File fdk-aac.changes of Package fdk-aac

-------------------------------------------------------------------
Fri Dec 22 10:13:42 UTC 2023 - marec@detebe.org

- Update to version 2.0.3:
  * Bump the versions and add changelog entries for v2.0.3
  * Update the list of paths in EXTRA_DIST
  * ci: Do "apt-get update" before installing anything
  * Add build support for s390x
  * Saturate additions in getEnvSfbEnergy in SBR encoding
  * cmake: Update the description of the BUILD_PROGRAMS option
  * gitignore: Ignore files with a ~ suffix
  * gitignore: Add the new test-encode-decode tool
  * Remove duplicate AACDECODER_LIB_VL*
  * ci: Run tests, add sanitizer build configurations that run tests too
  * Add a script for doing very rudimentary regression testing
  * Add a tool for testing decoding+encoding in a number of configurations
  * cmake: Link the library as C, not C++
  * github: Add -Wno-stringop-overflow to the GCC/CMake builds
  * ci: Disable warnings about deprecated functions on macOS
  * ci: Update the llvm-mingw job to the latest version
  * ci: Update the checkout steps to the latest, non deprecated version
  * Increase patchParam array size by one and fix out-of-bounce write in resetLppTransposer().
  * Increase patchParam array size by one and fix out-of-bounce write in resetLppTransposer().
  * Increase patchParam array size by one and fix out-of-bounce write in resetLppTransposer().
  * Increase patchParam array size by one and fix out-of-bounce write in resetLppTransposer().
  * Increase patchParam array size by one and fix out-of-bounce write in resetLppTransposer().
  * Increase patchParam array size by one and fix out-of-bounce write in resetLppTransposer().
  * Increase patchParam array size by one and fix out-of-bounce write in resetLppTransposer().
  * Increase patchParam array size by one and fix out-of-bounce write in resetLppTransposer().
  * Increase patchParam array size by one and fix out-of-bounce write in resetLppTransposer().
  * Disable MPEG-D DRC for legacy AOTs
  * Remove obsolete uni drc precedence handling
  * Remove obsolete uni drc precedence handling
  * Disable MPEG-D DRC for legacy AOTs
  * Correct comment spelling error.
  * Updated fuzz_config in Android.bp file
  * Update the version numbers in the public installed headers
  * Include the encoder library version in the publicly installed header
  * [Bluetooth apex] Use new apex name
  * Add riscv64 support
  * ci: Use sysctl instead of nproc on macOS
  * Reject invalid out of band config in transportDec_OutOfBandConfig() and skip re-allocation.
  * Reject invalid out of band config in transportDec_OutOfBandConfig() and skip re-allocation.
  * Reject invalid out of band config in transportDec_OutOfBandConfig() and skip re-allocation.
  * Reject invalid out of band config in transportDec_OutOfBandConfig() and skip re-allocation.
  * Reject invalid out of band config in transportDec_OutOfBandConfig() and skip re-allocation.
  * Reject invalid out of band config in transportDec_OutOfBandConfig() and skip re-allocation.
  * Reject invalid out of band config in transportDec_OutOfBandConfig() and skip re-allocation.
  * Add license_note
  * Add Github Actions CI build testing
  * Revert "Avoid failing asserts on broken bitstreams"
  * Renamed bluetooth apex name to com.android.bluetooth
  * Update aac-enc.c
  * Do not force inlining of indirect functions
  * Adjust eqSubbandGainFormat data type to prevent load of invalid value in _skipEqCoefficients().
  * Refine aacDecoder_DecodeFrame() output buffer parameter type description.
  * Add parameter pointer check to aacEncInfo() encoder API function.
  * Prevent too large shift exponent in apply_inter_tes() and merge two loops.
  * Use scaleValuesSaturate() to avoid integer overflows in QmfTransposerApply().
  * Pass flushing flag to SBR module for stereo config index 3 as for stereo config index 1 and 2.
  * Utilize dynamic scaling in slotAmp() to avoid signed integer overflows.
  * Adapt scaling in combineSignalCplxScale*() to prevent signed integer overflows.
  * Fix fixmadddiv2_DD() integer overflow in SpatialDecApplyM2_Mode212_ResidualsPlusPhaseCoding().
  * Avoid unintentional sign conversions in lppTransposer() and lppTransposerHBE().
  * Improve validation of channel indices in pcmDmx.
  * Fix stack buffer overflow in Pred_lt4().
