fenix
/
fdk-aac
mirror of https://github.com/mstorsjo/fdk-aac.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity
1,591 Commits 10 Branches 11 Tags 11 MiB
Commit Graph

55 Commits

Author SHA1 Message Date
Martin Storsjo 41ae84525e Merge remote-tracking branch 'aosp/master' 2022-01-10 13:55:34 +02:00
David Seifert 7f328b93ee
Do not force inlining of indirect functions 
* A function called indirectly cannot be decorated
  with `__attribute((always_inline))`, as this is
  guaranteed to only work with direct calls:
    https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63220#c1

Bug: https://bugs.gentoo.org/798045
 2021-07-10 13:06:57 +02:00
Fraunhofer IIS FDK a43c9f8822 Prevent undefined values for CODING_SCHEME type in huff_decode(). 
Bug: 186706541
Test: atest android.media.cts.DecoderTestAacFormat android.media.cts.DecoderTestXheAac android.media.cts.DecoderTestAacDrc
Change-Id: I82da4a5660289d1c96888d48c315f96a4a5c1c2a
 2021-04-30 15:00:27 +00:00
Rosen Penev d284d42e71 fix compilation with mips16 
Some users wrongly pass both -mips16 and -mdsp when compiling on MIPS
platforms. Handle such a case.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
 2021-01-14 14:56:46 +02:00
Martin Storsjo ea21799a7b Revert "Avoid undefined shifts in SATURATE_SHIFT" 
This reverts commit 3a831a5fbc.

This local fix doesn't seem to be needed any more after the latest
upstream update.
 2020-11-17 12:56:58 +02:00
Martin Storsjo c7a3808a2f Merge remote-tracking branch 'aosp/master' 2020-11-17 12:38:51 +02:00
Martin Storsjo 3a831a5fbc Avoid undefined shifts in SATURATE_SHIFT 
Make sure that the shift amount is less than the size of the shifted
value, otherwise return the saturation max values (for left shift, if
the source values was nonzero) or zero (for right shift, or zero
shifted left).

Fixes: 24376/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_LIBFDK_AAC_fuzzer-6529411206348800

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
 2020-09-02 12:56:41 +03:00
Fraunhofer IIS FDK e016635f0d Avoid decoder internal clipping by converting the whole audio sample data path from 16 to 32 bit data width (FDKdec v3.2.0). 
Bug: 149514474
Test: atest DecoderTestXheAac DecoderTestAacDrc
Change-Id: I8a504ab709e42e27a61fe29840212953742283a5
 2020-02-14 10:53:51 -08:00
Fraunhofer IIS FDK 90c29e0808 Revise memory overlay usage and remove deprecated buffers and overlay tags. 
Bug: 149514474
Test: atest DecoderTestXheAac DecoderTestAacDrc
Change-Id: I7aea2898a3c49e06209fae89d734939c100e1184
 2020-02-14 10:44:10 -08:00
Fraunhofer IIS FDK 5482443193 Restructure QMF implementation and provide 16/32 bit implementation in parallel. 
Bug: 149514474
Test: atest DecoderTestXheAac DecoderTestAacDrc
Change-Id: I2c6b332a0b5c4966a3f680d7a4d562d9bf18752c
 2020-02-14 08:51:24 -08:00
Fraunhofer IIS FDK e323bf6cab Avoid undefined shift exponent in f2Pow(). 
Bug: 146938418
Test: atest DecoderTestXheAac ; atest DecoderTestAacDrc
Change-Id: Id28608e049244968158900848c23bf7a8298083d
 2020-01-22 11:59:08 -08:00
Fraunhofer IIS FDK 1020e48d6e Fix IsLessThan() function for certain edge cases. 
Bug: 146936613
Test: atest DecoderTestXheAac ; atest DecoderTestAacDrc
Change-Id: Idbec38c1df01bd7a6a48ac4b6e5673c30627fc73
 2020-01-16 12:15:04 -08:00
Fraunhofer IIS FDK 7019272646 Prevent signed integer overflow in fMultIceil() for case m_e=-31. 
Bug: 146934600
Test: atest DecoderTestXheAac ; atest DecoderTestAacDrc
Change-Id: I8a67a3b17f3ec3af753b6463b72ae2947986b39c
 2020-01-14 14:35:18 -08:00
Fraunhofer IIS FDK 9d40b65b15 For all scaleValuesSaturate() variants make source pointer const. 
Bug: 145669510
Test: atest DecoderTestXheAac ; atest DecoderTestAacDrc
Change-Id: Ib69a4f12405026d3165193f0192eb25eaed7e797
 2019-12-27 10:24:28 -08:00
Martin Storsjo e365f4f278 Merge remote-tracking branch 'aosp/master' 2019-10-08 14:53:07 +03:00
Fraunhofer IIS FDK 1b5457fe93 Fix out of sync function declaration of scaleValues(). 
Bug: 132641988
Test: atest DecoderTestXheAac ; atest DecoderTestAacDrc

