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

1 files changed, 480 insertions, 0 deletions

diff --git a/libFDK/src/FDK_lpc.cpp b/libFDK/src/FDK_lpc.cpp
new file mode 100644
index 0000000..2ba7707
--- /dev/null
+++ b/libFDK/src/FDK_lpc.cpp
@@ -0,0 +1,480 @@
+/* -----------------------------------------------------------------------------
+Software License for The Fraunhofer FDK AAC Codec Library for Android
+
+© Copyright  1995 - 2018 Fraunhofer-Gesellschaft zur Förderung der angewandten
+Forschung e.V. All rights reserved.
+
+ 1.    INTRODUCTION
+The Fraunhofer FDK AAC Codec Library for Android ("FDK AAC Codec") is software
+that implements the MPEG Advanced Audio Coding ("AAC") encoding and decoding
+scheme for digital audio. This FDK AAC Codec software is intended to be used on
+a wide variety of Android devices.
+
+AAC's HE-AAC and HE-AAC v2 versions are regarded as today's most efficient
+general perceptual audio codecs. AAC-ELD is considered the best-performing
+full-bandwidth communications codec by independent studies and is widely
+deployed. AAC has been standardized by ISO and IEC as part of the MPEG
+specifications.
+
+Patent licenses for necessary patent claims for the FDK AAC Codec (including
+those of Fraunhofer) may be obtained through Via Licensing
+(www.vialicensing.com) or through the respective patent owners individually for
+the purpose of encoding or decoding bit streams in products that are compliant
+with the ISO/IEC MPEG audio standards. Please note that most manufacturers of
+Android devices already license these patent claims through Via Licensing or
+directly from the patent owners, and therefore FDK AAC Codec software may
+already be covered under those patent licenses when it is used for those
+licensed purposes only.
+
+Commercially-licensed AAC software libraries, including floating-point versions
+with enhanced sound quality, are also available from Fraunhofer. Users are
+encouraged to check the Fraunhofer website for additional applications
+information and documentation.
+
+2.    COPYRIGHT LICENSE
+
+Redistribution and use in source and binary forms, with or without modification,
+are permitted without payment of copyright license fees provided that you
+satisfy the following conditions:
+
+You must retain the complete text of this software license in redistributions of
+the FDK AAC Codec or your modifications thereto in source code form.
+
+You must retain the complete text of this software license in the documentation
+and/or other materials provided with redistributions of the FDK AAC Codec or
+your modifications thereto in binary form. You must make available free of
+charge copies of the complete source code of the FDK AAC Codec and your
+modifications thereto to recipients of copies in binary form.
+
+The name of Fraunhofer may not be used to endorse or promote products derived
+from this library without prior written permission.
+
+You may not charge copyright license fees for anyone to use, copy or distribute
+the FDK AAC Codec software or your modifications thereto.
+
+Your modified versions of the FDK AAC Codec must carry prominent notices stating
+that you changed the software and the date of any change. For modified versions
+of the FDK AAC Codec, the term "Fraunhofer FDK AAC Codec Library for Android"
+must be replaced by the term "Third-Party Modified Version of the Fraunhofer FDK
+AAC Codec Library for Android."
+
+3.    NO PATENT LICENSE
+
+NO EXPRESS OR IMPLIED LICENSES TO ANY PATENT CLAIMS, including without
+limitation the patents of Fraunhofer, ARE GRANTED BY THIS SOFTWARE LICENSE.
+Fraunhofer provides no warranty of patent non-infringement with respect to this
+software.
+
+You may use this FDK AAC Codec software or modifications thereto only for
+purposes that are authorized by appropriate patent licenses.
+
+4.    DISCLAIMER
+
+This FDK AAC Codec software is provided by Fraunhofer on behalf of the copyright
+holders and contributors "AS IS" and WITHOUT ANY EXPRESS OR IMPLIED WARRANTIES,
+including but not limited to the implied warranties of merchantability and
+fitness for a particular purpose. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR
+CONTRIBUTORS BE LIABLE for any direct, indirect, incidental, special, exemplary,
+or consequential damages, including but not limited to procurement of substitute
+goods or services; loss of use, data, or profits, or business interruption,
+however caused and on any theory of liability, whether in contract, strict
+liability, or tort (including negligence), arising in any way out of the use of
+this software, even if advised of the possibility of such damage.
