static lsmash_video_summary_t *vc1_create_summary( vc1_info_t *info, vc1_sequence_header_t *sequence, uint32_t max_au_length )
{
if( !info->sequence.present || !info->entry_point.present )
return NULL;
lsmash_video_summary_t *summary = (lsmash_video_summary_t *)lsmash_create_summary( LSMASH_SUMMARY_TYPE_VIDEO );
if( !summary )
return NULL;
lsmash_codec_specific_t *specific = lsmash_create_codec_specific_data( LSMASH_CODEC_SPECIFIC_DATA_TYPE_ISOM_VIDEO_VC_1,
LSMASH_CODEC_SPECIFIC_FORMAT_UNSTRUCTURED );
specific->data.unstructured = lsmash_create_vc1_specific_info( &info->dvc1_param, &specific->size );
if( !specific->data.unstructured
|| lsmash_add_entry( &summary->opaque->list, specific ) < 0 )
{
lsmash_cleanup_summary( (lsmash_summary_t *)summary );
lsmash_destroy_codec_specific_data( specific );
return NULL;
}
summary->sample_type = ISOM_CODEC_TYPE_VC_1_VIDEO;
summary->max_au_length = max_au_length;
summary->timescale = sequence->framerate_numerator;
summary->timebase = sequence->framerate_denominator;
summary->vfr = !sequence->framerate_flag;
summary->sample_per_field = 0;
summary->width = sequence->disp_horiz_size;
summary->height = sequence->disp_vert_size;
summary->par_h = sequence->aspect_width;
summary->par_v = sequence->aspect_height;
summary->color.primaries_index = sequence->color_prim;
summary->color.transfer_index = sequence->transfer_char;
summary->color.matrix_index = sequence->matrix_coef;
return summary;
}
int lsmash_add_codec_specific_data( lsmash_summary_t *summary, lsmash_codec_specific_t *specific )
{
if( !summary || !summary->opaque || !specific )
return LSMASH_ERR_FUNCTION_PARAM;
lsmash_codec_specific_t *dup = isom_duplicate_codec_specific_data( specific );
if( !dup )
return LSMASH_ERR_NAMELESS;
if( lsmash_add_entry( &summary->opaque->list, dup ) < 0 )
{
lsmash_destroy_codec_specific_data( dup );
return LSMASH_ERR_MEMORY_ALLOC;
}
return 0;
}
void lsmash_cleanup_summary( lsmash_summary_t *summary )
{
if( !summary )
return;
if( summary->opaque )
{
for( lsmash_entry_t *entry = summary->opaque->list.head; entry; )
{
lsmash_entry_t *next = entry->next;
lsmash_destroy_codec_specific_data( (lsmash_codec_specific_t *)entry->data );
lsmash_free( entry );
entry = next;
}
lsmash_free( summary->opaque );
}
lsmash_free( summary );
}
/* create AudioSpecificConfig as memory block from summary, and set it into that summary itself */
int lsmash_setup_AudioSpecificConfig( lsmash_audio_summary_t *summary )
{
if( !summary || !summary->opaque )
return LSMASH_ERR_FUNCTION_PARAM;
/* Remove an old one. */
for( lsmash_entry_t *entry = summary->opaque->list.head; entry; entry = entry->next )
{
lsmash_codec_specific_t *cs = (lsmash_codec_specific_t *)entry->data;
if( !cs || cs->type != LSMASH_CODEC_SPECIFIC_DATA_TYPE_MP4SYS_DECODER_CONFIG )
continue;
lsmash_remove_entry_direct( &summary->opaque->list, entry, lsmash_destroy_codec_specific_data );
}
/* Create and add a new one. */
uint32_t data_length;
uint8_t *data = mp4a_export_AudioSpecificConfig( summary->aot,
summary->frequency,
summary->channels,
summary->sbr_mode,
NULL, /* FIXME */
0, /* FIXME */
&data_length );
if( !data )
return LSMASH_ERR_NAMELESS;
lsmash_codec_specific_t *cs = lsmash_create_codec_specific_data( LSMASH_CODEC_SPECIFIC_DATA_TYPE_MP4SYS_DECODER_CONFIG,
LSMASH_CODEC_SPECIFIC_FORMAT_STRUCTURED );
if( !cs )
{
lsmash_free( data );
return LSMASH_ERR_MEMORY_ALLOC;
}
lsmash_mp4sys_decoder_parameters_t *param = (lsmash_mp4sys_decoder_parameters_t *)cs->data.structured;
param->objectTypeIndication = MP4SYS_OBJECT_TYPE_Audio_ISO_14496_3;