Change-Id: I1b9a204c52aee20adabc1baee7f5e5cdbf1ee39e
 2019-05-14 09:13:11 -07:00
Martin Storsjo 02103fd630 Merge remote-tracking branch 'aosp/master' 2019-01-21 15:54:25 +02:00
Fraunhofer IIS FDK edbb4b0a14 Fix SATURATE_LEFT_SHIFT_ALT() since it returned 0x80000000 in certain cases 
Bug: 112661911
Test: atest DecoderTestXheAac ; atest DecoderTestAacDrc
Change-Id: I16e522c975ad45ff8d47d4cabe44f1f2bf95b73b
 2018-12-26 18:17:27 -05:00
Fraunhofer IIS FDK 9744e41c40 Prevent bit buffer counter overflow. 
While long-term test we discovered a bit counter overflow in the bit buffer.
The bit buffer state was only used by HCR and RVLC tool and can easily be substituted with FDKgetValidBits() call.
The following patch completely removes the bit counter and all its obsolete functions.

Bug: 112662184
Test: atest DecoderTestXheAac ; atest DecoderTestAacDrc

Change-Id: Icee0519d26a2aa62367d2dece59cd3d60ffcade7
(cherry picked from commit 15292f7e96)
 2018-09-10 23:19:19 +00:00
Martin Storsjo 651ff34d8d Remove a duplicate definition of fMax/fMin for x86_64 mingw 
FIXP_SGL is a typedef for SHORT, so these extra definitions (which
only were enabled on x86_64 mingw) were duplicates with the fMin/fMax
functions defined for FIXP_SGL right above.
 2018-09-03 10:45:44 +03:00
Martin Storsjo 574bd5e407 Recognize the MSVC ARM64 compiler arch identification 2018-09-02 23:27:57 +03:00
Martin Storsjo b95b15e51d Merge remote-tracking branch 'aosp/master' 2018-09-02 23:16:58 +03:00
Fraunhofer IIS FDK a4d1f0ad52 FDKv2 ubsan patches 
Bug: 80053205
Test: see bug for repro with FB "wow"
   atest DecoderTestAacDrc

Fix signed integer overflows in CLpc_SynthesisLattice()

Change-Id: Icbddfcc8c5fc73382ae5bf8c2a7703802c688e06

Fix signed integer overflows in imlt

Change-Id: I687834fca2f1aab6210ed9862576b4f38fcdeb24

Fix overflow in addLowbandEnergies()

Change-Id: Iaa9fdf9deb49c33ec6ca7ed3081c4ddaa920e9aa

Concealment fix for audio frames containing acelp components

Change-Id: Ibe5e83a6efa75a48f729984a161a76b826878f4e

Fix out-of-bounds access in PS concealment

Change-Id: I08809a03a40d1feaf00e41278db314d67e1efe88

Fix potential memory leak in setup of qmf domain

Change-Id: Id9fc2448354dc7f1b439469128407305efa3def2

Reject channel config 13

Change-Id: Idf5236f6cd054df994e69c9c972c97f6768cf9e5

Fix unsigned integer overflow in configExtension()

Change-Id: I8a1668810b85e6237c3892891444ff08f04b019b

Fix unsigned integer overflow in CAacDecoder_DecodeFrame()

Change-Id: I79678c571690178e6c37680f70a9b94dd3cbc439

Fix unsigned integer overflow in aacDecoder_UpdateBitStreamCounters()

Change-Id: I3bff959da9f53fabb18cd0ae6c260e6256194526

Fix unsigned integer overflow in transportDec_readStream()

Change-Id: I6a6f9f4acaa32fae0b5de9641f8787bbc7f8286b
 2018-05-23 12:19:40 -07:00
Fraunhofer IIS FDK 44ac411683 FDK patches: fix overflows in decoder out-of-band config 
Bug: 71430241
Bug: 79220129
Test: cts-tradefed run commandAndExit cts-dev -m CtsMediaTestCases -t android.media.cts.DecoderTestXheAac
      cts-tradefed run commandAndExit cts-dev -m CtsMediaTestCases -t android.media.cts.DecoderTestAacDrc