+
+5.    CONTACT INFORMATION
+
+Fraunhofer Institute for Integrated Circuits IIS
+Attention: Audio and Multimedia Departments - FDK AAC LL
+Am Wolfsmantel 33
+91058 Erlangen, Germany
+
+www.iis.fraunhofer.de/amm
+amm-info@iis.fraunhofer.de
+----------------------------------------------------------------------------- */
+
+/******************* Library for basic calculation routines ********************
+
+   Author(s):   Manuel Jander
+
+   Description: LPC related functions
+
+*******************************************************************************/
+
+#include "FDK_lpc.h"
+
+/* Internal scaling of LPC synthesis to avoid overflow of filte states.
+   This depends on the LPC order, because the LPC order defines the amount
+   of MAC operations. */
+static SCHAR order_ld[LPC_MAX_ORDER] = {
+    /* Assume that Synthesis filter output does not clip and filter
+       accu does change no more than 1.0 for each iteration.
+       ceil(0.5*log((1:24))/log(2)) */
+    0, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3};
+
+/* IIRLattice */
+#ifndef FUNCTION_CLpc_SynthesisLattice_SGL
+void CLpc_SynthesisLattice(FIXP_DBL *signal, const int signal_size,
+                           const int signal_e, const int signal_e_out,
+                           const int inc, const FIXP_SGL *coeff,
+                           const int order, FIXP_DBL *state) {
+  int i, j;
+  FIXP_DBL *pSignal;
+  int shift;
+
+  FDK_ASSERT(order <= LPC_MAX_ORDER);
+  FDK_ASSERT(order > 0);
+
+  if (inc == -1)
+    pSignal = &signal[signal_size - 1];
+  else
+    pSignal = &signal[0];
+
+  /*
+    tmp = x(k) - K(M)*g(M);
+    for m=M-1:-1:1
+            tmp = tmp - K(m) * g(m);
+            g(m+1) = g(m) + K(m) * tmp;
+    endfor
+    g(1) = tmp;
+
+    y(k) = tmp;
+  */
+
+  shift = -order_ld[order - 1];
+
+  for (i = signal_size; i != 0; i--) {
+    FIXP_DBL *pState = state + order - 1;
+    const FIXP_SGL *pCoeff = coeff + order - 1;
+    FIXP_DBL tmp;
+
+    tmp = scaleValue(*pSignal, shift + signal_e) -
+          fMultDiv2(*pCoeff--, *pState--);
+    for (j = order - 1; j != 0; j--) {
+      tmp = fMultSubDiv2(tmp, pCoeff[0], pState[0]);
+      pState[1] = pState[0] + (fMultDiv2(*pCoeff--, tmp) << 2);
+      pState--;
+    }
+
+    *pSignal = scaleValueSaturate(tmp, -shift - signal_e_out);
+
+    /* exponent of state[] is -1 */
+    pState[1] = tmp << 1;
+    pSignal += inc;
+  }
+}
+#endif
+
+#ifndef FUNCTION_CLpc_SynthesisLattice_DBL
+void CLpc_SynthesisLattice(FIXP_DBL *signal, const int signal_size,
+                           const int signal_e, const int signal_e_out,
+                           const int inc, const FIXP_DBL *coeff,
+                           const int order, FIXP_DBL *state) {
+  int i, j;
+  FIXP_DBL *pSignal;
+
+  FDK_ASSERT(order <= LPC_MAX_ORDER);
+  FDK_ASSERT(order > 0);
+
+  if (inc == -1)
+    pSignal = &signal[signal_size - 1];
+  else
+    pSignal = &signal[0];
+
+  FDK_ASSERT(signal_size > 0);
+  for (i = signal_size; i != 0; i--) {
+    FIXP_DBL *pState = state + order - 1;
+    const FIXP_DBL *pCoeff = coeff + order - 1;
+    FIXP_DBL tmp;
+
+    tmp = scaleValue(*pSignal, signal_e) - fMult(*pCoeff--, *pState--);
+    for (j = order - 1; j != 0; j--) {
+      tmp = tmp - fMult(pCoeff[0], pState[0]);
+      pState[1] = pState[0] + fMult(*pCoeff--, tmp);
+      pState--;
+    }
+
+    *pSignal = scaleValue(tmp, -signal_e_out);
+
+    /* exponent of state[] is 0 */
+    pState[1] = tmp;
+    pSignal += inc;
+  }
+}
+
+#endif
+
