1
0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2024-12-12 19:18:44 +02:00
Commit Graph

54 Commits

Author SHA1 Message Date
Andreas Rheinhardt
790f793844 avutil/common: Don't auto-include mem.h
There are lots of files that don't need it: The number of object
files that actually need it went down from 2011 to 884 here.

Keep it for external users in order to not cause breakages.

Also improve the other headers a bit while just at it.

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
2024-03-31 00:08:43 +01:00
Michael Niedermayer
907743239d
avutil/tx_template: fix integer ovberflwo in fft3()
Fixes: signed integer overflow: -1028966111 + -1314089526 cannot be represented in type 'int'
Fixes: 63174/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_AAC_FIXED_fuzzer-5853273711837184

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Reviewed-by: Lynne <dev@lynne.ee>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2023-10-27 18:10:47 +02:00
Michael Niedermayer
c42a89309a
avutil/tx_template: Fix some signed integer overflows in DECL_FFT5()
Fixes: signed integer overflow: -1364715454 + -1468954671 cannot be represented in type 'int'
Fixes: 62093/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_AAC_FIXED_fuzzer-5538774254485504

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2023-10-03 20:09:27 +02:00
Lynne
d40672e661
lavu/tx: fix scaling of R2R transforms
Still slightly inaccurate, but it's good enough now.
2023-09-01 23:59:13 +02:00
Lynne
59b39d241e
lavu/tx: improve rdft table generation precision slightly 2023-09-01 23:59:13 +02:00
Lynne
ef8fd7bc3c
lavu/tx: add DCT-I and DST-I transforms
These are true, actual DCT-I and DST-I transforms, unlike the
libavcodec versions, which are plainly not.
2023-09-01 23:59:10 +02:00
Lynne
11e22730e1
lavu/tx: add real to real and real to imaginary RDFT transforms
These are in-place transforms, required for DCT-I and DST-I.

Templated as the mod2 variant requires minor modifications, and is
required specifically for DCT-I/DST-I.
2023-09-01 23:59:08 +02:00
Michael Niedermayer
8f48a62151
avutil/tx_template: extend to 2M
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2023-07-22 16:18:39 +02:00
Michael Niedermayer
9f04055669
avutil/tx_template: Use more unsigned ints to avoid undefined overflows
Fixes: signed integer overflow: 574590586 - -1875616554 cannot be represented in type 'int'
Fixes: 53914/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_AAC_FIXED_fuzzer-5037125846564864

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2022-12-25 19:33:24 +01:00
Lynne
710d83bdde
lavu/tx: zero-out imaginary of last coefficient in forward RDFTs
We didn't do this, because it's zero anyway, but it prevents users from using
uninitialized memory in calculations.
2022-12-03 21:02:00 +01:00
Michael Niedermayer
7792825ad6
avutil/tx: Use unsigned in ff_tx_fft_sr_combine() to avoid undefined behavior
Fixes: signed integer overflow: -1284837070 - 982101618 cannot be represented in type 'int'
Fixes: 53105/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_AC3_FIXED_fuzzer-4848015827664896

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2022-11-28 20:58:05 +01:00
Lynne
a56d7e0ca3
lavu/tx: add DCT-III implementation 2022-11-24 15:58:36 +01:00
Lynne
504b7bec1a
lavu/tx: add DCT-II implementation 2022-11-24 15:58:35 +01:00
Lynne
43d285a40f
lavu/tx: fix last coefficient scaling for R2C transforms
This was a typo.
2022-11-24 15:58:35 +01:00
Lynne
8547123f3b
lavu/tx: generalize PFA FFTs
This commit permits any stacking of FFTs of any size.
2022-11-24 15:58:34 +01:00
Lynne
87bae6b018
lavu/tx: refactor to explicitly track and convert lookup table order
Necessary for generalizing PFAs.
2022-11-24 15:58:34 +01:00
Lynne
6ddd10c3e2
lavu/tx: allow codelets to specify a minimum number of matching factors 2022-11-24 15:58:33 +01:00
Lynne
fbe4fd992f
lavu/tx: support output stride in naive transforms
Allows them to be used in general PFAs.
2022-11-24 15:58:31 +01:00
Lynne
68cabf8750
lavu/tx: add fft_inplace_small transforms
This is much faster than the loop.
2022-11-24 15:58:30 +01:00
Lynne
fff3e1d848
lavu/tx: support out-of-place transforms in fft_inplace
This makes testing easier, as a unified path can be used for in/out of
place transforms.
2022-11-24 15:58:30 +01:00
Lynne
d260796f11
lavu/tx: make C ptwo transforms in+out of place
We assume that _all_ in-place transforms can operate out of place,
which isn't true, because the C ptwo transforms were always in-place (dst).
2022-11-24 15:58:29 +01:00
Lynne
37008dc402
lavu/tx: add naive_small FFT
The same as naive but with precomputed tables. Makes it more useful
for odd-factors we don't support yet.
2022-11-24 15:58:29 +01:00
Lynne
e8a9b7b298
lavu/tx: list all odd-length FFT factors as regular codelets
Allows them to be picked just like any other transform.
2022-11-24 15:58:28 +01:00
Lynne
45bd4bf79f
lavu/tx: generalize single-factor transforms
Not that useful, but it gives us fast small odd-length transforms.
2022-11-24 15:58:28 +01:00
Lynne
79f11e2409
lavu/tx: make prime factor transforms truly in-place
They all overwrote in[0] and then used it as a DC.
2022-11-24 15:58:28 +01:00
Andreas Rheinhardt
f8efd890bf avutil/tx_template: Move function pointers to const memory
This can be achieved by moving the AVOnce out of the structure
containing the function pointers; the latter can then be made
const.
This also has the advantage of eliminating padding in the structure
(sizeof(AVOnce) is four here) and allowing the AVOnces to be put
into .bss (dependening upon the implementation).