Unsigned Integer Overflows in CDataStreamElement_Read()

Change-Id: Ic2f5b3ae111bf984d4d0db664823798957b0a979

Unsigned Integer Overflow in CProgramConfig_ReadHeightExt()

Change-Id: Iaebc458bb59504203e604a28ed6d5cecaa875c42

Unsigned Integer Overflow in transportDec_OutOfBandConfig()

Change-Id: I24a4b32d736f28c55147f0e2ca06fe5537da19c2

Unsigned Integer Overflows in CDKcrcEndReg() & crcCalc()

Change-Id: I6ebbe541a4d3b6bacbd5ace17264972951de7ca8

Unsigned Integer Overflows in ReadPsData()

Change-Id: Id36576fe545236860a06f17971494ecd4484c494

Unsigned Integer Overflow in SpatialDecParseSpecificConfig()

Change-Id: Ib468f129a951c69776b88468407f008ab4cfd2c7

Unsigned Integer Overflows in _readUniDrcConfigExtension() & _readLoudnessInfoSetExtension()

Change-Id: Ibcf7c6a23af49239206ea9301c58adac36e3ceba
 2018-05-09 15:15:28 -07:00
Fraunhofer IIS FDK 6cfabd3536 Upgrade to FDKv2 
Bug: 71430241
Test: CTS DecoderTest and DecoderTestAacDrc

original-Change-Id: Iaa20f749b8a04d553b20247cfe1a8930ebbabe30

Apply clang-format also on header files.

original-Change-Id: I14de1ef16bbc79ec0283e745f98356a10efeb2e4

Fixes for MPEG-D DRC

original-Change-Id: If1de2d74bbbac84b3f67de3b88b83f6a23b8a15c

Catch unsupported tw_mdct at an early stage

original-Change-Id: Ied9dd00d754162a0e3ca1ae3e6b854315d818afe

Fixing PVC transition frames

original-Change-Id: Ib75725abe39252806c32d71176308f2c03547a4e

Move qmf bands sanity check

original-Change-Id: Iab540c3013c174d9490d2ae100a4576f51d8dbc4

Initialize scaling variable

original-Change-Id: I3c4087101b70e998c71c1689b122b0d7762e0f9e

Add 16 qmf band configuration to getSlotNrgHQ()

original-Change-Id: I49a5d30f703a1b126ff163df9656db2540df21f1

Always apply byte alignment at the end of the AudioMuxElement

original-Change-Id: I42d560287506d65d4c3de8bfe3eb9a4ebeb4efc7

Setup SBR element only if no parse error exists

original-Change-Id: I1915b73704bc80ab882b9173d6bec59cbd073676

Additional array index check in HCR

original-Change-Id: I18cc6e501ea683b5009f1bbee26de8ddd04d8267

Fix fade-in index selection in concealment module

original-Change-Id: Ibf802ed6ed8c05e9257e1f3b6d0ac1162e9b81c1

Enable explicit backward compatible parser for AAC_LD

original-Change-Id: I27e9c678dcb5d40ed760a6d1e06609563d02482d

Skip spatial specific config in explicit backward compatible ASC

original-Change-Id: Iff7cc365561319e886090cedf30533f562ea4d6e

Update flags description in decoder API

original-Change-Id: I9a5b4f8da76bb652f5580cbd3ba9760425c43830

Add QMF domain reset function

original-Change-Id: I4f89a8a2c0277d18103380134e4ed86996e9d8d6

DRC upgrade v2.1.0

original-Change-Id: I5731c0540139dab220094cd978ef42099fc45b74

Fix integer overflow in sqrtFixp_lookup()

original-Change-Id: I429a6f0d19aa2cc957e0f181066f0ca73968c914

Fix integer overflow in invSqrtNorm2()

original-Change-Id: I84de5cbf9fb3adeb611db203fe492fabf4eb6155

Fix integer overflow in GenerateRandomVector()

original-Change-Id: I3118a641008bd9484d479e5b0b1ee2b5d7d44d74

Fix integer overflow in adjustTimeSlot_EldGrid()

original-Change-Id: I29d503c247c5c8282349b79df940416a512fb9d5

Fix integer overflow in FDKsbrEnc_codeEnvelope()