+/* LPC_SYNTHESIS_IIR version */
+void CLpc_Synthesis(FIXP_DBL *signal, const int signal_size, const int signal_e,
+                    const int inc, const FIXP_LPC_TNS *lpcCoeff_m,
+                    const int lpcCoeff_e, const int order, FIXP_DBL *state,
+                    int *pStateIndex) {
+  int i, j;
+  FIXP_DBL *pSignal;
+  int stateIndex = *pStateIndex;
+
+  FIXP_LPC_TNS coeff[2 * LPC_MAX_ORDER];
+  FDKmemcpy(&coeff[0], lpcCoeff_m, order * sizeof(FIXP_LPC_TNS));
+  FDKmemcpy(&coeff[order], lpcCoeff_m, order * sizeof(FIXP_LPC_TNS));
+
+  FDK_ASSERT(order <= LPC_MAX_ORDER);
+  FDK_ASSERT(stateIndex < order);
+
+  if (inc == -1)
+    pSignal = &signal[signal_size - 1];
+  else
+    pSignal = &signal[0];
+
+  /* y(n) = x(n) - lpc[1]*y(n-1) - ... - lpc[order]*y(n-order) */
+
+  for (i = 0; i < signal_size; i++) {
+    FIXP_DBL x;
+    const FIXP_LPC_TNS *pCoeff = coeff + order - stateIndex;
+
+    x = scaleValue(*pSignal, -(lpcCoeff_e + 1));
+    for (j = 0; j < order; j++) {
+      x -= fMultDiv2(state[j], pCoeff[j]);
+    }
+    x = SATURATE_SHIFT(x, -lpcCoeff_e - 1, DFRACT_BITS);
+
+    /* Update states */
+    stateIndex = ((stateIndex - 1) < 0) ? (order - 1) : (stateIndex - 1);
+    state[stateIndex] = x;
+
+    *pSignal = scaleValue(x, signal_e);
+    pSignal += inc;
+  }
+
+  *pStateIndex = stateIndex;
+}
+/* default version */
+void CLpc_Synthesis(FIXP_DBL *signal, const int signal_size, const int signal_e,
+                    const int inc, const FIXP_LPC *lpcCoeff_m,
+                    const int lpcCoeff_e, const int order, FIXP_DBL *state,
+                    int *pStateIndex) {
+  int i, j;
+  FIXP_DBL *pSignal;
+  int stateIndex = *pStateIndex;
+
+  FIXP_LPC coeff[2 * LPC_MAX_ORDER];
+  FDKmemcpy(&coeff[0], lpcCoeff_m, order * sizeof(FIXP_LPC));
+  FDKmemcpy(&coeff[order], lpcCoeff_m, order * sizeof(FIXP_LPC));
+
+  FDK_ASSERT(order <= LPC_MAX_ORDER);
+  FDK_ASSERT(stateIndex < order);
+
+  if (inc == -1)
+    pSignal = &signal[signal_size - 1];
+  else
+    pSignal = &signal[0];
+
+  /* y(n) = x(n) - lpc[1]*y(n-1) - ... - lpc[order]*y(n-order) */
+
+  for (i = 0; i < signal_size; i++) {
+    FIXP_DBL x;
+    const FIXP_LPC *pCoeff = coeff + order - stateIndex;
+
+    x = scaleValue(*pSignal, -(lpcCoeff_e + 1));
+    for (j = 0; j < order; j++) {
+      x -= fMultDiv2(state[j], pCoeff[j]);
+    }
+    x = SATURATE_SHIFT(x, -lpcCoeff_e - 1, DFRACT_BITS);
+
+    /* Update states */
+    stateIndex = ((stateIndex - 1) < 0) ? (order - 1) : (stateIndex - 1);
+    state[stateIndex] = x;
+
+    *pSignal = scaleValue(x, signal_e);
+    pSignal += inc;
+  }
+
+  *pStateIndex = stateIndex;
+}
+
+/* FIR */
+void CLpc_Analysis(FIXP_DBL *RESTRICT signal, const int signal_size,
+                   const FIXP_LPC lpcCoeff_m[], const int lpcCoeff_e,
+                   const int order, FIXP_DBL *RESTRICT filtState,
+                   int *filtStateIndex) {
+  int stateIndex;
+  INT i, j, shift = lpcCoeff_e + 1; /* +1, because fMultDiv2 */
+  FIXP_DBL tmp;
+
+  if (order <= 0) {
+    return;
+  }
+  if (filtStateIndex != NULL) {
+    stateIndex = *filtStateIndex;
+  } else {
+    stateIndex = 0;
+  }
+
+  /* keep filter coefficients twice and save memory copy operation in
+     modulo state buffer */
+  FIXP_LPC coeff[2 * LPC_MAX_ORDER];
+  FIXP_LPC *pCoeff;
+  FDKmemcpy(&coeff[0], lpcCoeff_m, order * sizeof(FIXP_LPC));
+  FDKmemcpy(&coeff[order], lpcCoeff_m, order * sizeof(FIXP_LPC));
+
+  /*
+      # Analysis filter, obtain residual.