  * Prevent signed integer overflow in complex path of calc_qmfBuffer().
  * Prevent signed integer overflow in calc_qmfBufferReal().
  * Revise synthesis QMF scaling for PS to avoid integer overflow in qmfSynPrototypeFirSlot().
  * Avoid signed integer overflow in combineSignalCplxScale2().
  * Prevent undefined values for CODING_SCHEME type in huff_decode().
  * Avoid load of undefined SPATIALDEC_TREE_CONFIG enum value.
  * Adjust data types to prevent load of invalid value
  * Avoid signed integer overflow in pcmLimiter_Apply().
  * Validate whether all PCE listed element instance tags are present in raw_data_block.
  * Avoid integer overflows in CLatmDemux_ReadAuChunkLengthInfo() and FDK_get32() to prevent endless loop.
  * Check the number of available escapes in rvlcDecodeBackward() to avoid out-of-bounds access.
  * Sbr syncstate may only be set to upsampling if the sbr decoder was sucessfully initialized.
  * Perform parameter check before parameter assignment to avoid incorrect initialization in resetFreqBandTables().
  * Use local variables for sbr and mps state in CAacDecoder_Init() to avoid inconsistencies in case of failing initialization.
  * Follow-up on: Improve decoder robustness by storing flags and elFlags temporarily.
  * Use dynamic scaling depending on autocorr length to avoid signed integer overflow.
  * Avoid signed integer overflow in pcmLimiter_Apply().
  * Stop using __DATE__/__TIME__ on all builds
  * Solve issue regarding config changes between Loudness-only vs. Loudness+DRC.
  * Overcome SBR encoder coupling quality issue in ELD.
  * Fix input buffer flushing for PS encoder.
  * Fix wrong input buffer feed for PS encoder at less input samples than framesize.
  * Read uniDrcGainExtension element only if all DRC gain sequences are parsed correctly.
  * Consider TNS headroom for complex prediction.
  * Fix USAC time domain limiter latency at config change.
  * Introduce additional sanity checks to validate program config element.
  * Validate DRC compression factor and DRC boost factor value range in aacDecoder_SetParam().
  * Fix unsigned integer overflow in Hcr_State_BODY_SIGN_ESC__ESC_WORD().
  * Make sure that alphaValue is correctly initialized in sbrDecoder_drcApplySlot().
  * Introduce aacDecoder_drcDisable() and always disable legacy DRC for USAC.
  * Evaluate and return StoreConfigAsBitstream() error state.
  * Check transportDec_OutOfBandConfig() input buffer size parameter.
  * ELD downscale factor 3 is only allowed for framesize 480.
  * Improve decoder robustness by storing flags and elFlags temporarily.
  * Revise bypass mode in SpatialDecApplyParameterSets() to prevent an assert in fDivNorm().
  * Check number of core channels and SAC decoder input channels to avoid a channel mismatch.
  * Prevent integer overflow in subbandTPApply() energy update.
  * Limit smoothedNoise values to avoid integer overflows in adjustTimeSlotHQ() and adjustTimeSlotHQ_GainAndNoise().
  * Avoid integer overflow in dct_II().
  * Prevent integer overflows in dualChannelFiltering() and eightChannelFiltering().
  * Adjust VBR mode depending on given peak bitrate and fix crash recovery usage.
  * Overcome SBR encoder coupling quality issue in ELD.
  * Fix input buffer flushing for PS encoder.
  * Fix wrong input buffer feed for PS encoder at less input samples than framesize.
  * Solve issue regarding config changes between Loudness-only vs. Loudness+DRC.
  * Read uniDrcGainExtension element only if all DRC gain sequences are parsed correctly.
  * Fix USAC time domain limiter latency at config change.
  * Consider TNS headroom for complex prediction.
  * Introduce additional sanity checks to validate program config element.
  * Check number of core channels and SAC decoder input channels to avoid a channel mismatch.
  * Improve decoder robustness by storing flags and elFlags temporarily.
  * Make sure that alphaValue is correctly initialized in sbrDecoder_drcApplySlot().
  * Fix unsigned integer overflow in Hcr_State_BODY_SIGN_ESC__ESC_WORD().
  * Revise bypass mode in SpatialDecApplyParameterSets() to prevent an assert in fDivNorm().
  * ELD downscale factor 3 is only allowed for framesize 480.
  * Check transportDec_OutOfBandConfig() input buffer size parameter.
  * Evaluate and return StoreConfigAsBitstream() error state.