original-Change-Id: I6b34b61ebb9d525b0c651ed08de2befc1f801449

Follow-up on: Fix integer overflow in adjustTimeSlot_EldGrid()

original-Change-Id: I6f8f578cc7089e5eb7c7b93e580b72ca35ad689a

Fix integer overflow in get_pk_v2()

original-Change-Id: I63375bed40d45867f6eeaa72b20b1f33e815938c

Fix integer overflow in Syn_filt_zero()

original-Change-Id: Ie0c02fdfbe03988f9d3b20d10cd9fe4c002d1279

Fix integer overflow in CFac_CalcFacSignal()

original-Change-Id: Id2d767c40066c591b51768e978eb8af3b803f0c5

Fix integer overflow in FDKaacEnc_FDKaacEnc_calcPeNoAH()

original-Change-Id: Idcbd0f4a51ae2550ed106aa6f3d678d1f9724841

Fix integer overflow in sbrDecoder_calculateGainVec()

original-Change-Id: I7081bcbe29c5cede9821b38d93de07c7add2d507

Fix integer overflow in CLpc_SynthesisLattice()

original-Change-Id: I4a95ddc18de150102352d4a1845f06094764c881

Fix integer overflow in Pred_Lt4()

original-Change-Id: I4dbd012b2de7d07c3e70a47b92e3bfae8dbc750a

Fix integer overflow in FDKsbrEnc_InitSbrFastTransientDetector()

original-Change-Id: I788cbec1a4a00f44c2f3a72ad7a4afa219807d04

Fix unsigned integer overflow in FDKaacEnc_WriteBitstream()

original-Change-Id: I68fc75166e7d2cd5cd45b18dbe3d8c2a92f1822a

Fix unsigned integer overflow in FDK_MetadataEnc_Init()

original-Change-Id: Ie8d025f9bcdb2442c704bd196e61065c03c10af4

Fix overflow in pseudo random number generators

original-Change-Id: I3e2551ee01356297ca14e3788436ede80bd5513c

Fix unsigned integer overflow in sbrDecoder_Parse()

original-Change-Id: I3f231b2f437e9c37db4d5b964164686710eee971

Fix unsigned integer overflow in longsub()

original-Change-Id: I73c2bc50415cac26f1f5a29e125bbe75f9180a6e

Fix unsigned integer overflow in CAacDecoder_DecodeFrame()

original-Change-Id: Ifce2db4b1454b46fa5f887e9d383f1cc43b291e4

Fix overflow at CLpdChannelStream_Read()

original-Change-Id: Idb9d822ce3a4272e4794b643644f5434e2d4bf3f

Fix unsigned integer overflow in Hcr_State_BODY_SIGN_ESC__ESC_WORD()

original-Change-Id: I1ccf77c0015684b85534c5eb97162740a870b71c

Fix unsigned integer overflow in UsacConfig_Parse()

original-Change-Id: Ie6d27f84b6ae7eef092ecbff4447941c77864d9f

Fix unsigned integer overflow in aacDecoder_drcParse()

original-Change-Id: I713f28e883eea3d70b6fa56a7b8f8c22bcf66ca0

Fix unsigned integer overflow in aacDecoder_drcReadCompression()

original-Change-Id: Ia34dfeb88c4705c558bce34314f584965cafcf7a

Fix unsigned integer overflow in CDataStreamElement_Read()

original-Change-Id: Iae896cc1d11f0a893d21be6aa90bd3e60a2c25f0

Fix unsigned integer overflow in transportDec_AdjustEndOfAccessUnit()

original-Change-Id: I64cf29a153ee784bb4a16fdc088baabebc0007dc

Fix unsigned integer overflow in transportDec_GetAuBitsRemaining()

original-Change-Id: I975b3420faa9c16a041874ba0db82e92035962e4

Fix unsigned integer overflow in extractExtendedData()

original-Change-Id: I2a59eb09e2053cfb58dfb75fcecfad6b85a80a8f

Fix signed integer overflow in CAacDecoder_ExtPayloadParse()

original-Change-Id: I4ad5ca4e3b83b5d964f1c2f8c5e7b17c477c7929

Fix unsigned integer overflow in CAacDecoder_DecodeFrame()

original-Change-Id: I29a39df77d45c52a0c9c5c83c1ba81f8d0f25090

Follow-up on: Fix integer overflow in CLpc_SynthesisLattice()

