs[k], sce->ics.predictor_present && sce->ics.prediction_used[sfb]); @@ -1573,14 +1664,14 @@ static int decode_ics(AACContext *ac, SingleChannelElement *sce, return -1; if (get_bits1(gb)) { av_log_missing_feature(ac->avctx, "SSR", 1); - return -1; + return AVERROR_PATCHWELCOME; } } if (decode_spectrum_and_dequant(ac, out, gb, sce->sf, pulse_present, &pulse, ics, sce->band_type) < 0) return -1; - if (ac->m4ac.object_type == AOT_AAC_MAIN && !common_window) + if (ac->oc[1].m4ac.object_type == AOT_AAC_MAIN && !common_window) apply_prediction(ac, sce); return 0; @@ -1601,9 +1692,9 @@ static void apply_mid_side_stereo(AACContext *ac, ChannelElement *cpe) if (cpe->ms_mask[idx] && cpe->ch[0].band_type[idx] < NOISE_BT && cpe->ch[1].band_type[idx] < NOISE_BT) { for (group = 0; group < ics->group_len[g]; group++) { - ac->dsp.butterflies_float(ch0 + group * 128 + offsets[i], - ch1 + group * 128 + offsets[i], - offsets[i+1] - offsets[i]); + ac->fdsp.butterflies_float(ch0 + group * 128 + offsets[i], + ch1 + group * 128 + offsets[i], + offsets[i+1] - offsets[i]); } } } @@ -1638,10 +1729,10 @@ static void apply_intensity_stereo(AACContext *ac, ChannelElement *cpe, int ms_p c *= 1 - 2 * cpe->ms_mask[idx]; scale = c * sce1->sf[idx]; for (group = 0; group < ics->group_len[g]; group++) - ac->dsp.vector_fmul_scalar(coef1 + group * 128 + offsets[i], - coef0 + group * 128 + offsets[i], - scale, - offsets[i + 1] - offsets[i]); + ac->fdsp.vector_fmul_scalar(coef1 + group * 128 + offsets[i], + coef0 + group * 128 + offsets[i], + scale, + offsets[i + 1] - offsets[i]); } } else { int bt_run_end = sce1->band_type_run_end[idx]; @@ -1670,9 +1761,9 @@ static int decode_cpe(AACContext *ac, GetBitContext *gb, ChannelElement *cpe) i = cpe->ch[1].ics.use_kb_window[0]; cpe->ch[1].ics = cpe->ch[0].ics; cpe->ch[1].ics.use_kb_window[1] = i; - if (cpe->ch[1].ics.predictor_present && (ac->m4ac.object_type != AOT_AAC_MAIN)) + if (cpe->ch[1].ics.predictor_present && (ac->oc[1].m4ac.object_type != AOT_AAC_MAIN)) if ((cpe->ch[1].ics.ltp.present = get_bits(gb, 1))) - decode_ltp(ac, &cpe->ch[1].ics.ltp, gb, cpe->ch[1].ics.max_sfb); + decode_ltp(&cpe->ch[1].ics.ltp, gb, cpe->ch[1].ics.max_sfb); ms_present = get_bits(gb, 2); if (ms_present == 3) { av_log(ac->avctx, AV_LOG_ERROR, "ms_present = 3 is reserved.\n"); @@ -1688,7 +1779,7 @@ static int decode_cpe(AACContext *ac, GetBitContext *gb, ChannelElement *cpe) if (common_window) { if (ms_present) apply_mid_side_stereo(ac, cpe); - if (ac->m4ac.object_type == AOT_AAC_MAIN) { + if (ac->oc[1].m4ac.object_type == AOT_AAC_MAIN) { apply_prediction(ac, &cpe->ch[0]); apply_prediction(ac, &cpe->ch[1]); } @@ -1799,12 +1890,10 @@ static int decode_drc_channel_exclusions(DynamicRangeControl *che_drc, /** * Decode dynamic range information; reference: table 4.52. * - * @param cnt length of TYPE_FIL syntactic element in bytes - * * @return Returns number of bytes consumed. */ static int decode_dynamic_range(DynamicRangeControl *che_drc, - GetBitContext *gb, int cnt) + GetBitContext *gb) { int n = 1; int drc_num_bands = 1; @@ -1869,26 +1958,26 @@ static int decode_extension_payload(AACContext *ac, GetBitContext *gb, int cnt, if (!che) { av_log(ac->avctx, AV_LOG_ERROR, "SBR was found before the first channel element.