Anton Khirnov [Mon, 7 Nov 2016 13:05:49 +0000 (14:05 +0100)]
hwcontext_dxva2: split transfer_data() into upload/download functions
Just the presence of a hw frames context is not enough to detect whether
the transfer is an upload or a download, because hw frames mapped to
system memory will have a hw frames context attached.
Anton Khirnov [Mon, 7 Nov 2016 12:42:50 +0000 (13:42 +0100)]
hwcontext_dxva2: fix handling of the mapping flags
D3DLOCK_READONLY properly corresponds to the absence of the write flag,
not to the presence of the read flag, while D3DLOCK_DISCARD is
equivalent to the overwrite flag.
A number of new pix_fmts* have been added to AviSynth+:
16-bit packed RGB and RGBA
10-, 12-, 14, and 16-bit YUV 4:2:0, 4:2:2, and 4:4:4
8-, 10-, 12-, 14-, and 16-bit Planar RGB
8-, 10-, 12-, 14-, and 16-bit Planar YUVA and Planar RGBA
10-, 12-, 14-, and 16-bit GRAY variants
32-bit floating point Planar YUV(A), Planar RGB(A), and GRAY
*some of which are not currently available pix_fmts here and were
not added to the demuxer due to this
Vittorio Giovara [Tue, 15 Nov 2016 15:14:30 +0000 (10:14 -0500)]
mov: Fix identity matrix boolean logic
This prevented the code from correctly exporting the rotation matrix
which caused a few samples to be displayed wrong.
Introduced in ecd2ec69ce10e13f6ede353d2def7c.
Martin Storsjö [Mon, 14 Nov 2016 21:44:06 +0000 (23:44 +0200)]
checkasm: vp9dsp: Benchmark the dc-only version of idct_idct separately
The dc-only mode is already checked to work correctly above, but this
allows benchmarking this mode for performance tuning, and allows making
sure that it actually is correctly hooked up.
Janne Grunau [Mon, 14 Nov 2016 21:35:13 +0000 (22:35 +0100)]
aarch64: vp9: loop filter: replace 'orr; cbn?z' with 'adds; b.{eq,ne};
The latter is 1 cycle faster on a cortex-53 and since the operands are
bytewise (or larger) bitmask (impossible to overflow to zero) both are
equivalent.
Janne Grunau [Sun, 13 Nov 2016 23:13:34 +0000 (00:13 +0100)]
aarch64: vp9: use alternative returns in the core loop filter function
Since aarch64 has enough free general purpose registers use them to
branch to the appropiate storage code. 1-2 cycles faster for the
functions using loop_filter 8/16, ... on a cortex-a53. Mixed results
(up to 2 cycles faster/slower) on a cortex-a57.
Gianluigi Tiesi [Fri, 19 Sep 2014 02:49:36 +0000 (04:49 +0200)]
libilbc: support for latest git of libilbc
In the latest git commits of libilbc developers removed WebRtc_xxx typedefs.
This commit uses int types instead. It's safe to apply also for previous
versions since WebRtc_Word16 was always a typedef of int16_t and
WebRtc_UWord16 a typedef of uint16_t.
Reviewed-by: Timothy Gu <timothygu99@gmail.com> Signed-off-by: Diego Biurrun <diego@biurrun.de>
Diego Biurrun [Mon, 4 Jan 2016 10:14:51 +0000 (11:14 +0100)]
af_asyncts: Use llabs instead of labs for 64-bit variable
libavfilter/af_asyncts.c:212:9: warning: absolute value function 'labs' given an argument of type 'int64_t' (aka 'long long') but has parameter of type 'long' which may cause truncation of value [-Wabsolute-value]
Mark Thompson [Wed, 26 Oct 2016 20:26:10 +0000 (21:26 +0100)]
qsvdec: Pass the correct profile to libmfx
This was correct for H.26[45], because libmfx uses the same values
derived from profile_idc and the constraint_set flags, but it is
wrong for other codecs.
Also avoid passing FF_LEVEL_UNKNOWN (-99) as the level, as this is
certainly invalid.
Martin Storsjö [Sun, 13 Nov 2016 22:07:27 +0000 (00:07 +0200)]
aarch64: vp9: Implement NEON loop filters
This work is sponsored by, and copyright, Google.
These are ported from the ARM version; thanks to the larger
amount of registers available, we can do the loop filters with
16 pixels at a time. The implementation is fully templated, with
a single macro which can generate versions for both 8 and
16 pixels wide, for both 4, 8 and 16 pixels loop filters
(and the 4/8 mixed versions as well).