original-Change-Id: I8fb194ffc073a3432a380845be71036a272d388f

Fix signed integer overflow in _interpolateDrcGain()

original-Change-Id: I879ec9ab14005069a7c47faf80e8bc6e03d22e60

Fix unsigned integer overflow in FDKreadBits()

original-Change-Id: I1f47a6a8037ff70375aa8844947d5681bb4287ad

Fix unsigned integer overflow in FDKbyteAlign()

original-Change-Id: Id5f3a11a0c9e50fc6f76ed6c572dbd4e9f2af766

Fix unsigned integer overflow in FDK_get32()

original-Change-Id: I9d33b8e97e3d38cbb80629cb859266ca0acdce96

Fix unsigned integer overflow in FDK_pushBack()

original-Change-Id: Ic87f899bc8c6acf7a377a8ca7f3ba74c3a1e1c19

Fix unsigned integer overflow in FDK_pushForward()

original-Change-Id: I3b754382f6776a34be1602e66694ede8e0b8effc

Fix unsigned integer overflow in ReadPsData()

original-Change-Id: I25361664ba8139e32bbbef2ca8c106a606ce9c37

Fix signed integer overflow in E_UTIL_residu()

original-Change-Id: I8c3abd1f437ee869caa8fb5903ce7d3d641b6aad

REVERT: Follow-up on: Integer overflow in CLpc_SynthesisLattice().

original-Change-Id: I3d340099acb0414795c8dfbe6362bc0a8f045f9b

Follow-up on: Fix integer overflow in CLpc_SynthesisLattice()

original-Change-Id: I4aedb8b3a187064e9f4d985175aa55bb99cc7590

Follow-up on: Fix unsigned integer overflow in aacDecoder_drcParse()

original-Change-Id: I2aa2e13916213bf52a67e8b0518e7bf7e57fb37d

Fix integer overflow in acelp

original-Change-Id: Ie6390c136d84055f8b728aefbe4ebef6e029dc77

Fix unsigned integer overflow in aacDecoder_UpdateBitStreamCounters()

original-Change-Id: I391ffd97ddb0b2c184cba76139bfb356a3b4d2e2

Adjust concealment default settings

original-Change-Id: I6a95db935a327c47df348030bcceafcb29f54b21

Saturate estimatedStartPos

original-Change-Id: I27be2085e0ae83ec9501409f65e003f6bcba1ab6

Negative shift exponent in _interpolateDrcGain()

original-Change-Id: I18edb26b26d002aafd5e633d4914960f7a359c29

Negative shift exponent in calculateICC()

original-Change-Id: I3dcd2ae98d2eb70ee0d59750863cbb2a6f4f8aba

Too large shift exponent in FDK_put()

original-Change-Id: Ib7d9aaa434d2d8de4a13b720ca0464b31ca9b671

Too large shift exponent in CalcInvLdData()

original-Change-Id: I43e6e78d4cd12daeb1dcd5d82d1798bdc2550262

Member access within null pointer of type SBR_CHANNEL

original-Change-Id: Idc5e4ea8997810376d2f36bbdf628923b135b097

Member access within null pointer of type CpePersistentData

original-Change-Id: Ib6c91cb0d37882768e5baf63324e429589de0d9d

Member access within null pointer FDKaacEnc_psyMain()

original-Change-Id: I7729b7f4479970531d9dc823abff63ca52e01997

Member access within null pointer FDKaacEnc_GetPnsParam()

original-Change-Id: I9aa3b9f3456ae2e0f7483dbd5b3dde95fc62da39

Member access within null pointer FDKsbrEnc_EnvEncodeFrame()

original-Change-Id: I67936f90ea714e90b3e81bc0dd1472cc713eb23a

Add HCR sanity check

original-Change-Id: I6c1d9732ebcf6af12f50b7641400752f74be39f7

Fix memory issue for HBE edge case with 8:3 SBR

original-Change-Id: I11ea58a61e69fbe8bf75034b640baee3011e63e9

Additional SBR parametrization sanity check for ELD

original-Change-Id: Ie26026fbfe174c2c7b3691f6218b5ce63e322140

Add MPEG-D DRC channel layout check

original-Change-Id: Iea70a74f171b227cce636a9eac4ba662777a2f72

Additional out-of-bounds checks in MPEG-D DRC

original-Change-Id: Ife4a8c3452c6fde8a0a09e941154a39a769777d4

Change-Id: Ic63cb2f628720f54fe9b572b0cb528e2599c624e
 2018-04-19 11:21:15 -07:00
Martin Storsjo 5891b75b31 Fix warnings with MSVC about mismatch between signed and unsigned in a header 2018-03-06 12:49:20 +02:00
Martin Storsjo 52c2660c26 Make sure at least one bit exists before reading further in FDKreadBit 
Fixes: 2709/clusterfuzz-testcase-minimized-6160249369133056