\n"); return res; - } else if (!ac->m4ac.sbr) { + } else if (!ac->oc[1].m4ac.sbr) { av_log(ac->avctx, AV_LOG_ERROR, "SBR signaled to be not-present but was found in the bitstream.\n"); skip_bits_long(gb, 8 * cnt - 4); return res; - } else if (ac->m4ac.sbr == -1 && ac->output_configured == OC_LOCKED) { + } else if (ac->oc[1].m4ac.sbr == -1 && ac->oc[1].status == OC_LOCKED) { av_log(ac->avctx, AV_LOG_ERROR, "Implicit SBR was found with a first occurrence after the first frame.\n"); skip_bits_long(gb, 8 * cnt - 4); return res; - } else if (ac->m4ac.ps == -1 && ac->output_configured < OC_LOCKED && ac->avctx->channels == 1) { - ac->m4ac.sbr = 1; - ac->m4ac.ps = 1; - output_configure(ac, ac->layout_map, ac->layout_map_tags, - ac->m4ac.chan_config, ac->output_configured); + } else if (ac->oc[1].m4ac.ps == -1 && ac->oc[1].status < OC_LOCKED && ac->avctx->channels == 1) { + ac->oc[1].m4ac.sbr = 1; + ac->oc[1].m4ac.ps = 1; + output_configure(ac, ac->oc[1].layout_map, ac->oc[1].layout_map_tags, + ac->oc[1].status, 1); } else { - ac->m4ac.sbr = 1; + ac->oc[1].m4ac.sbr = 1; } res = ff_decode_sbr_extension(ac, &che->sbr, gb, crc_flag, cnt, elem_type); break; case EXT_DYNAMIC_RANGE: - res = decode_dynamic_range(&ac->che_drc, gb, cnt); + res = decode_dynamic_range(&ac->che_drc, gb); break; case EXT_FILL: case EXT_FILL_DATA: @@ -1913,7 +2002,7 @@ static void apply_tns(float coef[1024], TemporalNoiseShaping *tns, int w, filt, m, i; int bottom, top, order, start, end, size, inc; float lpc[TNS_MAX_ORDER]; - float tmp[TNS_MAX_ORDER]; + float tmp[TNS_MAX_ORDER + 1]; for (w = 0; w < ics->num_windows; w++) { bottom = ics->num_swb; @@ -1971,15 +2060,15 @@ static void windowing_and_mdct_ltp(AACContext *ac, float *out, const float *swindow_prev = ics->use_kb_window[1] ? ff_aac_kbd_short_128 : ff_sine_128; if (ics->window_sequence[0] != LONG_STOP_SEQUENCE) { - ac->dsp.vector_fmul(in, in, lwindow_prev, 1024); + ac->fdsp.vector_fmul(in, in, lwindow_prev, 1024); } else { memset(in, 0, 448 * sizeof(float)); - ac->dsp.vector_fmul(in + 448, in + 448, swindow_prev, 128); + ac->fdsp.vector_fmul(in + 448, in + 448, swindow_prev, 128); } if (ics->window_sequence[0] != LONG_START_SEQUENCE) { - ac->dsp.vector_fmul_reverse(in + 1024, in + 1024, lwindow, 1024); + ac->fdsp.vector_fmul_reverse(in + 1024, in + 1024, lwindow, 1024); } else { - ac->dsp.vector_fmul_reverse(in + 1024 + 448, in + 1024 + 448, swindow, 128); + ac->fdsp.vector_fmul_reverse(in + 1024 + 448, in + 1024 + 448, swindow, 128); memset(in + 1024 + 576, 0, 448 * sizeof(float)); } ac->mdct_ltp.mdct_calc(&ac->mdct_ltp, out, in); @@ -2032,17 +2121,17 @@ static void update_ltp(AACContext *ac, SingleChannelElement *sce) if (ics->window_sequence[0] == EIGHT_SHORT_SEQUENCE) { memcpy(saved_ltp, saved, 512 * sizeof(float)); memset(saved_ltp + 576, 0, 448 * sizeof(float)); - ac->dsp.vector_fmul_reverse(saved_ltp + 448, ac->buf_mdct + 960, &swindow[64], 64); + ac->fdsp.vector_fmul_reverse(saved_ltp + 448, ac->buf_mdct + 960, &swindow[64], 64); for (i = 0; i < 64; i++) saved_ltp[i + 512] = ac->buf_mdct[1023 - i] * swindow[63 - i]; } else if (ics->window_sequence[0] == LONG_START_SEQUENCE) { memcpy(saved_ltp, ac->buf_mdct + 512, 448 * sizeof(float)); memset(saved_ltp + 576, 0, 448 * sizeof(float)); - ac->dsp.vector_fmul_reverse(saved_ltp + 448, ac->buf_mdct + 960, &swindow[64], 64); + ac->fdsp.vector_fmul_reverse(saved_ltp + 448, ac->buf_mdct + 960, &swindow[64], 64); for (i = 0; i < 64; i++) saved_ltp[i + 512] = ac->buf_mdct[1023 - i] * swindow[63 - i]; } else { // LONG_STOP or ONLY_LONG - ac->dsp.vector_fmul_reverse(saved_ltp, ac->buf_mdct + 512, &lwindow[512], 512); + ac->fdsp.vector_fmul_reverse(saved_ltp, ac->buf_mdct + 512, &lwindow[512], 512); for (i = 0; i < 512; i++) saved_ltp[i + 512] = ac->buf_mdct[1023 - i] * lwindow[511 - i]; } @@ -2083,35 +2172,35 @@ static void imdct_and_windowing(AACContext *ac, SingleChannelElement *sce) */ if ((ics->window_sequence[1] == ONLY_LONG_SEQUENCE || ics->window_sequence[1] == LONG_STOP_SEQUENCE) && (ics->window_sequence[0] == ONLY_LONG_SEQUENCE || ics->window_sequence[0] == LONG_START_SEQUENCE)) { - ac->dsp.vector_fmul_window( out, saved, buf, lwindow_prev, 512); + ac->fdsp.vector_fmul_window( out, saved, buf, lwindow_prev, 512); } else { - memcpy( out, saved, 448 * sizeof(float)); + memcpy( out, saved, 448 * sizeof(float)); if (ics->window_sequence[0] == EIGHT_SHORT_SEQUENCE) { - ac->dsp.vector_fmul_window(out + 448 + 0*128, saved + 448, buf + 0*128, swindow_prev, 64); - ac->dsp.vector_fmul_window(out + 448 + 1*128, buf + 0*128 + 64, buf + 1*128, swindow, 64); - ac->dsp.vector_fmul_window(out + 448 + 2*128, buf + 1*128 + 64, buf + 2*128, swindow, 64); - ac->dsp.vector_fmul_window(out + 448 + 3*128, buf + 2*128 + 64, buf + 3*128, swindow, 64); - ac->dsp.vector_fmul_window(temp, buf + 3*128 + 64, buf + 4*128, swindow, 64); - memcpy( out + 448 + 4*128, temp, 64 * sizeof(float)); + ac->fdsp.vector_fmul_window(out + 448 + 0*128, saved + 448, buf + 0*128, swindow_prev, 64); + ac->fdsp.vector_fmul_window(out + 448 + 1*128, buf + 0*128 + 64, buf + 1*128, swindow, 64); + ac->fdsp.vector_fmul_window(out + 448 + 2*128, buf + 1*128 + 64, buf + 2*128, swindow, 64); + ac->fdsp.vector_fmul_window(out + 448 + 3*128, buf + 2*128 + 64, buf + 3*128, swindow, 64); + ac->fdsp.vector_fmul_window(temp, buf + 3*128 + 64, buf + 4*128, swindow, 64); + memcpy( out + 448 + 4*128, temp, 64 * sizeof(float)); } else { - ac->dsp.vector_fmul_window(out + 448, saved + 448, buf, swindow_prev, 64); - memcpy( out + 576, buf + 64, 448 * sizeof(float)); + ac->fdsp.vector_fmul_window(out + 448, saved + 448, buf, swindow_prev, 64); + memcpy( out + 576, buf + 64, 448 * sizeof(float)); } } // buffer update if (ics->window_sequence[0] == EIGHT_SHORT_SEQUENCE) { - memcpy( saved, temp + 64, 64 * sizeof(float)); - ac->dsp.vector_fmul_window(saved + 64, buf + 4*128 + 64, buf + 5*128, swindow, 64); - ac->dsp.vector_fmul_window(saved + 192, buf + 5*128 + 64, buf + 6*128, swindow, 64); - ac->dsp.vector_fmul_window(saved + 320, buf + 6*128 + 64, buf + 7*128, swindow, 64); - memcpy( saved + 448, buf + 7*128 + 64, 64 * sizeof(float)); + memcpy( saved, temp + 64, 64 * sizeof(float)); + ac->fdsp.vector_fmul_window(saved + 64, buf + 4*128 + 64, buf + 5*128, swindow, 64); + ac->fdsp.vector_fmul_window(saved + 192, buf + 5*128 + 64, buf + 6*128, swindow, 64); + ac->fdsp.vector_fmul_window(saved + 320, buf + 6*128 + 64, buf + 7*128, swindow, 64); + memcpy( saved + 448, buf + 7*128 + 64, 64 * sizeof(float)); } else if (ics->window_sequence[0] == LONG_START_SEQUENCE) { - memcpy( saved, buf + 512, 448 * sizeof(float)); - memcpy( saved + 448, buf + 7*128 + 64, 64 * sizeof(float)); + memcpy( saved, buf + 512, 448 * sizeof(float)); + memcpy( saved + 448, buf + 7*128 + 64, 64 * sizeof(float)); } else { // LONG_STOP or ONLY_LONG - memcpy( saved, buf + 512, 512 * sizeof(float)); + memcpy( saved, buf + 512, 512 * sizeof(float)); } } @@ -2129,7 +2218,7 @@ static void apply_dependent_coupling(AACContext *ac, float *dest = target->coeffs; const float *src = cce->ch[0].coeffs; int g, i, group, k, idx = 0; - if (ac->m4ac.object_type == AOT_AAC_LTP) { + if (ac->oc[1].m4ac.object_type == AOT_AAC_LTP) { av_log(ac->avctx, AV_LOG_ERROR, "Dependent coupling is not supported together with LTP\n"); return; @@ -2164,7 +2253,7 @@ static void apply_independent_coupling(AACContext *ac, const float gain = cce->coup.gain[index][0]; const float *src = cce->ch[0].ret; float *dest = target->ret; - const int len = 1024 << (ac->m4ac.sbr == 1); + const int len = 1024 << (ac->oc[1].m4ac.sbr == 1); for (i = 0; i < len; i++) dest[i] += gain * src[i]; @@ -2217,7 +2306,7 @@ static void spectral_to_sample(AACContext *ac) if (che) { if (type <= TYPE_CPE) apply_channel_coupling(ac, che, type, i, BEFORE_TNS, apply_dependent_coupling); - if (ac->m4ac.object_type == AOT_AAC_LTP) { + if (ac->oc[1].m4ac.object_type == AOT_AAC_LTP) { if (che->ch[0].ics.predictor_present) { if (che->ch[0].ics.ltp.present) apply_ltp(ac, &che->ch[0]); @@ -2233,14 +2322,14 @@ static void spectral_to_sample(AACContext *ac) apply_channel_coupling(ac, che, type, i, BETWEEN_TNS_AND_IMDCT, apply_dependent_coupling); if (type != TYPE_CCE || che->coup.coupling_point == AFTER_IMDCT) { imdct_and_windowing(ac, &che->ch[0]); - if (ac->m4ac.object_type == AOT_AAC_LTP) + if (ac->oc[1].m4ac.object_type == AOT_AAC_LTP) update_ltp(ac, &che->ch[0]); if (type == TYPE_CPE) { imdct_and_windowing(ac, &che->ch[1]); - if (ac->m4ac.object_type == AOT_AAC_LTP) + if (ac->oc[1].m4ac.object_type == AOT_AAC_LTP) update_ltp(ac, &che->ch[1]); } - if (ac->m4ac.sbr > 0) { + if (ac->oc[1].m4ac.sbr > 0) { ff_sbr_apply(ac, &che->sbr, type, che->ch[0].ret, che->ch[1].ret); } } @@ -2260,35 +2349,33 @@ static int parse_adts_frame_header(AACContext *ac, GetBitContext *gb) size = avpriv_aac_parse_header(gb, &hdr_info); if (size > 0) { + if (hdr_info.num_aac_frames != 1) { + av_log_missing_feature(ac->avctx, "More than one AAC RDB per ADTS frame", 0); + return AVERROR_PATCHWELCOME; + } + push_output_configuration(ac); if (hdr_info.chan_config) { - ac->m4ac.chan_config = hdr_info.chan_config; + ac->oc[1].m4ac.chan_config = hdr_info.chan_config; if (set_default_channel_config(ac->avctx, layout_map, &layout_map_tags, hdr_info.chan_config)) return -7; if (output_configure(ac, layout_map, layout_map_tags, - hdr_info.chan_config, - FFMAX(ac->output_configured, OC_TRIAL_FRAME))) + FFMAX(ac->oc[1].status, OC_TRIAL_FRAME), 0)) return -7; - } else if (ac->output_configured != OC_LOCKED) { - ac->m4ac.chan_config = 0; - ac->output_configured = OC_NONE; - } - if (ac->output_configured != OC_LOCKED) { - ac->m4ac.sbr = -1; - ac->m4ac.ps = -1; - ac->m4ac.sample_rate = hdr_info.sample_rate; - ac->m4ac.sampling_index = hdr_info.sampling_index; - ac->m4ac.object_type = hdr_info.object_type; - } - if (!ac->avctx->sample_rate) - ac->avctx->sample_rate = hdr_info.sample_rate; - if (hdr_info.num_aac_frames == 1) { - if (!hdr_info.crc_absent) - skip_bits(gb, 16); } else { - av_log_missing_feature(ac->avctx, "More than one AAC RDB per ADTS frame is", 0); - return -1; + ac->oc[1].m4ac.chan_config = 0; + } + ac->oc[1].m4ac.sample_rate = hdr_info.sample_rate; + ac->oc[1].m4ac.sampling_index = hdr_info.sampling_index; + ac->oc[1].m4ac.object_type = hdr_info.object_type; + if (ac->oc[0].status != OC_LOCKED || + ac->oc[0].m4ac.chan_config != hdr_info.chan_config || + ac->oc[0].m4ac.sample_rate != hdr_info.sample_rate) { + ac->oc[1].m4ac.sbr = -1; + ac->oc[1].m4ac.ps = -1; } + if (!hdr_info.crc_absent) + skip_bits(gb, 16); } return size; } @@ -2300,19 +2387,26 @@ static int aac_decode_frame_int(AVCodecContext *avctx, void *data, ChannelElement *che = NULL, *che_prev = NULL; enum RawDataBlockType elem_type, elem_type_prev = TYPE_END; int err, elem_id; - int samples = 0, multiplier, audio_found = 0; + int samples = 0, multiplier, audio_found = 0, pce_found = 0; if (show_bits(gb, 12) == 0xfff) { if (parse_adts_frame_header(ac, gb) < 0) { av_log(avctx, AV_LOG_ERROR, "Error decoding AAC frame header.