+      for k = 0:BL-1
+              err(i-BL+k) = a * inputSignal(i-BL+k:-1:i-BL-M+k);
+      endfor
+   */
+
+  FDK_ASSERT(shift >= 0);
+
+  for (j = 0; j < signal_size; j++) {
+    pCoeff = &coeff[(order - stateIndex)];
+
+    tmp = signal[j] >> shift;
+    for (i = 0; i < order; i++) {
+      tmp = fMultAddDiv2(tmp, pCoeff[i], filtState[i]);
+    }
+
+    stateIndex =
+        ((stateIndex - 1) < 0) ? (stateIndex - 1 + order) : (stateIndex - 1);
+    filtState[stateIndex] = signal[j];
+
+    signal[j] = tmp << shift;
+  }
+
+  if (filtStateIndex != NULL) {
+    *filtStateIndex = stateIndex;
+  }
+}
+
+/* For the LPC_SYNTHESIS_IIR version */
+INT CLpc_ParcorToLpc(const FIXP_LPC_TNS reflCoeff[], FIXP_LPC_TNS LpcCoeff[],
+                     INT numOfCoeff, FIXP_DBL workBuffer[]) {
+  INT i, j;
+  INT shiftval,
+      par2LpcShiftVal = 6; /* 6 should be enough, bec. max(numOfCoeff) = 20 */
+  FIXP_DBL maxVal = (FIXP_DBL)0;
+
+  workBuffer[0] = FX_LPC_TNS2FX_DBL(reflCoeff[0]) >> par2LpcShiftVal;
+  for (i = 1; i < numOfCoeff; i++) {
+    for (j = 0; j < i / 2; j++) {
+      FIXP_DBL tmp1, tmp2;
+
+      tmp1 = workBuffer[j];
+      tmp2 = workBuffer[i - 1 - j];
+      workBuffer[j] += fMult(reflCoeff[i], tmp2);
+      workBuffer[i - 1 - j] += fMult(reflCoeff[i], tmp1);
+    }
+    if (i & 1) {
+      workBuffer[j] += fMult(reflCoeff[i], workBuffer[j]);
+    }
+
+    workBuffer[i] = FX_LPC_TNS2FX_DBL(reflCoeff[i]) >> par2LpcShiftVal;
+  }
+
+  /* calculate exponent */
+  for (i = 0; i < numOfCoeff; i++) {
+    maxVal = fMax(maxVal, fAbs(workBuffer[i]));
+  }
+
+  shiftval = fMin(fNorm(maxVal), par2LpcShiftVal);
+
+  for (i = 0; i < numOfCoeff; i++) {
+    LpcCoeff[i] = FX_DBL2FX_LPC_TNS(workBuffer[i] << shiftval);
+  }
+
+  return (par2LpcShiftVal - shiftval);
+}
+/* Default version */
+INT CLpc_ParcorToLpc(const FIXP_LPC reflCoeff[], FIXP_LPC LpcCoeff[],
+                     INT numOfCoeff, FIXP_DBL workBuffer[]) {
+  INT i, j;
+  INT shiftval,
+      par2LpcShiftVal = 6; /* 6 should be enough, bec. max(numOfCoeff) = 20 */
+  FIXP_DBL maxVal = (FIXP_DBL)0;
+
+  workBuffer[0] = FX_LPC2FX_DBL(reflCoeff[0]) >> par2LpcShiftVal;
+  for (i = 1; i < numOfCoeff; i++) {
+    for (j = 0; j < i / 2; j++) {
+      FIXP_DBL tmp1, tmp2;
+
+      tmp1 = workBuffer[j];
+      tmp2 = workBuffer[i - 1 - j];
+      workBuffer[j] += fMult(reflCoeff[i], tmp2);
+      workBuffer[i - 1 - j] += fMult(reflCoeff[i], tmp1);
+    }