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
 2017-08-03 14:22:30 +03:00
Martin Storsjo 50922e3dbd Try to properly handle the case when the bitstream reader runs out of bits to read 2017-04-26 23:37:11 +03:00
Robert Kausch 6b0d8201b1 Add checks to avoid overreading supplied buffers and fix issue #61. 2017-04-23 21:31:36 +02:00
Martin Storsjo 72037da0ba Fix building with MSVC for ARM 
cmnintrin.h only existed for Windows CE, and doesn't seem to actually
be necessary for building this.
 2016-09-11 13:28:47 +03:00
Martin Storsjo 45915b67e4 Detect MSVC x64 2016-09-11 13:28:44 +03:00
Lexyan 1d686c3a23 Add aarch64 assembly optimization (ARMv8a 64 bits) 
The fixmuldiv functions don't need inline assembly to be fast
in this architecture; the compiler (both clang and GCC) figure
out to use the optimal instructions for this (which is 2 instruction
sequence), and when letting the compiler emit the instructions
instead of using inline assembly, the compiler is able to
interleave those instructions with other instructions,
improving scheduling, making it even faster than when using
inline assembly.

Overall, this gives about 50% speedup.
 2016-09-10 14:33:39 +03:00
Martin Storsjo f9d0f65e65 Merge remote-tracking branch 'aosp/master' 2016-08-25 22:12:15 +03:00
Guillaume Smaha f132ac5f3d Invert inline and void to be compatible with gcc 3.3.4 2016-05-18 08:07:32 +02:00
Jean-Michel Trivi 46ba3676b8 AAC/SBR encoder improvements 
* AAC-Encoder

  - AAC-ELD core encoder audio quality tuning. Update tuning tables, configure
    bitreservoir size and adapt afterburner iteration value.
     Modified file(s):
        libAACenc/src/aacenc.h
        libAACenc/src/aacenc_lib.cpp
        libAACenc/src/adj_thr.cpp
        libAACenc/src/adj_thr.h
        libAACenc/src/adj_thr_data.h
        libAACenc/src/bandwidth.cpp
        libAACenc/src/pnsparam.cpp
        libAACenc/src/qc_main.cpp

   - Introduze dead zone quantizer for ELD to improve audio quality at certain
     configurations.
     Modified file(s):
        libAACenc/src/aacenc_lib.cpp
        libAACenc/src/adj_thr.cpp
        libAACenc/src/adj_thr.h
        libAACenc/src/qc_data.h
        libAACenc/src/qc_main.cpp
        libAACenc/src/quantize.cpp
        libAACenc/src/quantize.h
        libAACenc/src/sf_estim.cpp
        libAACenc/src/sf_estim.h

   - Revise TNS module to improve ELD audio quality.
   - Use new window function and separate prediction gain according TNS filters.
   - Add missing memory initilization to TNS configuration.
     Modified file(s):
        libAACenc/src/aacenc_lib.cpp
        libAACenc/src/aacenc_tns.cpp
        libAACenc/src/aacenc_tns.h
        libAACenc/src/psy_main.cpp
        libAACenc/src/tns_func.h

* SBR-Encoder

   - Revise frequency resolution calculation and handle differently depending
     on number of envelopes and split frames decision.
   - Add and adjust ELD SBR tuning tables.
     Modified file(s):
        libSBRenc/include/sbr_encoder.h
        libSBRenc/src/bit_sbr.h
        libSBRenc/src/env_est.cpp
        libSBRenc/src/fram_gen.cpp
        libSBRenc/src/fram_gen.h
        libSBRenc/src/mh_det.cpp
        libSBRenc/src/sbr_def.h
        libSBRenc/src/sbr_encoder.cpp
        libSBRenc/src/sbr_rom.cpp
        libSBRenc/src/tran_det.cpp

