From 5c54fa53876a92d4c7b923c5a6312f427cfc8c86 Mon Sep 17 00:00:00 2001 From: Fraunhofer IIS FDK Date: Thu, 20 Dec 2018 15:52:46 +0100 Subject: [PATCH] Add AOT specific number of qmf bands sanity check in SpatialSpecificConfig() Test: atest DecoderTestXheAac ; atest DecoderTestAacDrc Change-Id: Ic106c4371c5ac17cb832c7d9db042bcc9d1e7a09 --- libAACdec/src/aacdecoder_lib.cpp | 6 +- libAACenc/src/aacenc_lib.cpp | 3 +- libMpegTPDec/include/tp_data.h | 3 +- libMpegTPDec/src/tpdec_asc.cpp | 27 ++++-- libMpegTPEnc/include/tp_data.h | 3 +- libMpegTPEnc/src/tpenc_asc.cpp | 4 +- libSACdec/include/sac_dec_lib.h | 6 +- libSACdec/src/sac_bitdec.cpp | 4 + libSACdec/src/sac_dec_lib.cpp | 138 +++++++++++++++++++++++-------- 9 files changed, 144 insertions(+), 50 deletions(-) diff --git a/libAACdec/src/aacdecoder_lib.cpp b/libAACdec/src/aacdecoder_lib.cpp index 2860bf6..cbcd404 100644 --- a/libAACdec/src/aacdecoder_lib.cpp +++ b/libAACdec/src/aacdecoder_lib.cpp @@ -387,7 +387,7 @@ static INT aacDecoder_SbrCallback( static INT aacDecoder_SscCallback(void *handle, HANDLE_FDK_BITSTREAM hBs, const AUDIO_OBJECT_TYPE coreCodec, - const INT samplingRate, + const INT samplingRate, const INT frameSize, const INT stereoConfigIndex, const INT coreSbrFrameLengthIndex, const INT configBytes, const UCHAR configMode, @@ -398,8 +398,8 @@ static INT aacDecoder_SscCallback(void *handle, HANDLE_FDK_BITSTREAM hBs, err = mpegSurroundDecoder_Config( (CMpegSurroundDecoder *)hAacDecoder->pMpegSurroundDecoder, hBs, coreCodec, - samplingRate, stereoConfigIndex, coreSbrFrameLengthIndex, configBytes, - configMode, configChanged); + samplingRate, frameSize, stereoConfigIndex, coreSbrFrameLengthIndex, + configBytes, configMode, configChanged); switch (err) { case MPS_UNSUPPORTED_CONFIG: diff --git a/libAACenc/src/aacenc_lib.cpp b/libAACenc/src/aacenc_lib.cpp index 9e0630c..08cd57a 100644 --- a/libAACenc/src/aacenc_lib.cpp +++ b/libAACenc/src/aacenc_lib.cpp @@ -1216,7 +1216,8 @@ static INT aacenc_SbrCallback(void *self, HANDLE_FDK_BITSTREAM hBs, INT aacenc_SscCallback(void *self, HANDLE_FDK_BITSTREAM hBs, const AUDIO_OBJECT_TYPE coreCodec, - const INT samplingRate, const INT stereoConfigIndex, + const INT samplingRate, const INT frameSize, + const INT stereoConfigIndex, const INT coreSbrFrameLengthIndex, const INT configBytes, const UCHAR configMode, UCHAR *configChanged) { HANDLE_AACENCODER hAacEncoder = (HANDLE_AACENCODER)self; diff --git a/libMpegTPDec/include/tp_data.h b/libMpegTPDec/include/tp_data.h index b4ab802..b015332 100644 --- a/libMpegTPDec/include/tp_data.h +++ b/libMpegTPDec/include/tp_data.h @@ -367,7 +367,8 @@ typedef INT (*cbFreeMem_t)(void *, const CSAudioSpecificConfig *); typedef INT (*cbCtrlCFGChange_t)(void *, const CCtrlCFGChange *); typedef INT (*cbSsc_t)(void *, HANDLE_FDK_BITSTREAM, const AUDIO_OBJECT_TYPE coreCodec, - const