+    if (i & 1) {
+      workBuffer[j] += fMult(reflCoeff[i], workBuffer[j]);
+    }
+
+    workBuffer[i] = FX_LPC2FX_DBL(reflCoeff[i]) >> par2LpcShiftVal;
+  }
+
+  /* calculate exponent */
+  for (i = 0; i < numOfCoeff; i++) {
+    maxVal = fMax(maxVal, fAbs(workBuffer[i]));
+  }
+
+  shiftval = fMin(fNorm(maxVal), par2LpcShiftVal);
+
+  for (i = 0; i < numOfCoeff; i++) {
+    LpcCoeff[i] = FX_DBL2FX_LPC(workBuffer[i] << shiftval);
+  }
+
+  return (par2LpcShiftVal - shiftval);
+}
+
+void CLpc_AutoToParcor(FIXP_DBL acorr[], const int acorr_e,
+                       FIXP_LPC reflCoeff[], const int numOfCoeff,
+                       FIXP_DBL *pPredictionGain_m, INT *pPredictionGain_e) {
+  INT i, j, scale = 0;
+  FIXP_DBL parcorWorkBuffer[LPC_MAX_ORDER];
+
+  FIXP_DBL *workBuffer = parcorWorkBuffer;
+  FIXP_DBL autoCorr_0 = acorr[0];
+
+  FDKmemclear(reflCoeff, numOfCoeff * sizeof(FIXP_LPC));
+
+  if (autoCorr_0 == FL2FXCONST_DBL(0.0)) {
+    if (pPredictionGain_m != NULL) {
+      *pPredictionGain_m = FL2FXCONST_DBL(0.5f);
+      *pPredictionGain_e = 1;
+    }
+    return;
+  }
+
+  FDKmemcpy(workBuffer, acorr + 1, numOfCoeff * sizeof(FIXP_DBL));
+  for (i = 0; i < numOfCoeff; i++) {
+    LONG sign = ((LONG)workBuffer[0] >> (DFRACT_BITS - 1));
+    FIXP_DBL tmp = (FIXP_DBL)((LONG)workBuffer[0] ^ sign);
+
+    /* Check preconditions for division function: num<=denum             */
+    /* For 1st iteration acorr[0] cannot be 0, it is checked before loop */
+    /* Due to exor operation with "sign", num(=tmp) is greater/equal 0   */
+    if (acorr[0] < tmp) break;
+
+    /* tmp = div(num, denum, 16) */
+    tmp = (FIXP_DBL)((LONG)schur_div(tmp, acorr[0], FRACT_BITS) ^ (~sign));
+
+    reflCoeff[i] = FX_DBL2FX_LPC(tmp);
+
+    for (j = numOfCoeff - i - 1; j >= 0; j--) {
+      FIXP_DBL accu1 = fMult(tmp, acorr[j]);
+      FIXP_DBL accu2 = fMult(tmp, workBuffer[j]);
+      workBuffer[j] += accu1;
+      acorr[j] += accu2;
+    }
+    /* Check preconditions for division function: denum (=acorr[0]) > 0 */
+    if (acorr[0] == (FIXP_DBL)0) break;
+
+    workBuffer++;
+  }
+
+  if (pPredictionGain_m != NULL) {
+    if (acorr[0] > (FIXP_DBL)0) {
+      /* prediction gain = signal power / error (residual) power */
+      *pPredictionGain_m = fDivNormSigned(autoCorr_0, acorr[0], &scale);
+      *pPredictionGain_e = scale;
+    } else {
+      *pPredictionGain_m = (FIXP_DBL)0;
+      *pPredictionGain_e = 0;
+    }
+  }
+}