  * Introduce aacDecoder_drcDisable() and always disable legacy DRC for USAC.
  * Adjust VBR mode depending on given peak bitrate and fix crash recovery usage.
  * Validate DRC compression factor and DRC boost factor value range in aacDecoder_SetParam().
  * Prevent integer overflow in subbandTPApply() energy update.
  * Limit smoothedNoise values to avoid integer overflows in adjustTimeSlotHQ() and adjustTimeSlotHQ_GainAndNoise().
  * Avoid integer overflow in dct_II().
  * Prevent integer overflows in dualChannelFiltering() and eightChannelFiltering().
  * Fix fuzzer's use of aacDecoder_DecodeFrame
  * Fix heap buffer overflow in sbrDecoder_AssignQmfChannels2SbrChannels().

-------------------------------------------------------------------
Wed Apr 28 13:37:25 UTC 2021 - marec@detebe.org

- Update to version 2.0.2:
  * Bump the versions and add changelog entries for v2.0.2
  * Add files missing from EXTRA_DIST
  * Specify the macOS compatibility version to make it a drop-in replacement for the libtool built
  * Don't use enums for values read directly from the bitstream
  * Made this compatible with cmake version 3.5.1
  * Remove the separate nmakefile for MSVC
  * Fixed a typo in the documentation
  * Fix improper assignment in aac_enc_fuzzer
  * [LSC] Add LOCAL_LICENSE_KINDS to external/aac
  * Avoid failing asserts on broken bitstreams
  * aac_enc_fuzzer: Improve code coverage
  * Add CMake project
  * Don't use enums for values read directly from the bitstream
  * fix compilation with mips16
  * aac remove diag
  * Adjust VBR mode depending on given peak bitrate and fix crash recovery usage.
  * Added aac_enc_fuzzer
  * Enable cfi for aac
  * Don't use an enum for a value read directly from the bitstream
  * Revert "Avoid a failed assert on invalid channel indices"
  * Revert "Don't use an enum for a value read directly from the bitstream"
  * Revert "Don't use an enum for a value read directly from the bitstream"
  * Revert "Don't use an enum for a value read directly from the bitstream"
  * Revert "Avoid undefined shifts in SATURATE_SHIFT"
  * Update the version numbers in the public installed headers
  * Remove references to sbr_crc.cpp that was removed upstream
  * Avoid undefined shifts in calculateSbrEnvelope
  * Avoid undefined shifts in SBR calcCRC
  * Fix heap buffer overflow in sbrDecoder_AssignQmfChannels2SbrChannels().
  * Avoid undefined shifts in SATURATE_SHIFT
  * Don't use an enum for a value read directly from the bitstream
  * Avoid undefined shifts
  * Improve code coverage for aac_dec_fuzzer
  * Don't store a value read directly from the bitstream in an enum
  * Don't use an enum for a value read directly from the bitstream
  * point to new repo hooks for mainline uploads
  * Fix license type s/b BY_EXCEPTION_ONLY
  * Added fuzz_config field in aac_dec_fuzzer
  * Set min_sdk_version to be part of mainline modules
  * Add METADATA to aac: Apache2=NOTICE
  * Set min_sdk_version to be part of mainline modules
  * Added aac_dec_fuzzer
  * aacdec: Add host support
  * Stop using __DATE__/__TIME__ on all builds
  * Remove outdated apply_normalization parameter from legacy DRC module.
  * Avoid decoder internal clipping by converting the whole audio sample data path from 16 to 32 bit data width (FDKdec v3.2.0).
  * Create new applyDrcLevelNormalization() function for subsequent usage.
  * Revise memory overlay usage and remove deprecated buffers and overlay tags.
  * Restructure QMF implementation and provide 16/32 bit implementation in parallel.
  * Enhance encoder audio quality for vbr (FDKenc v4.0.1).
  * Improve encoder audio quality for low bitreservoir configuration.
  * Fix SBR encoder element mapping for channel configurations with more than one layer
  * Ensure that QC_OUT_CHANNEL struct in encoder always uses aligned memory
  * Revise and straighten out pointless description from encoder documentation
  * Revise and straighten out useless references from decoder documentation.
  * Extend decoder API with audio output loudness info (FDKdec v3.1.3).
  * Don't use an enum for a value read directly from the bitstream
  * Request for Expert and Anchor loudness by default.
  * Fix AACDEC_INTR handling for USAC configuration with multiple pre-roll AUs
  * Do not allow channel configuration change within PCE for ADTS. Fixes assert.