Reviewed-by: Lynne <dev@lynne.ee>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
2022-10-28 09:30:10 +02:00
Andreas Rheinhardt
188216581b avutil/tx_template: Avoid code duplication
Reviewed-by: Lynne <dev@lynne.ee>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
2022-10-28 09:30:10 +02:00
Andreas Rheinhardt
2af5f55b2e avutil/tx_template: Don't waste space for inexistent factors
It is possible to avoid the factors array for the power-of-two
tables for which said array is unused by using a different
structure for initialization for power-of-two tables than for
non-power-of-two-tables. This saves 3*15*16B from .data.

Reviewed-by: Lynne <dev@lynne.ee>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
2022-10-28 09:29:41 +02:00
Lynne
ace42cf581
x86/tx_float: add 15xN PFA FFT AVX SIMD
~4x faster than the C version.
The shuffles in the 15pt dim1 are seriously expensive. Not happy with it,
but I'm contempt.

Can be easily converted to pure AVX by removing all vpermpd/vpermps
instructions.
2022-09-23 12:35:27 +02:00
Lynne
7e7baf8ab8
lavu/tx: do not steal lookup tables of subcontexts in the iMDCT
As it happens, some still need their contexts.
2022-09-23 12:33:31 +02:00
Lynne
f1b35fc8f0
lavu/tx: remove av_cold from table definitions
How did this get here?
2022-09-11 03:18:40 +02:00
Lynne
c92edd969a
lavu/tx: rotate 3 & 15-point exptabs
This just inverts their signs. Simplifies SIMD.
2022-09-10 02:37:17 +02:00
Lynne
51172223fd
lavu/tx: generalize MDCTs
The same code can perform any-length MDCTs with minimal changes.
2022-09-10 02:37:16 +02:00
Lynne
645a1f4422
lavu/tx: add the inplace flag to PFA FFTs
They support in-place, because they have to use a temporary buffer.
2022-09-10 02:37:14 +02:00
Lynne
ae66a9db7b
lavu/tx: optimize and simplify inverse MDCTs
Convert the input from a scatter to a gather instead,
which is faster and better for SIMD.
Also, add a pre-shuffled exptab version to avoid
gathering there at all. This doubles the exptab size,
but the speedup makes it worth it. In SIMD, the
exptab will likely be purged to a higher cache
anyway because of the FFT in the middle, and
the amount of loads stays identical.

For a 960-point inverse MDCT, the speedup is 10%.

This makes it possible to write sane and fast SIMD
versions of inverse MDCTs.
2022-08-16 01:22:38 +02:00
Lynne
af94ab7c7c
lavu/tx: add an RDFT implementation
RDFTs are full of conventions that vary between implementations.
What I've gone for here is what's most common between
both fftw, avcodec's rdft and what we use, the equivalent of
which is DFT_R2C for forward and IDFT_C2R for inverse. The
other 2 conventions (IDFT_R2C and DFT_C2R) were not used at
all in our code, and their names are also not appropriate.
If there's a use for either, we can easily add a flag which
would just flip the sign on one exptab.