INT samplingRate, const INT stereoConfigIndex, + const INT samplingRate, const INT frameSize, + const INT stereoConfigIndex, const INT coreSbrFrameLengthIndex, const INT configBytes, const UCHAR configMode, UCHAR *configChanged); diff --git a/libMpegTPDec/src/tpdec_asc.cpp b/libMpegTPDec/src/tpdec_asc.cpp index 5270f2b..28bc22d 100644 --- a/libMpegTPDec/src/tpdec_asc.cpp +++ b/libMpegTPDec/src/tpdec_asc.cpp @@ -1413,7 +1413,9 @@ static TRANSPORTDEC_ERROR EldSpecificConfig_Parse(CSAudioSpecificConfig *asc, esc->m_useLdQmfTimeAlign = 1; if (cb->cbSsc != NULL) { ErrorStatus = (TRANSPORTDEC_ERROR)cb->cbSsc( - cb->cbSscData, hBs, asc->m_aot, asc->m_extensionSamplingFrequency, + cb->cbSscData, hBs, asc->m_aot, + asc->m_samplingFrequency << esc->m_sbrSamplingRate, + asc->m_samplesPerFrame << esc->m_sbrSamplingRate, 1, /* stereoConfigIndex */ -1, /* nTimeSlots: read from bitstream */ eldExtLen, asc->configMode, &asc->SacConfigChanged); @@ -1812,9 +1814,16 @@ static TRANSPORTDEC_ERROR UsacRsv60DecoderConfig_Parse( if (usc->element[i].m_stereoConfigIndex > 0) { if (cb->cbSsc != NULL) { + int samplesPerFrame = asc->m_samplesPerFrame; + + if (usc->m_sbrRatioIndex == 1) samplesPerFrame <<= 2; + if (usc->m_sbrRatioIndex == 2) + samplesPerFrame = (samplesPerFrame * 8) / 3; + if (usc->m_sbrRatioIndex == 3) samplesPerFrame <<= 1; + /* Mps212Config() ISO/IEC FDIS 23003-3 */ if (cb->cbSsc(cb->cbSscData, hBs, asc->m_aot, - asc->m_extensionSamplingFrequency, + asc->m_extensionSamplingFrequency, samplesPerFrame, usc->element[i].m_stereoConfigIndex, usc->m_coreSbrFrameLengthIndex, 0, /* don't know the length */ @@ -2181,8 +2190,9 @@ TRANSPORTDEC_ERROR AudioSpecificConfig_Parse( case AOT_MPEGS: if (cb->cbSsc != NULL) { if (cb->cbSsc(cb->cbSscData, bs, self->m_aot, self->m_samplingFrequency, - 1, -1, /* nTimeSlots: read from bitstream */ - 0, /* don't know the length */ + self->m_samplesPerFrame, 1, + -1, /* nTimeSlots: read from bitstream */ + 0, /* don't know the length */ self->configMode, &self->SacConfigChanged)) { return TRANSPORTDEC_UNSUPPORTED_FORMAT; } @@ -2365,10 +2375,17 @@ static TRANSPORTDEC_ERROR Drm_xHEAACDecoderConfig( /*usc->element[elemIdx].m_stereoConfigIndex =*/FDKreadBits(hBs, 2); if (usc->element[elemIdx].m_stereoConfigIndex > 0) { if (cb->cbSsc != NULL) { + int samplesPerFrame = asc->m_samplesPerFrame; + + if (usc->m_sbrRatioIndex == 1) samplesPerFrame <<= 2; + if (usc->m_sbrRatioIndex == 2) + samplesPerFrame = (samplesPerFrame * 8) / 3; + if (usc->m_sbrRatioIndex == 3) samplesPerFrame <<= 1; + ErrorStatus = (TRANSPORTDEC_ERROR)cb->cbSsc( cb->cbSscData, hBs, AOT_DRM_USAC, /* syntax differs from MPEG Mps212Config() */ - asc->m_extensionSamplingFrequency, + asc->m_extensionSamplingFrequency, samplesPerFrame, usc->element[elemIdx].m_stereoConfigIndex, usc->m_coreSbrFrameLengthIndex, 0, /* don't know the length */ asc->configMode, &asc->SacConfigChanged); diff --git a/libMpegTPEnc/include/tp_data.h b/libMpegTPEnc/include/tp_data.h index 6f032d1..00de356 100644 --- a/libMpegTPEnc/include/tp_data.h +++ b/libMpegTPEnc/include/tp_data.h @@ -367,7 +367,8 @@ typedef INT (*cbFreeMem_t)(void *, const CSAudioSpecificConfig *); typedef INT (*cbCtrlCFGChange_t)(void *, const CCtrlCFGChange *); typedef INT (*cbSsc_t)(void *, HANDLE_FDK_BITSTREAM, const AUDIO_OBJECT_TYPE coreCodec, - const INT samplingRate, const INT stereoConfigIndex, + const INT samplingRate, const INT frameSize, + const INT stereoConfigIndex, const INT coreSbrFrameLengthIndex, const INT configBytes, const UCHAR configMode, UCHAR *configChanged); diff --git a/libMpegTPEnc/src/tpenc_asc.cpp b/libMpegTPEnc/src/tpenc_asc.cpp index ce4e364..0b484a0 100644 --- a/libMpegTPEnc/src/tpenc_asc.cpp +++ b/libMpegTPEnc/src/tpenc_asc.cpp @@ -795,7 +795,7 @@ static int transportEnc_writeELDSpecificConfig(HANDLE_FDK_BITSTREAM hBs, const INT eldExtLen = (cb->cbSsc(cb->cbSscData, NULL, config->aot, config->extSamplingRate, 0, - 0, 0, 0, NULL) + + 0, 0, 0, 0, NULL) + 7) >> 3; INT cnt = eldExtLen; @@ -818,7 +818,7 @@ static int transportEnc_writeELDSpecificConfig(HANDLE_FDK_BITSTREAM hBs, } cb->cbSsc(cb->cbSscData, hBs, config->aot, config->extSamplingRate, 0, 0, 0, - 0, NULL); + 0, 0, NULL); } if (config->downscaleSamplingRate != 0 && diff --git a/libSACdec/include/sac_dec_lib.h b/libSACdec/include/sac_dec_lib.h index 2797a10..9913279 100644 --- a/libSACdec/include/sac_dec_lib.h +++ b/libSACdec/include/sac_dec_lib.h @@ -315,9 +315,9 @@ SACDEC_ERROR mpegSurroundDecoder_Init( */ SACDEC_ERROR mpegSurroundDecoder_Config( CMpegSurroundDecoder *pMpegSurroundDecoder, HANDLE_FDK_BITSTREAM hBs, - AUDIO_OBJECT_TYPE coreCodec, INT samplingRate, INT stereoConfigIndex, - INT coreSbrFrameLengthIndex, INT configBytes, const UCHAR configMode, - UCHAR *configChanged); + AUDIO_OBJECT_TYPE coreCodec, INT samplingRate, INT frameSize, + INT stereoConfigIndex, INT coreSbrFrameLengthIndex, INT configBytes, + const UCHAR configMode, UCHAR *configChanged); SACDEC_ERROR mpegSurroundDecoder_ConfigureQmfDomain( diff --git a/libSACdec/src/sac_bitdec.cpp b/libSACdec/src/sac_bitdec.cpp index 1049c3d..883e1e8 100644 --- a/libSACdec/src/sac_bitdec.cpp +++ b/libSACdec/src/sac_bitdec.cpp @@ -517,6 +517,10 @@ SACDEC_ERROR SpatialDecParseSpecificConfig( pSpatialSpecificConfig->tempShapeConfig = (SPATIALDEC_TS_CONF)FDKreadBits(bitstream, 2); + if (pSpatialSpecificConfig->tempShapeConfig > 2) { + return MPS_PARSE_ERROR; /* reserved value */ + } + pSpatialSpecificConfig->decorrConfig = (SPATIALDEC_DECORR_CONF)FDKreadBits(bitstream, 2); if (pSpatialSpecificConfig->decorrConfig > 2) { diff --git a/libSACdec/src/sac_dec_lib.cpp b/libSACdec/src/sac_dec_lib.cpp index 7549e51..5792858 100644 --- a/libSACdec/src/sac_dec_lib.cpp +++ b/libSACdec/src/sac_dec_lib.cpp @@ -237,6 +237,11 @@ struct MpegSurroundDecoder { SPATIAL_DEC_CONFIG decConfig; }; +SACDEC_ERROR +static sscCheckOutOfBand(const SPATIAL_SPECIFIC_CONFIG *pSsc, + const INT coreCodec, const INT sampleRate, + const INT frameSize); + static SACDEC_ERROR sscParseCheck(const SPATIAL_SPECIFIC_CONFIG *pSsc); /** @@ -694,11 +699,13 @@ bail: **/ SACDEC_ERROR mpegSurroundDecoder_Config( CMpegSurroundDecoder *pMpegSurroundDecoder, HANDLE_FDK_BITSTREAM hBs, - AUDIO_OBJECT_TYPE coreCodec, INT samplingRate, INT stereoConfigIndex, - INT coreSbrFrameLengthIndex, INT configBytes, const UCHAR configMode, - UCHAR *configChanged) { + AUDIO_OBJECT_TYPE coreCodec, INT samplingRate, INT frameSize, + INT stereoConfigIndex, INT coreSbrFrameLengthIndex, INT configBytes, + const UCHAR configMode, UCHAR *configChanged) { SACDEC_ERROR err = MPS_OK; SPATIAL_SPECIFIC_CONFIG spatialSpecificConfig; + SPATIAL_SPECIFIC_CONFIG *pSsc = + &pMpegSurroundDecoder->spatialSpecificConfigBackup; switch (coreCodec) { case AOT_DRM_USAC: @@ -709,6 +716,7 @@ SACDEC_ERROR mpegSurroundDecoder_Config( err = SpatialDecParseMps212Config( hBs, &spatialSpecificConfig, samplingRate, coreCodec, stereoConfigIndex, coreSbrFrameLengthIndex); + pSsc = &spatialSpecificConfig; } else { err = SpatialDecParseMps212Config( hBs, &pMpegSurroundDecoder->spatialSpecificConfigBackup, @@ -723,6 +731,7 @@ SACDEC_ERROR mpegSurroundDecoder_Config( * into temporarily allocated structure */ err = SpatialDecParseSpecificConfig(hBs, &spatialSpecificConfig, configBytes, coreCodec); + pSsc = &spatialSpecificConfig; } else { err = SpatialDecParseSpecificConfig( hBs, &pMpegSurroundDecoder->spatialSpecificConfigBackup, @@ -738,14 +747,21 @@ SACDEC_ERROR mpegSurroundDecoder_Config( goto bail; } + err = sscCheckOutOfBand(pSsc, coreCodec, samplingRate, frameSize); + + if (err != MPS_OK) { + goto bail; + } + if (configMode & AC_CM_DET_CFG_CHANGE) { return err; } if (configMode & AC_CM_ALLOC_MEM) { if (*configChanged) { - if ((err = mpegSurroundDecoder_Open(&pMpegSurroundDecoder, - stereoConfigIndex, NULL))) { + err = mpegSurroundDecoder_Open(&pMpegSurroundDecoder, stereoConfigIndex, + NULL); + if (err) { return err; } } @@ -815,29 +831,9 @@ static MPEGS_OPMODE mpegSurroundOperationMode( * \return MPS_OK on sucess, and else on parse error. */ static SACDEC_ERROR sscParseCheck(const SPATIAL_SPECIFIC_CONFIG *pSsc) { - SACDEC_ERROR err = MPS_OK; - if (pSsc->samplingFreq > 96000) return MPS_PARSE_ERROR; if (pSsc->samplingFreq < 8000) return MPS_PARSE_ERROR; - switch (pSsc->freqRes) { - case SPATIALDEC_FREQ_RES_28: - case SPATIALDEC_FREQ_RES_20: - case SPATIALDEC_FREQ_RES_14: - case SPATIALDEC_FREQ_RES_10: - case SPATIALDEC_FREQ_RES_23: - case SPATIALDEC_FREQ_RES_15: - case SPATIALDEC_FREQ_RES_12: - case SPATIALDEC_FREQ_RES_9: - case SPATIALDEC_FREQ_RES_7: - case SPATIALDEC_FREQ_RES_5: - case SPATIALDEC_FREQ_RES_4: - break; - case SPATIALDEC_FREQ_RES_40: /* 40 doesn't exist in ISO/IEC 23003-1 */ - default: - return MPS_PARSE_ERROR; - } - if ((pSsc->treeConfig < 0) || (pSsc->treeConfig > 7)) { return MPS_PARSE_ERROR; } @@ -846,17 +842,9 @@ static SACDEC_ERROR sscParseCheck(const SPATIAL_SPECIFIC_CONFIG *pSsc) { return MPS_PARSE_ERROR; } - if (pSsc->tempShapeConfig == 3) { - return MPS_PARSE_ERROR; - } - - if (pSsc->decorrConfig == 3) { - return MPS_PARSE_ERROR; - } - /* now we are sure there were no parsing errors */ - return err; + return MPS_OK; } /** @@ -1024,6 +1012,11 @@ static SACDEC_ERROR sscCheckInBand(SPATIAL_SPECIFIC_CONFIG *pSsc, FDK_ASSERT(pSsc != NULL); + /* check ssc for parse errors */ + if (sscParseCheck(pSsc) != MPS_OK) { + err = MPS_PARSE_ERROR; + } + /* core fs and mps fs must match */ if (pSsc->samplingFreq != sampleRate) { err = MPS_PARSE_ERROR /* MPEGSDEC_SSC_PARSE_ERROR */; @@ -1093,6 +1086,83 @@ mpegSurroundDecoder_ConfigureQmfDomain( return err; } +/** + * \brief Check out-of-band config + * + * \param pSsc spatial specific config handle. + * \param coreCodec core codec. + * \param sampleRate sampling frequency. + * + * \return errorStatus + */ +SACDEC_ERROR +sscCheckOutOfBand(const SPATIAL_SPECIFIC_CONFIG *pSsc, const INT coreCodec, + const INT sampleRate, const INT frameSize) { + FDK_ASSERT(pSsc != NULL); + int qmfBands = 0; + + /* check ssc for parse errors */ + if (sscParseCheck(pSsc) != MPS_OK) { + return MPS_PARSE_ERROR; + } + + switch (coreCodec) { + case AOT_USAC: + case AOT_DRM_USAC: + /* ISO/IEC 23003-1:2007(E), Chapter 6.3.3, Support for lower and higher + * sampling frequencies */ + if (pSsc->samplingFreq >= 55426) { + return MPS_PARSE_ERROR; + } + break; + case AOT_ER_AAC_LD: + case AOT_ER_AAC_ELD: + /* core fs and mps fs must match */ + if (pSsc->samplingFreq != sampleRate) { + return MPS_PARSE_ERROR; + } + + /* ISO/IEC 14496-3:2009 FDAM 3: Chapter 1.5.2.3, Levels for the Low Delay + * AAC v2 profile */ + if (pSsc->samplingFreq > 48000) { + return MPS_PARSE_ERROR; + } + + qmfBands = mpegSurroundDecoder_GetNrOfQmfBands(pSsc, pSsc->samplingFreq); + switch (frameSize) { + case 480: + if (!((qmfBands == 32) && (pSsc->nTimeSlots == 15))) { + return MPS_PARSE_ERROR; + } + break; + case 960: + if (!((qmfBands == 64) && (pSsc->nTimeSlots == 15))) { + return MPS_PARSE_ERROR; + } + break; + case 512: + if (!(((qmfBands == 32) && (pSsc->nTimeSlots == 16)) || + ((qmfBands == 64) && (pSsc->nTimeSlots == 8)))) { + return MPS_PARSE_ERROR; + } + break; + case 1024: + if (!((qmfBands == 64) && (pSsc->nTimeSlots == 16))) { + return MPS_PARSE_ERROR; + } + break; + default: + return MPS_PARSE_ERROR; + } + break; + default: + return MPS_PARSE_ERROR; + break; + } + + return MPS_OK; +} + /** * \brief Decode MPEG Surround frame. **/