  * Avoid negative valid bits in adtsRead_DecodeHeader()
  * Reject HE-AAC SBR with a ratio greater than 2:1 and reject USAC streams containing legacy DRC info.
  * Reject undefined fPow() input data in _compressorIO_sigmoid_common().
  * Avoid undefined shift exponent in f2Pow().
  * Perform sanity check on DRC sets and improve the selection process
  * Fix too large shift exponent in CConcealment_InterpolateBuffer().
  * Prevent too large shift exponents in SpatialDecReshapeBBEnv().
  * Avoid too large shift exponent in apply_inter_tes() by using fIsLessThan.
  * Fix IsLessThan() function for certain edge cases.
  * Avoid unsigned integer overflows in FDK_Feed().
  * Prevent undefined syncLayerFrameBits in synchronization().
  * Avoid unsigned integer overflow in transportDec_InBandConfig() and transportDec_ReadAccessUnit().
  * Prevent negation of INT_MIN in CConcealment_ApplyRandomSign().
  * Revise scaling in extractBBEnv() to avoid potential signed integer overflows.
  * Avoid signed integer overflow in combineSignalCplx*().
  * Prevent signed integer overflow in TsdApply().
  * Revise scaling for USAC combined with phase coding in SpatialDecApplyM2() to prevent signed integer overflow.
  * Avoid signed integer overflows in SpatialDecApplyParameterSets() when adding dry and wet signal.
  * Introduce scaling in FDKhybridSynthesisApply() to avoid potential signed integer overflow.
  * Prevent signed integer overflow in fMultIceil() for case m_e=-31.
  * Don't use an enum for a value read directly from the bitstream
  * Revert "Add TEST_MAPPING for aac project"
  * Add TEST_MAPPING for aac project
  * Fix for OPD smoothing in USAC decoder.
  * Move array definition out of for loop in MPEG-D DRC _skipEqInstructions().
  * Use saturate shift to prevent signed integer overflow in multiple qmfInverseModulation*() functions.
  * For all scaleValuesSaturate() variants make source pointer const.
  * Avoid negation of INT_MIN in dst_IV().
  * Add sanity check for huffman escape sequences in HCR tool.
  * Prevent negation of INT_MIN in pcmLimiter_Apply().
  * Adjust scaling in lppTransposerHBE() to prevent signed integer overflow in autocorrelation.
  * Fix signed integer overflow in DuckerCalcEnergy().
  * Avoid signed integer overflow in subbandTPApply() wet energy calculation.
  * Avoid signed integer overflow in subbandTPApply() dry energy calculation.
  * Fix dry and wet energy initialization in subbandTPInit().
  * Add loop abort criterion to prevent timeout in EldSpecificConfig_Parse().
  * Prevent signed integer overflows in ELD filterbank, multE2_DinvF_fdk().
  * Avoid signed integer overflows in multiple adjustTimeSlot*() functions.
  * Improve robustness of limiter band table calculation in ResetLimiterBands().
  * Fix signed integer overflow in pvcDecodeTimeSlot().
  * Use one additional bit headroom to prevent signed integer overflow in BuildAdaptiveExcitation().
  * FDK: USAC CLpd_Reset uses rate constants
  * Initialize scale variable in qmfForwardModulationLP_even() and qmfInverseModulationLP_even() which is updated in dct/fft processing.
  * Limit too large shift value in apply_inter_tes().
  * Fix nrgGain_e update in equalizeFiltBufferExp(). Prevents negative shift exponents in calculateSbrEnvelope().
  * Prevent signed integer overflows in CLpd_FAC_Mdct2Acelp().
  * Fix integer overflow in imlt_block().
  * Prevent signed integer overflows in CLpd_FAC_Acelp2Mdct().
  * Fix integer overflow in complex multiplication called from CLpd_FAC_Acelp2Mdct().
  * Avoid signed integer overflow in bass_pf_1sf_delay().
  * Prevent signed integer overflow in filtLP().
  * Fix signed integer overflow in calcFactorPerBand() calculation.
  * Prevent integer overflow in sbrDecoder_calculateGainVec().
  * Suppress integer overflow in lsf_weight_2st() by using 64 bit multiplication with explicit 32 bit integer result.
  * Prevent signed integer overflow in RE8_PPV().
  * Avoid signed integer overflow in E_LPC_f_lsp_a_conversion().
  * Use SATURATE_LEFT_SHIFT in get_lsppol() to prevent integer overflow.
