Was accidently disabled in 14c6bb798d
(in January 2019)
(Had to make a minor change due argument differences)
Change-Id: If7c128cdeaa9ed82b2b33de1b75ca7cc4a95abdd
* Only codecs and plugins are enabled
* Only native builds (so far)
* Only tested on xDuoo X3 (MIPS, monochrome)
* opus & speex generate some warnings
* Significant compile time impact
Change-Id: I519b0d179631a54b2103cd67225bd5ec6ad3bd2f
Annoyingly, this makes all of the '.S' files we compile get treated as
divided syntax, so we need to make the syntax in them explicit.
Change-Id: I56a3916b7b24c84a1214a5d6bc4ed4d651f002cf
...Both appear to be false positives, and will hopefully be resolved
in a later GCC point release. I will periodically revisit this.
Change-Id: I7388932f5be052d39dd8c4195b438bf275dd72d9
Allocates several SBR + PS arrays and variables statically
Prevents out of memory exceptions for long audiobooks
Note from original commit a602f46d why it was disabled:
For now malloc is not fully removed but used by a few arrays needed for AAC-HE SBR+PS only. Reason to keep malloc is to have this amount of memory available for AAC-LC files which might require large m4a tables.
But now with "shrinkable" seek table 2c6dfd06 it's no longer a critical issue for aac-lc files (we just load less seek entries) but might be critical for aac-he files.
Fixes issue described here https://www.rockbox.org/tracker/task/13049#comment44587
Change-Id: Id9c20bad4c46034299a2a4de95d41c807b3af412
The supplied time may be more accurate than the calculated time from the bitrate (can be quite inaccurate for long vbr files and even for cbr due to padding)
Change-Id: I1f8291eab45fbd91a161b51776c391a0f7886a5c
Captured ADTS streams might start from some garbage data from previous frame. We should check for sync errors even for first frame.
Change-Id: I70171298b79713aeedf9fa2e6098a03063487649
first_frame_offset is not set for mp4 files so seeking to it doesn't make any sense. It can lead only to additional re-buffer request. So instead let's just allow seek back for m4a_check_sample_offset check. And do seek_buffer(0) before reading metadata (fixes possible hang introduced in fc65bdab)
Change-Id: Ia7fae14b0137d73b5e263390be5f143deb7ca789
Includes ipod video (5G) and earlier models, sansa c200 and others players not capable to decode AAC-HE.
Allows to play backward compatible files as AAC-LC.
Change-Id: Ic9f5c0f255d9a4308c3414d402f8f27f4328ca94
In files with gaps between chunks and reduced lookup_table we can't properly detect all gaps in m4a_check_sample_offset. So just ignore decode errors till next chunk present in lookup_table
Change-Id: I317864dce6a2251cdb6ddb8c0ad4d7c1640cb7a1
Instead of FOURCC it needs ignore-case text match. Also value can contain \0 ('m4a\0' instead of expected 'm4a ').But let's simply skip it and let decoder handle it.
Change-Id: I87eefcabbc9010481286257c26cee09e61d1221c
Codecs mostly use custom LOGF define for logging (i.e. see aac.c). Now such logging can be enabled in single file with #define LOGF_ENABLE
Change-Id: I36312fbcd2d9166fb1fe5ead31e7354342d8828d
Make optimization from 2358fabb actually work.
Fix potential out of bound access.
Remove redundant zero offset check.
Change-Id: I0a0ba04670b612d410ac17a761bd08c2915721b9
Don't store sample_to_chunk table and read data on demand instead (it's required only once for building lookup table). It allows to store 2x bigger lookup table.
Change-Id: Ida79d0c281040300d6561e124fe10ebacb0e4679
pretty unlikely but I think in a memory exaustion scenario
you might roll the uint8 over to zero and crash
Change-Id: If73dc64484f2d363446be068aa7bf8025407a2a7
GCC 4.9 always emits assembly with divided syntax. Setting unified
syntax in inline assembly causes the assembler to complain about
GCC's generated code, because the directive extends past the scope
of the inline asm. Fix this by setting divided mode at the end of
the inline assembly block.
The assembler directives are hidden behind macros because later
versions of GCC won't need this workaround: they can be told to
use the unified syntax with -masm-syntax-unified.
Change-Id: Ic09e729e5bbb6fd44d08dac348daf6f55c75d7d8
Replace the minimum version bound with a check on the size of
the API struct. The version only needs to be incremented for
ABI breaking changes. Additions to the API won't need to touch
the version number, resulting in fewer merge conflicts.
Change-Id: I916a04a7bf5890dcf5d615ce30087643165f8e1f
Plain import of the library parts first. Adaptions to Rockbox will
follow. A *lot* of kudos go to Mihaly Horvath for creating this library
from his already lightweight cSID-light, mainly for Rockbox. Besides a
lot of other things, he made his algorithms integer-only and
significantly improved the C64 emulation, so finally RSIDs could be
played as well as PSIDs. TinySID was nice for what it is, but this is a
quantum leap in SID playback quality for Rockbox. Check for example:
https://hvsc.csdb.dk/MUSICIANS/P/Page_Jason/Eighth.sidhttps://hvsc.csdb.dk/MUSICIANS/J/Jeff/Blowing.sid
Change-Id: I353e12fbfd7cd8696b834616e55743e7b844a73e
It turns out removing DSP_INIT broke the codec ABI and caused
old codecs to crash; the loop and mdelay() was a red herring.
This reverts commit 541960a110.
Change-Id: I020d826e7b4beb006d093d9c3d4f45fa5eaac717
Commit 6bcd830490 ported an optimization to decode_subframe_fixed()
from FFmpeg (upstream commit 08965b22e2). This contains an out of
bounds read, which doesn't affect the decoder output, but makes ASAN
complain.
FFmpeg fixed the out of bounds read (upstream commit 0ec7b71de8) but
that appears to increase code size a lot.
Inlining the initialization of a, b, c, d into the switch produces
similar code as the non-bounds-checked version with only a handful
of instructions of overhead (checked on MIPS & ARM).
Change-Id: I053fac4efc4676b133eb7545c80e23f37fb00d86
Resume by offset was obviously inaccurate for ALAC -- it tried
to convert the offset to an elapsed time using the approximate
bitrate, which is going to be wrong for VBR files. This became
a problem since commit 26ffcd8f9f restored the ability to resume
by offset.
It turns out that m4a_seek_raw() has terrible resolution since
it can only seek to chunk boundaries, and lies about the real
sample position; basically the same issue that affected seeking
described in commit 4dd3c2b33e. Resuming by offset is still not
very accurate because of this. Prefer to resume by time first,
which is normally highly accurate (and never worse than offset)
but use the file offset if it's the only thing we have.
There were a couple time calculations still using 32-bit math,
so clean those up too to reduce issues due to rounding errors.
Change-Id: Idd3bccd67505f4e59e784d92e45ea80a273975bb
