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.
This fixes the following warning:
libSBRenc/src/sbr_encoder.cpp:1942:14: warning: equality comparison
with extraneous parentheses [-Wparentheses-equality]
Change-Id: I2c833d19348574473c752076a4efff6642c0d653
GCC 6 defaults to C++14 (or more precisely, gnu++14) instead of C++98
(or gnu++98).
The fdk-aac source doesn't support being built in this mode at the
moment, since it relies on narrowing conversion from unsigned 32 bit
integers to FIXP_DBL (which is a signed data type of the same size).
The same approach is used upstream in Android as well, since
d52f374768.
This fixes buliding with GCC 6.
am: 2a923c9
* commit '2a923c92869efb277ca7d67ab6c27617ec26c128':
Disable warnings that are not easily fixable
Change-Id: I920b16bdaa1ce7a00c91e129b950f0d486a1fa37
am: 2b2e97b
* commit '2b2e97b893b16b69c7728d16e8653b9c94944294':
Left shift of negative int is undefined
Change-Id: Ie90b850b7adf787dfb47f8b7c43a22744c3a7fa0
am: 8192b1a
* commit '8192b1a98f60cc39253c658ba1ca70a70c2e0dfa':
Disable warnings that are not easily fixable
Change-Id: I5dc1fabd29cfb7024f8362359a796efd1764032d
am: 6ae5480
* commit '6ae54806ab88efd76e884447680fd30b340936fd':
Left shift of negative int is undefined
Change-Id: If6a4abe686d3ab6b7f825058255b03aa46055484
Parse DVB DRC data only when numThreads is below
MAX_DRC_THREADS. The post-increment is necessary as
it is used in fill element DRC data section.
This solution parses as many DRC payloads as allowed by
MAX_DRC_THREADS and skips all remaining DRC elements in the stream.
Bug 27792766
Bug 26751339
Change-Id: Ie1641888bac1757c4d1491119f977fc5d436eaea
* AAC-Decoder
- Add support for AOT 20 (ER-AAC scalable) (base layer only)
- Add support for AAC as used in Digital Radio Mondiale (DRM30/DRM+)
Modified file(s):
libAACdec/src/aacdecoder.cpp
libAACdec/src/aacdecoder_lib.cpp
libFDK/src/FDK_core.cpp
libFDK/src/FDK_tools_rom.cpp
libMpegTPDec/src/tpdec_asc.cpp
libMpegTPDec/src/tpdec_lib.cpp
libMpegTPDec/src/version
libSBRdec/include/sbrdecoder.h
libSBRdec/src/env_extr.h
libSBRdec/src/sbrdecoder.cpp
Added file(s):
libMpegTPDec/src/tpdec_drm.cpp
libMpegTPDec/src/tpdec_drm.h
- Fix sanity check in HCR module that was performed at the wrong point in time.
Modified file(s):
libAACdec/src/aacdecoder_lib.cpp
libAACdec/src/block.cpp
- Extend core sampling rate support up to 96 kHz.
Modified file(s):
libAACdec/src/aac_rom.cpp
libAACdec/src/aacdecoder.cpp
libAACdec/src/aacdecoder_lib.cpp
- Return correct audio output channel description according number of output
channels.
Modified file(s):
libAACdec/src/aacdecoder_lib.cpp
- Indroduce decoder intern output buffer. This change allows to use framework
output buffer with the actual size of the deocder output channels.
Modified file(s):
libAACdec/include/aacdecoder_lib.h
libAACdec/src/aacdecoder.h
libAACdec/src/aacdecoder_lib.cpp
* SBR-Decoder
- Increase robustness for erroneous input data.
- Improve error concealment performance.
- Fix handling of lowest sub-band for LD-SBR
Modified file(s):
libAACdec/src/aacdecoder.cpp
libAACdec/src/aacdecoder_lib.cpp
libSBRdec/src/env_calc.cpp
libSBRdec/src/env_dec.cpp
libSBRdec/src/env_extr.cpp
libSBRdec/src/env_extr.h
libSBRdec/src/sbr_dec.cpp
libSBRdec/src/sbr_rom.cpp
libSBRdec/src/sbr_rom.h
libSBRdec/src/sbrdecoder.cpp
- Add QMF delay compensation for ELD v2 streams decoded with the complex
low delay filter-bank.
Modified file(s):
libSBRdec/src/sbr_dec.cpp
libSBRdec/src/sbr_dec.h
libSBRdec/src/sbrdecoder.cpp
- Introduce a different handling of frames to be flushed
dependent on whether there are delayed frames available or not.
Modified file(s):
libSBRdec/src/sbr_ram.h
libSBRdec/src/sbrdecoder.cpp
- Calculate the correct number of samples for dual-mono copy in case of no
available PS data.
Modified file(s):
libSBRdec/src/sbrdecoder.cpp
* SYS-Library
- Change include order of genericStds.h to prevent conflict with definitions
which are also used in math.h.
Modified file(s):
libSYS/src/genericStds.cpp
Change-Id: I3ecffbad85f39b056213107955cfadbeb3f4b6e1