For some unknown reason, possibly to allow reusing FFT's exp tables,
av_rdft's C2R output is 0.5x lower than what it should be to ensure
a proper back-and-forth conversion.
This code outputs its real samples at the correct level, which
matches FFTW's level, and allows the user to change the level
and insert arbitrary multiplies for free by setting the scale option.
2022-01-26 04:12:46 +01:00
Lynne
ef4bd81615
lavu/tx: rewrite internal code as a tree-based codelet constructor
This commit rewrites the internal transform code into a constructor
that stitches transforms (codelets).
This allows for transforms to reuse arbitrary parts of other
transforms, and allows transforms to be stacked onto one
another (such as a full iMDCT using a half-iMDCT which in turn
uses an FFT). It also permits for each step to be individually
replaced by assembly or a custom implementation (such as an ASIC).
2022-01-26 04:12:44 +01:00
Lynne
1978b143eb
checkasm: add av_tx FFT SIMD testing code
This sadly required making changes to the code itself,
due to the same context needing to be reused for both versions.
The lookup table had to be duplicated for both versions.
2021-04-24 17:19:17 +02:00
Lynne
0072a42388
lavu/tx: add full-sized iMDCT transform flag 2021-04-24 17:17:27 +02:00
Lynne
8c55c82583
lavu/tx: add a 9-point FFT and (i)MDCT 2021-04-24 17:17:25 +02:00
Lynne
bd9ea917a3
lavu/tx: add a 7-point FFT and (i)MDCT 2021-04-24 17:17:23 +02:00
Lynne
89da62f2fc
lavu/tx: refactor power-of-two FFT
This commit refactors the power-of-two FFT, making it faster and
halving the size of all tables, making the code much smaller on
all systems.
This removes the big/small pass split, because on modern systems
the "big" pass is always faster, and even on older machines there
is no measurable speed difference.
2021-04-24 17:17:20 +02:00
Lynne
e20a39a375
lavu/tx: do not invert permutes on MDCTs 2021-02-27 05:01:17 +01:00
Lynne
8e94b7cff0
lavu/tx: invert permutation lookups
out[lut[i]] = in[i] lookups were 4.04 times(!) slower than
out[i] = in[lut[i]] lookups for an out-of-place FFT of length 4096.

The permutes remain unchanged for anything but out-of-place monolithic
FFT, as those benefit quite a lot from the current order (it means
there's only 1 lookup necessary to add to an offset, rather than
a full gather).

The code was based around non-power-of-two FFTs, so this wasn't
benchmarked early on.
2021-02-27 04:21:05 +01:00
Lynne
10341743d2
lavu/tx: require output argument to match input for inplace transforms
This simplifies some assembly code by a lot, by either saving a branch
or saving an entire duplicated function.
2021-02-26 05:42:24 +01:00
Lynne
5ca40d6d94
lavu/tx: support in-place FFT transforms
This commit adds support for in-place FFT transforms. Since our
internal transforms were all in-place anyway, this only changes
the permutation on the input.

Unfortunately, research papers were of no help here. All focused
on dry hardware implementations, where permutes are free, or on
software implementations where binary bloat is of no concern so
storing dozen times the transforms for each permutation and version
is not considered bad practice.
Still, for a pure C implementation, it's only around 28% slower
than the multi-megabyte FFTW3 in unaligned mode.

Unlike a closed permutation like with PFA, split-radix FFT bit-reversals
contain multiple NOPs, multiple simple swaps, and a few chained swaps,
so regular single-loop single-state permute loops were not possible.
Instead, we filter out parts of the input indices which are redundant.
This allows for a single branch, and with some clever AVX512 asm,
could possibly be SIMD'd without refactoring.

The inplace_idx array is guaranteed to never be larger than the
revtab array, and in practice only requires around log2(len) entries.

The power-of-two MDCTs can be done in-place as well. And it's
possible to eliminate a copy in the compound MDCTs too, however
it'll be slower than doing them out of place, and we'd need to dirty
the input array.
2021-02-21 17:05:16 +01:00
James Almer
f6477ac9f4 avutil/tx: use ENOSYS instead of ENOTSUP
It's the standard error code used across the codebase to signal unimplemented
or unsupported features.

Signed-off-by: James Almer <jamrial@gmail.com>
2021-01-13 23:02:47 -03:00
Lynne
06a8596825
lavu: support arbitrary-point FFTs and all even (i)MDCT transforms
This patch adds support for arbitrary-point FFTs and all even MDCT
transforms.
Odd MDCTs are not supported yet as they're based on the DCT-II and DCT-III
and they're very niche.

With this we can now write tests.
2021-01-13 17:34:13 +01:00
Lynne
2465fe1302 lavu/tx: add 2-point FFT transform
By itself, this allows 6-point, 10-point and 30-point transforms.
When the 9-point transform is added it allows for 18-point FFT,
and also for a 36-point MDCT (used by MP3).
2020-03-23 21:26:25 +00:00
Lynne
e1c84856bb lavu/tx: improve 3-point fixed precision
There's just no reason not to when its so easy (albeit messy) and its also
reducing the precision of all non-power-of-two transforms that use it.
2020-02-14 19:58:14 +00:00