param->streamType = MP4SYS_STREAM_TYPE_AudioStream;
int err = lsmash_set_mp4sys_decoder_specific_info( param, data, data_length );
lsmash_free( data );
if( err < 0 || (err = lsmash_add_entry( &summary->opaque->list, cs )) < 0 )
{
lsmash_destroy_codec_specific_data( cs );
return err;
}
return 0;
}
static int write_headers( hnd_t handle, x264_nal_t *p_nal )
{
mp4_hnd_t *p_mp4 = handle;
uint32_t sps_size = p_nal[0].i_payload - H264_NALU_LENGTH_SIZE;
uint32_t pps_size = p_nal[1].i_payload - H264_NALU_LENGTH_SIZE;
uint32_t sei_size = p_nal[2].i_payload;
uint8_t *sps = p_nal[0].p_payload + H264_NALU_LENGTH_SIZE;
uint8_t *pps = p_nal[1].p_payload + H264_NALU_LENGTH_SIZE;
uint8_t *sei = p_nal[2].p_payload;
lsmash_codec_specific_t *cs = lsmash_create_codec_specific_data( LSMASH_CODEC_SPECIFIC_DATA_TYPE_ISOM_VIDEO_H264,
LSMASH_CODEC_SPECIFIC_FORMAT_STRUCTURED );
lsmash_h264_specific_parameters_t *param = (lsmash_h264_specific_parameters_t *)cs->data.structured;
param->lengthSizeMinusOne = H264_NALU_LENGTH_SIZE - 1;
/* SPS
* The remaining parameters are automatically set by SPS. */
if( lsmash_append_h264_parameter_set( param, H264_PARAMETER_SET_TYPE_SPS, sps, sps_size ) )
{
MP4_LOG_ERROR( "failed to append SPS.\n" );
return -1;
}
/* PPS */
if( lsmash_append_h264_parameter_set( param, H264_PARAMETER_SET_TYPE_PPS, pps, pps_size ) )
{
MP4_LOG_ERROR( "failed to append PPS.\n" );
return -1;
}
if( lsmash_add_codec_specific_data( (lsmash_summary_t *)p_mp4->summary, cs ) )
{
MP4_LOG_ERROR( "failed to add H.264 specific info.\n" );
return -1;
}
lsmash_destroy_codec_specific_data( cs );
/* Additional extensions */
/* Bitrate info */
cs = lsmash_create_codec_specific_data( LSMASH_CODEC_SPECIFIC_DATA_TYPE_ISOM_VIDEO_H264_BITRATE,
LSMASH_CODEC_SPECIFIC_FORMAT_STRUCTURED );
if( cs )
lsmash_add_codec_specific_data( (lsmash_summary_t *)p_mp4->summary, cs );
lsmash_destroy_codec_specific_data( cs );
p_mp4->i_sample_entry = lsmash_add_sample_entry( p_mp4->p_root, p_mp4->i_track, p_mp4->summary );
MP4_FAIL_IF_ERR( !p_mp4->i_sample_entry,
"failed to add sample entry for video.\n" );
/* SEI */
p_mp4->p_sei_buffer = malloc( sei_size );
MP4_FAIL_IF_ERR( !p_mp4->p_sei_buffer,
"failed to allocate sei transition buffer.\n" );
memcpy( p_mp4->p_sei_buffer, sei, sei_size );
p_mp4->i_sei_size = sei_size;
return sei_size + sps_size + pps_size;
}
static lsmash_audio_summary_t *wave_create_summary( waveformat_extensible_t *fmt )
{
lsmash_audio_summary_t *summary = (lsmash_audio_summary_t *)lsmash_create_summary( LSMASH_SUMMARY_TYPE_AUDIO );
if( !summary )
return NULL;
waveformat_extended_t *wfx = &fmt->wfx;
summary->sample_type = QT_CODEC_TYPE_LPCM_AUDIO;
summary->aot = MP4A_AUDIO_OBJECT_TYPE_NULL;
summary->frequency = wfx->nSamplesPerSec;
summary->channels = wfx->nChannels;
summary->sample_size = wfx->wFormatTag == WAVE_FORMAT_TYPE_ID_EXTENSIBLE
? fmt->Samples.wValidBitsPerSample
: wfx->wBitsPerSample;
summary->samples_in_frame = 1000; /* arbitrary */
summary->sbr_mode = MP4A_AAC_SBR_NOT_SPECIFIED;
summary->bytes_per_frame = wfx->nBlockAlign * summary->samples_in_frame;
summary->max_au_length = summary->bytes_per_frame;
lsmash_codec_specific_t *cs = lsmash_create_codec_specific_data( LSMASH_CODEC_SPECIFIC_DATA_TYPE_QT_AUDIO_FORMAT_SPECIFIC_FLAGS,
LSMASH_CODEC_SPECIFIC_FORMAT_STRUCTURED );
if( !cs )
goto fail;
lsmash_qt_audio_format_specific_flags_t *lpcm = (lsmash_qt_audio_format_specific_flags_t *)cs->data.structured;
if( (summary->sample_size & 7) == 0 )