  * Fix integer overflow in E_UTIL_preemph().
  * Overcome potential integer overflows in M/S module for complex prediction.
  * Avoid integer overflows with pseudoLR in CAacDecoder_DecodeFrame().
  * Validate correct ascLen in CLatmDemux_ReadStreamMuxConfig() to overcome integer overflow in FDK_get32().
  * Restrict size of huffman escape sequence to prevent shift with too large exponent in DecodeEscapeSequence().
  * Limit too large shift exponent in CJointStereo_ApplyMS()
  * Saturate shift value in calculateSbrEnvelope(). Prevents too large shift values.
  * Limit shift exponent in adjustTimeSlot_EldGrid().
  * Fix shift value assignment in addHighBandPart().
  * Replace obsolete SBR CRC with superior FDK CRC implementation. Fixes ELD SBR CRC decoding and prevents negative shift exponents in calcCRC().
  * Solve potential dereference of null pointers in factorCLD().
  * Reject unsupported channel configuration / audio object type combinations.
  * Fix possible stack buffer overflow in mapIndexData().
  * Fix CpePersistentData memory leak.
  * Validate aacSampleRate after applying ELD downscale factor. Avoid division by zero.
  * Fix assert in synchronization() for loas streams with multiple sub frames.
  * Restrict sampling rates parsed from bitstream to maximum of 96kHz.
  * Revert "Reapply: Avoid reading out of bounds due to negative aaIccIndexMapped"
  * Revert "Reapply: Avoid reading out of bounds due to too large aaIidIndexMapped"
  * Revert "Avoid index-out-of-bounds in prepareDrcGain"
  * Revert "Avoid index-out-of-bounds in prepareDrcGain"
  * Revert "Avoid index-out-of-bounds in processDrcTime"
  * Update the version numbers in the public installed headers
  * Integer overflow fix in FDKaacEnc_initAvoidHoleFlag()
  * Fix integer overflow in FDKaacEnc_IntensityStereoProcessing().
  * Improve phase coding accuracy by use of splitAngle lookup table.
  * Improve random phase table accuracy.
  * Increase scale factor gain in TES tool.
  * Improve accuracy of division in calcSubbandGain().
  * Avoid -1.0 autocorrelation input data.
  * Increase scale factor to avoid overflows in function factorCLD().
  * Fix out of sync function declaration of scaleValues().
  * Fetch multiple samples of input channels from encoder input buffer.
  * Fix HE-AAC using MPEG-D DRC with implicit SBR signaling.
  * MPEG-4/D DRC unification, loudness norm default at -24
  * Add support for Album Mode (FDKdec v3.1.0).
  * Reset parameter smoothing tool while concealment.
  * Discard SBR data in case of unsuccessful SBR header update.
  * Trigger config change after a change in SBR activity has been recognized.
  * Clear config in case of parsing error in drcDec_readUniDrc()
  * Add MPEG-D DRC sanity checks
  * Always call startSelectionProcess() independent on an error in FDK_drcDec_ReadUniDrc()
  * AAC decoder: fix use of uninitialized value, check index
  * libSBRdec: prevent OOB access
  * libMpegTPDec: fix use of uninitialized value in OutOfBandConfig
  * Don't explicitly enable CFI in libFraunhoferAAC
  * Always check whether given channel config is supported.
  * FDK sanitization fixes

-------------------------------------------------------------------
Wed Jul 29 11:23:52 UTC 2020 - marec@detebe.org

- Update to version 2.0.1:
  * Bump the versions and add a changelog entry for v2.0.1
  * Avoid index-out-of-bounds in processDrcTime
  * Avoid index-out-of-bounds in prepareDrcGain
  * Avoid index-out-of-bounds in prepareDrcGain
  * Avoid a failed assert on invalid channel indices
  * Remove an unused version declaration from Makefile.vc
  * Fix FDK_FALLTHROUGH for apple clang
  * Add QMF analysis band sanity check
  * Prevent out of bounds accesses in lppTransposer() and lppTransposerHBE()
  * Always check whether given channel config is supported.
  * Unify audio element loop abort criterion in ER syntax
  * Prevent overflow in concealment clipping check
  * Break audio element loop in case element_count becomes too large.
  * Add sampling rate sanity check
  * Prevent bit buffer counter overflow.

-------------------------------------------------------------------
Wed Jul 29 11:15:14 UTC 2020 - marec <marec@detebe.org>

- initial package build