This patch replaces the transform used in AAC with lavu/tx and removes
the limitation on only being able to decode 960-sample files
with the float decoder.
This commit also removes a whole bunch of unnecessary and slow
lifting steps the decoder did to compensate for the poor accuracy
of the old integer transformation code.
Overall float decoder speedup on Zen 3 for 64kbps: 32%
Since the removal of the 16-bit FFT said define is unnecessary as
FFT_FIXED_32 is always !FFT_FLOAT. But one wouldn't believe it when
looking at the code.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
The fixed-point AAC decoder is the only user of the fixed-point sinewin
tables from sinewin; and it only uses a few of them (about 10% when
counting by size). This means that guarding initializing these tables by
an AVOnce (as done in 3719122065) is
unnecessary for them. Furthermore the array of pointers to the
individual arrays is also unneeded.
Therefore this commit moves these tables directly into aacdec_fixed.c;
this is done by ridding the original sinewin.h and sinewin_tablegen.h
headers completely of any fixed-point code at the cost of a bit of
duplicated code (the alternative is an ugly ifdef-mess).
This saves about 58KB from the binary when using hardcoded tables (as
these tables are hardcoded in this scenario); when not using hardcoded
tables, most of these savings only affect the .bss segment, but the rest
(< 1KB) contains relocations (i.e. savings in .data.rel.ro).
Reviewed-by: Lynne <dev@lynne.ee>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
The floating point kbd tables for 120 and 960 samples are only used by
the floating point decoder whereas the fixed point kbd tables for 128
and 1024 samples are only used by the fixed point AAC decoder. So move
these tables to their only users. This ensures that they are not
accidentally used somewhere else without ensuring that initializing
these tables stays thread-safe (as it is now because the only place from
where they are initialized is guarded by an AVOnce).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Fixes: runtime error: signed integer overflow: -2147483648 - 1202286525 cannot be represented in type 'int'
Fixes: 2071/clusterfuzz-testcase-minimized-6036414271586304
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Fixes: runtime error: shift exponent 1073741848 is too large for 32-bit type 'INTFLOAT' (aka 'int')
Fixes: 1880/clusterfuzz-testcase-minimized-4900645322620928
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
This may improve the precision of the fixed point encoder/decoder for some
compilers and architectures.
Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Marton Balint <cus@passwd.hu>
Andreas Rheinhardt
Lynne
Andreas Rheinhardt
Michael Niedermayer
Marton Balint
Ganesh Ajjanagadde
Djordje Pesut