lpcm->format_flags |= QT_AUDIO_FORMAT_FLAG_PACKED;
else
lpcm->format_flags |= QT_AUDIO_FORMAT_FLAG_ALIGNED_HIGH;
if( summary->sample_size > 8 )
lpcm->format_flags |= QT_AUDIO_FORMAT_FLAG_SIGNED_INTEGER;
if( lsmash_add_entry( &summary->opaque->list, cs ) < 0 )
{
lsmash_destroy_codec_specific_data( cs );
goto fail;
}
if( wfx->wFormatTag == WAVE_FORMAT_TYPE_ID_EXTENSIBLE || wfx->nChannels > 2 )
{
cs = lsmash_create_codec_specific_data( LSMASH_CODEC_SPECIFIC_DATA_TYPE_QT_AUDIO_CHANNEL_LAYOUT,
LSMASH_CODEC_SPECIFIC_FORMAT_STRUCTURED );
if( !cs )
goto fail;
lsmash_qt_audio_channel_layout_t *layout = (lsmash_qt_audio_channel_layout_t *)cs->data.structured;
if( wfx->wFormatTag == WAVE_FORMAT_TYPE_ID_EXTENSIBLE )
{
layout->channelLayoutTag = QT_CHANNEL_LAYOUT_USE_CHANNEL_BITMAP;
layout->channelBitmap = fmt->dwChannelMask;
}
else
{
layout->channelLayoutTag = QT_CHANNEL_LAYOUT_UNKNOWN | wfx->nChannels;
layout->channelBitmap = 0;
}
if( lsmash_add_entry( &summary->opaque->list, cs ) < 0 )
{
lsmash_destroy_codec_specific_data( cs );
goto fail;
}
}
return summary;
fail:
lsmash_cleanup_summary( (lsmash_summary_t *)summary );
return NULL;
}
static lsmash_audio_summary_t *dts_create_summary( dts_info_t *info )
{
lsmash_audio_summary_t *summary = (lsmash_audio_summary_t *)lsmash_create_summary( LSMASH_SUMMARY_TYPE_AUDIO );
if( !summary )
return NULL;
lsmash_dts_specific_parameters_t *param = &info->ddts_param;
lsmash_codec_specific_t *specific = lsmash_create_codec_specific_data( LSMASH_CODEC_SPECIFIC_DATA_TYPE_ISOM_AUDIO_DTS,
LSMASH_CODEC_SPECIFIC_FORMAT_UNSTRUCTURED );
specific->data.unstructured = lsmash_create_dts_specific_info( param, &specific->size );
if( !specific->data.unstructured
|| lsmash_add_entry( &summary->opaque->list, specific ) < 0 )
{
lsmash_cleanup_summary( (lsmash_summary_t *)summary );
lsmash_destroy_codec_specific_data( specific );
return NULL;
}
/* The CODEC identifiers probably should not be the combination of 'mp4a' and
* the objectTypeIndications for DTS audio since there is no public specification
* which defines the encapsulation of the stream as the MPEG-4 Audio context yet.
* In the world, there are muxers which is using such doubtful implementation.
* The objectTypeIndications are registered at MP4RA, but this does not always
* mean we can mux by using those objectTypeIndications.
* If available, there shall be the specification which defines the existence of
* DecoderSpecificInfo and its semantics, and what access unit consists of. */
summary->sample_type = lsmash_dts_get_codingname( param );
summary->aot = MP4A_AUDIO_OBJECT_TYPE_NULL; /* make no sense */
summary->sbr_mode = MP4A_AAC_SBR_NOT_SPECIFIED; /* make no sense */
switch( param->DTSSamplingFrequency )
{
case 12000 : /* Invalid? (No reference in the spec) */
case 24000 :
case 48000 :
case 96000 :
case 192000 :
case 384000 : /* Invalid? (No reference in the spec) */
summary->frequency = 48000;
break;
case 22050 :
case 44100 :
case 88200 :
case 176400 :
case 352800 : /* Invalid? (No reference in the spec) */
summary->frequency = 44100;
break;
case 8000 : /* Invalid? (No reference in the spec) */
case 16000 :
case 32000 :
case 64000 :
case 128000 :
summary->frequency = 32000;
break;
default :
summary->frequency = 0;
break;
}
summary->samples_in_frame = (summary->frequency * info->frame_duration) / param->DTSSamplingFrequency;
summary->max_au_length = DTS_MAX_CORE_SIZE + DTS_MAX_NUM_EXSS * DTS_MAX_EXSS_SIZE;
summary->sample_size = param->pcmSampleDepth;
summary->channels = dts_get_max_channel_count( info );
return summary;
}