\n"); - return -1; + err = -1; + goto fail; } - if (ac->m4ac.sampling_index > 12) { - av_log(ac->avctx, AV_LOG_ERROR, "invalid sampling rate index %d\n", ac->m4ac.sampling_index); - return -1; + if (ac->oc[1].m4ac.sampling_index > 12) { + av_log(ac->avctx, AV_LOG_ERROR, "invalid sampling rate index %d\n", ac->oc[1].m4ac.sampling_index); + err = -1; + goto fail; } } + if (frame_configure_elements(avctx) < 0) { + err = -1; + goto fail; + } + ac->tags_mapped = 0; // parse while ((elem_type = get_bits(gb, 3)) != TYPE_END) { @@ -2322,7 +2416,8 @@ static int aac_decode_frame_int(AVCodecContext *avctx, void *data, if (!(che=get_che(ac, elem_type, elem_id))) { av_log(ac->avctx, AV_LOG_ERROR, "channel element %d.%d is not allocated\n", elem_type, elem_id); - return -1; + err = -1; + goto fail; } samples = 1024; } @@ -2355,16 +2450,20 @@ static int aac_decode_frame_int(AVCodecContext *avctx, void *data, case TYPE_PCE: { uint8_t layout_map[MAX_ELEM_ID*4][3]; int tags; - tags = decode_pce(avctx, &ac->m4ac, layout_map, gb); + push_output_configuration(ac); + tags = decode_pce(avctx, &ac->oc[1].m4ac, layout_map, gb); if (tags < 0) { err = tags; break; } - if (ac->output_configured > OC_TRIAL_PCE) + if (pce_found) { av_log(avctx, AV_LOG_ERROR, "Not evaluating a further program_config_element as this construct is dubious at best.\n"); - else - err = output_configure(ac, layout_map, tags, 0, OC_TRIAL_PCE); + pop_output_configuration(ac); + } else { + err = output_configure(ac, layout_map, tags, OC_TRIAL_PCE, 1); + pce_found = 1; + } break; } @@ -2373,7 +2472,8 @@ static int aac_decode_frame_int(AVCodecContext *avctx, void *data, elem_id += get_bits(gb, 8) - 1; if (get_bits_left(gb) < 8 * elem_id) { av_log(avctx, AV_LOG_ERROR, overread_err); - return -1; + err = -1; + goto fail; } while (elem_id > 0) elem_id -= decode_extension_payload(ac, gb, elem_id, che_prev, elem_type_prev); @@ -2389,48 +2489,36 @@ static int aac_decode_frame_int(AVCodecContext *avctx, void *data, elem_type_prev = elem_type; if (err) - return err; + goto fail; if (get_bits_left(gb) < 3) { av_log(avctx, AV_LOG_ERROR, overread_err); - return -1; + err = -1; + goto fail; } } spectral_to_sample(ac); - multiplier = (ac->m4ac.sbr == 1) ? ac->m4ac.ext_sample_rate > ac->m4ac.sample_rate : 0; + multiplier = (ac->oc[1].m4ac.sbr == 1) ? ac->oc[1].m4ac.ext_sample_rate > ac->oc[1].m4ac.sample_rate : 0; samples <<= multiplier; - if (ac->output_configured < OC_LOCKED) { - avctx->sample_rate = ac->m4ac.sample_rate << multiplier; - avctx->frame_size = samples; - } if (samples) { - /* get output buffer */ ac->frame.nb_samples = samples; - if ((err = avctx->get_buffer(avctx, &ac->frame)) < 0) { - av_log(avctx, AV_LOG_ERROR, "get_buffer() failed\n"); - return err; - } - - if (avctx->sample_fmt == AV_SAMPLE_FMT_FLT) - ac->fmt_conv.float_interleave((float *)ac->frame.data[0], - (const float **)ac->output_data, - samples, avctx->channels); - else - ac->fmt_conv.float_to_int16_interleave((int16_t *)ac->frame.data[0], - (const float **)ac->output_data, - samples, avctx->channels); - *(AVFrame *)data = ac->frame; } *got_frame_ptr = !!samples; - if (ac->output_configured && audio_found) - ac->output_configured = OC_LOCKED; + if (ac->oc[1].status && audio_found) { + avctx->sample_rate = ac->oc[1].m4ac.sample_rate << multiplier; + avctx->frame_size = samples; + ac->oc[1].status = OC_LOCKED; + } return 0; +fail: + pop_output_configuration(ac); + return err; } static int aac_decode_frame(AVCodecContext *avctx, void *data, @@ -2456,10 +2544,13 @@ static int aac_decode_frame(AVCodecContext *avctx, void *data, return AVERROR(ENOMEM); avctx->extradata_size = new_extradata_size; memcpy(avctx->extradata, new_extradata, new_extradata_size); - if (decode_audio_specific_config(ac, ac->avctx, &ac->m4ac, + push_output_configuration(ac); + if (decode_audio_specific_config(ac, ac->avctx, &ac->oc[1].m4ac, avctx->extradata, - avctx->extradata_size*8, 1) < 0) + avctx->extradata_size*8, 1) < 0) { + pop_output_configuration(ac); return AVERROR_INVALIDDATA; + } } init_get_bits(&gb, buf, buf_size * 8); @@ -2519,7 +2610,7 @@ static int latm_decode_audio_specific_config(struct LATMContext *latmctx, { AACContext *ac = &latmctx->aac_ctx; AVCodecContext *avctx = ac->avctx; - MPEG4AudioConfig m4ac = {0}; + MPEG4AudioConfig m4ac = { 0 }; int config_start_bit = get_bits_count(gb); int sync_extension = 0; int bits_consumed, esize; @@ -2531,9 +2622,9 @@ static int latm_decode_audio_specific_config(struct LATMContext *latmctx, asclen = get_bits_left(gb); if (config_start_bit % 8) { - av_log_missing_feature(latmctx->aac_ctx.avctx, "audio specific " - "config not byte aligned.\n", 1); - return AVERROR_INVALIDDATA; + av_log_missing_feature(latmctx->aac_ctx.avctx, + "Non-byte-aligned audio-specific config", 1); + return AVERROR_PATCHWELCOME; } if (asclen <= 0) return AVERROR_INVALIDDATA; @@ -2544,8 +2635,8 @@ static int latm_decode_audio_specific_config(struct LATMContext *latmctx, if (bits_consumed < 0) return AVERROR_INVALIDDATA; - if (ac->m4ac.sample_rate != m4ac.sample_rate || - ac->m4ac.chan_config != m4ac.chan_config) { + if (ac->oc[1].m4ac.sample_rate != m4ac.sample_rate || + ac->oc[1].m4ac.chan_config != m4ac.chan_config) { av_log(avctx, AV_LOG_INFO, "audio config changed\n"); latmctx->initialized = 0; @@ -2587,7 +2678,7 @@ static int read_stream_mux_config(struct LATMContext *latmctx, // numPrograms if (get_bits(gb, 4)) { // numPrograms av_log_missing_feature(latmctx->aac_ctx.avctx, - "multiple programs are not supported\n", 1); + "Multiple programs", 1); return AVERROR_PATCHWELCOME; } @@ -2596,7 +2687,7 @@ static int read_stream_mux_config(struct LATMContext *latmctx, // for each layer (which there is only on in DVB) if (get_bits(gb, 3)) { // numLayer av_log_missing_feature(latmctx->aac_ctx.avctx, - "multiple layers are not supported\n", 1); + "Multiple layers", 1); return AVERROR_PATCHWELCOME; } @@ -2726,10 +2817,13 @@ static int latm_decode_frame(AVCodecContext *avctx, void *out, *got_frame_ptr = 0; return avpkt->size; } else { + push_output_configuration(&latmctx->aac_ctx); if ((err = decode_audio_specific_config( - &latmctx->aac_ctx, avctx, &latmctx->aac_ctx.m4ac, - avctx->extradata, avctx->extradata_size*8, 1)) < 0) + &latmctx->aac_ctx, avctx, &latmctx->aac_ctx.oc[1].m4ac, + avctx->extradata, avctx->extradata_size*8, 1)) < 0) { + pop_output_configuration(&latmctx->aac_ctx); return err; + } latmctx->initialized = 1; } } @@ -2747,7 +2841,7 @@ static int latm_decode_frame(AVCodecContext *avctx, void *out, return muxlength; } -av_cold static int latm_decode_init(AVCodecContext *avctx) +static av_cold int latm_decode_init(AVCodecContext *avctx) { struct LATMContext *latmctx = avctx->priv_data; int ret = aac_decode_init(avctx); @@ -2760,18 +2854,18 @@ av_cold static int latm_decode_init(AVCodecContext *avctx) AVCodec ff_aac_decoder = { - .name = "aac", - .type = AVMEDIA_TYPE_AUDIO, - .id = CODEC_ID_AAC, - .priv_data_size = sizeof(AACContext), - .init = aac_decode_init, - .close = aac_decode_close, - .decode = aac_decode_frame, - .long_name = NULL_IF_CONFIG_SMALL("Advanced Audio Coding"), - .sample_fmts = (const enum AVSampleFormat[]) { - AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_NONE + .name = "aac", + .type = AVMEDIA_TYPE_AUDIO, + .id = AV_CODEC_ID_AAC, + .priv_data_size = sizeof(AACContext), + .init = aac_decode_init, + .close = aac_decode_close, + .decode = aac_decode_frame, + .long_name = NULL_IF_CONFIG_SMALL("AAC (Advanced Audio Coding)"), + .sample_fmts = (const enum AVSampleFormat[]) { + AV_SAMPLE_FMT_FLTP, AV_SAMPLE_FMT_NONE }, - .capabilities = CODEC_CAP_CHANNEL_CONF | CODEC_CAP_DR1, + .capabilities = CODEC_CAP_CHANNEL_CONF | CODEC_CAP_DR1, .channel_layouts = aac_channel_layout, }; @@ -2781,17 +2875,17 @@ AVCodec ff_aac_decoder = { To do a more complex LATM demuxing a separate LATM demuxer should be used. */ AVCodec ff_aac_latm_decoder = { - .name = "aac_latm", - .type = AVMEDIA_TYPE_AUDIO, - .id = CODEC_ID_AAC_LATM, - .priv_data_size = sizeof(struct LATMContext), - .init = latm_decode_init, - .close = aac_decode_close, - .decode = latm_decode_frame, - .long_name = NULL_IF_CONFIG_SMALL("AAC LATM (Advanced Audio Codec LATM syntax)"), - .sample_fmts = (const enum AVSampleFormat[]) { - AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_NONE + .name = "aac_latm", + .type = AVMEDIA_TYPE_AUDIO, + .id = AV_CODEC_ID_AAC_LATM, + .priv_data_size = sizeof(struct LATMContext), + .init = latm_decode_init, + .close = aac_decode_close, + .decode = latm_decode_frame, + .long_name = NULL_IF_CONFIG_SMALL("AAC LATM (Advanced Audio Coding LATM syntax)"), + .sample_fmts = (const enum AVSampleFormat[]) { + AV_SAMPLE_FMT_FLTP, AV_SAMPLE_FMT_NONE }, - .capabilities = CODEC_CAP_CHANNEL_CONF | CODEC_CAP_DR1, + .capabilities = CODEC_CAP_CHANNEL_CONF | CODEC_CAP_DR1, .channel_layouts = aac_channel_layout, };