   - Replace ELD transient detector with fast implementation.
     Modified file(s):
        libSBRenc/src/env_est.cpp
        libSBRenc/src/env_est.h
        libSBRenc/src/fram_gen.cpp
        libSBRenc/src/sbr_def.h
        libSBRenc/src/sbr_encoder.cpp
        libSBRenc/src/tran_det.cpp
        libSBRenc/src/tran_det.h

* FDK-Library
   - Introduce generic compare function in tools library.
     Modified file(s):
        libFDK/include/fixpoint_math.h
        libFDK/src/FDK_core.cpp

* SBR-Encoder

  -  Revise ELD frame splitter to improve bit distribution.
     Modified file(s):
        libSBRenc/include/sbr_encoder.h
        libSBRenc/src/bit_sbr.h
        libSBRenc/src/env_est.cpp
        libSBRenc/src/fram_gen.cpp
        libSBRenc/src/fram_gen.h
        libSBRenc/src/sbr_encoder.cpp
        libSBRenc/src/tran_det.cpp
        libSBRenc/src/tran_det.h

   - Configure amplitude resolution according the tonality of the audio signal.
     Modified file(s):
        libSBRenc/include/sbr_encoder.h
        libSBRenc/src/bit_sbr.h
        libSBRenc/src/env_est.cpp
        libSBRenc/src/nf_est.cpp
        libSBRenc/src/nf_est.h
        libSBRenc/src/sbr_def.h
        libSBRenc/src/sbr_encoder.cpp
        libSBRenc/src/ton_corr.cpp
        libSBRenc/src/ton_corr.h
        libSBRenc/src/tran_det.cpp
        libSBRenc/src/tran_det.h

Change-Id: Ie0672b989a06ee63b50240616b8d1d4b790b6cb2
 2016-04-08 10:52:42 -07:00
Jean-Michel Trivi e1c78ed73f SBR/AAC encoder updates, code clean up 
* SBR-Encoder

   - Prevent noise level overflow in noise floor detection.
   - Saturate threshold calculation in transient detection.
     Modified file(s):
        libSBRenc/src/nf_est.cpp
        libSBRenc/src/sbr_encoder.cpp
        libSBRenc/src/tran_det.cpp

* AAC-Encoder

   - Expand input data range of GetInvInt() function. There was an encoder
     assert observed in non-default bitrate configuration.
     Modified file(s):
        libAACenc/src/aacenc_lib.cpp
        libAACenc/src/intensity.cpp
        libFDK/include/fixpoint_math.h
        libFDK/src/FDK_core.cpp
        libFDK/src/FDK_tools_rom.cpp

   - Make sure that the encoder is stable with regard to very low audio bandwidth
     confguration parameter value.
   - Fix lowdelay blending for low audio bandwidth.
     Modified file(s):
        libAACenc/src/aacenc.cpp
        libAACenc/src/aacenc_lib.cpp
        libAACenc/src/adj_thr.cpp
        libAACenc/src/psy_configuration.cpp
        libAACenc/src/psy_main.cpp

   - Disable pseudo surround flag in case metadata matrix mixdown index is
     present in program config element.
     Modified file(s):
        libAACenc/src/aacenc_lib.cpp

   - Enable variable bitrate mode in encoder api.
   - Add AACENC_PEAK_BITRATE parameter to encoder api.
   - Add AACENC_AUDIOMUXVER parameter to encoder api.
     Modified file(s):
        libAACenc/include/aacenc_lib.h
        libAACenc/src/aacenc.cpp
        libAACenc/src/aacenc.h
        libAACenc/src/aacenc_lib.cpp
        libAACenc/src/qc_main.cpp
        libMpegTPEnc/src/tpenc_latm.cpp
        libMpegTPEnc/src/version

* FDK-Sources

   - Code clean up. Remove unneeded pseudo audio object types and transport types.
     Modified file(s):
        libAACdec/src/aacdecoder.cpp
        libAACdec/src/aacdecoder_lib.cpp
        libAACenc/include/aacenc_lib.h
        libAACenc/src/aacenc.cpp
        libAACenc/src/aacenc_lib.cpp
        libFDK/src/FDK_tools_rom.cpp
        libMpegTPDec/src/tpdec_lib.cpp
        libMpegTPDec/src/version
        libMpegTPEnc/src/tpenc_latm.cpp
        libMpegTPEnc/src/version
        libSBRdec/src/sbrdecoder.cpp
        libSBRenc/src/sbr_encoder.cpp
        libSYS/include/FDK_audio.h
        libSYS/src/genericStds.cpp

Change-Id: I807a53cb7f48c9ee7563cb8da1d0c52221576ca6
 2016-04-04 17:25:36 -07:00
Vicente Olivert Riera 047376af8c libFDK/include/mips/cplx_mul.h: use C code instead of ASM 
Replace the assembler code by the equivalent in C. This way it will work
for every MIPS ISA. Now the check for __mips_isa_rev < 6 is no longer
necessary, so remove it.

Signed-off-by: Vicente Olivert Riera <Vincent.Riera@imgtec.com>
 2015-09-21 11:59:02 +01:00
Jasper Lievisse Adriaanse 2585f10da2 Fix the preprocessor statements controlling the use of mips{32,64}r1 instructions. 
Only enable code using mips32/mips64 instructions if the compiler is targetting
this ISA. (integer madd and msub instructions aren't available in the
``canonical'' mips ISAs)
 2015-08-13 10:53:07 +02:00
Martin Storsjo 6ccecbc4b0 Merge remote-tracking branch 'aosp/master' 2014-11-07 09:32:48 +02:00
Martin Storsjo 9a3234055a Extend the GetInvInt table to 80 elements 
This is a bug fix patch from Fraunhofer, amending/replacing/extending
the previous commit which only extended the table to 55 elements.
(In my previous fix, 55 seemed to be enough for the issue at hand at
least. A few of the coefficients had slightly different rounding than
what Fraunhofer had used though).

This will show up in AOSP in the next code drop as well.
 2014-08-01 10:36:54 +03:00
Duane Sand 5c51a5898e [MIPSR6] Skip assembler code using MFHI/MFLO on mips32r6 
Change-Id: I3dacd96cf9d5cf9c3d34d612ebb0456d64bc23bc
 2014-07-31 15:40:38 -07:00
Martin Storsjo 6522e87ad6 Extend GetInvInt to support values up to 54 
For sample rates >= 64 kHz with short blocks, the largest sfb
width is 36, which after grouping can be multiplied with 3,
ending up as 108. GetInvInt needs to be able to support
the largest grouped sfb width, divided by two (3 * 36 / 2 = 54).

This fixes VBR encoding with sample rates >= 64 kHz.
 2014-07-29 15:00:56 +03:00
Mark Harris c15b653b99 Add ppc optimization for clz and fixmul 
Optimize ppc/ppc64 similar to i386/x86_64, approximately doubling
encoding speed on a Power Mac G5 running OS X 10.5.8.
 2014-05-26 12:23:56 -07:00
Mark Harris 5e154c7ed1 Make LONG 32 bits for ppc64 as it is for x86_64 2014-05-26 12:14:40 -07:00
Mark Harris 72fd73dbdd Fix PowerPC Mac build 2014-05-26 12:12:03 -07:00
Glenn Kasten 35f30c5ab8 Merge "AArch64: Make LONG 4 bytes" 2014-04-09 15:49:10 +00:00
Glenn Kasten 8609bfea36 Temporary workaround for 64-bit build error 
Replace #error by #warning so we can build 64-bit

Change-Id: Ib52fbfb0e5e945ff04ba51ae06b4e3393886d7b7
 2014-03-12 11:19:32 -07:00
Marcus Oakland f2f96fb86d AArch64: Make LONG 4 bytes 
The libSYS/include/machine_type.h header file states LONG should be
"Data type representing 4 byte signed integer on all supported
platforms" but only provided for defining LONG as INT and ULONG as
UINT when __x86_64__ was defined. This has been changed to when
__LP64__ is defined, so that it also applies to AArch64. The change
to libFDK/include/common_fix.h is then needed to get the project to
compile.

Change-Id: Iea42d7eca97dcc9da772a05b207d134cb999a72a
Signed-off-by: Marcus Oakland <marcus.oakland@arm.com>
 2014-03-11 12:01:00 +00:00
Jean-Michel Trivi 3a0a695565 FDK update for downsampled SBR 
Add missing support for downsampled SBR in libFDK

Bug 9428126

Change-Id: Idb732f8d31a115d36dd4b22916599db7fab98cae
 2013-08-27 15:57:28 -07:00
Jean-Michel Trivi 4f0d97057c Disclaimer update, lib version increment 
- Update disclaimer in all source files.
- Increment library version in each FDK sub-module.

Bug 9428126

Change-Id: I490b96d4ee472246b01483202b0bb4f1e9c2a5d7
 2013-08-20 14:38:55 -07:00