For the 8 pixel wide versions, it is pretty close in speed (the
v_4_8 and v_8_8 filters are the best examples of this; the h_4_8
and h_8_8 filters seem to get some gain in the load/transpose/store
part). For the 16 pixels wide ones, we get a speedup of around
1.2-1.4x compared to the 32 bit version.
Examples of runtimes vs the 32 bit version, on a Cortex A53:
ARM AArch64
vp9_loop_filter_h_4_8_neon: 144.0 127.2
vp9_loop_filter_h_8_8_neon: 207.0 182.5
vp9_loop_filter_h_16_8_neon: 415.0 328.7
vp9_loop_filter_h_16_16_neon: 672.0 558.6
vp9_loop_filter_mix2_h_44_16_neon: 302.0 203.5
vp9_loop_filter_mix2_h_48_16_neon: 365.0 305.2
vp9_loop_filter_mix2_h_84_16_neon: 365.0 305.2
vp9_loop_filter_mix2_h_88_16_neon: 376.0 305.2
vp9_loop_filter_mix2_v_44_16_neon: 193.2 128.2
vp9_loop_filter_mix2_v_48_16_neon: 246.7 218.4
vp9_loop_filter_mix2_v_84_16_neon: 248.0 218.5
vp9_loop_filter_mix2_v_88_16_neon: 302.0 218.2
vp9_loop_filter_v_4_8_neon: 89.0 88.7
vp9_loop_filter_v_8_8_neon: 141.0 137.7
vp9_loop_filter_v_16_8_neon: 295.0 272.7
vp9_loop_filter_v_16_16_neon: 546.0 453.7
The speedup vs C code in checkasm tests is around 2-7x, which is
pretty much the same as for the 32 bit version. Even if these functions
are faster than their 32 bit equivalent, the C version that we compare
to also became around 1.3-1.7x faster than the C version in 32 bit.
Based on START_TIMER/STOP_TIMER wrapping around a few individual
functions, the speedup vs C code is around 4-5x.
Examples of runtimes vs C on a Cortex A57 (for a slightly older version
of the patch):
A57 gcc-5.3 neon
loop_filter_h_4_8_neon: 256.6 93.4
loop_filter_h_8_8_neon: 307.3 139.1
loop_filter_h_16_8_neon: 340.1 254.1
loop_filter_h_16_16_neon: 827.0 407.9
loop_filter_mix2_h_44_16_neon: 524.5 155.4
loop_filter_mix2_h_48_16_neon: 644.5 173.3
loop_filter_mix2_h_84_16_neon: 630.5 222.0
loop_filter_mix2_h_88_16_neon: 697.3 222.0
loop_filter_mix2_v_44_16_neon: 598.5 100.6
loop_filter_mix2_v_48_16_neon: 651.5 127.0
loop_filter_mix2_v_84_16_neon: 591.5 167.1
loop_filter_mix2_v_88_16_neon: 855.1 166.7
loop_filter_v_4_8_neon: 271.7 65.3
loop_filter_v_8_8_neon: 312.5 106.9
loop_filter_v_16_8_neon: 473.3 206.5
loop_filter_v_16_16_neon: 976.1 327.8
The speed-up compared to the C functions is 2.5 to 6 and the cortex-a57
is again 30-50% faster than the cortex-a53.
Martin Storsjö [Sun, 13 Nov 2016 21:53:08 +0000 (23:53 +0200)]
aarch64: vp9: Add NEON itxfm routines
This work is sponsored by, and copyright, Google.
These are ported from the ARM version; thanks to the larger
amount of registers available, we can do the 16x16 and 32x32
transforms in slices 8 pixels wide instead of 4. This gives
a speedup of around 1.4x compared to the 32 bit version.
The fact that aarch64 doesn't have the same d/q register
aliasing makes some of the macros quite a bit simpler as well.
Examples of runtimes vs the 32 bit version, on a Cortex A53:
ARM AArch64
vp9_inv_adst_adst_4x4_add_neon: 90.0 87.7
vp9_inv_adst_adst_8x8_add_neon: 400.0 354.7
vp9_inv_adst_adst_16x16_add_neon: 2526.5 1827.2
vp9_inv_dct_dct_4x4_add_neon: 74.0 72.7
vp9_inv_dct_dct_8x8_add_neon: 271.0 256.7
vp9_inv_dct_dct_16x16_add_neon: 1960.7 1372.7
vp9_inv_dct_dct_32x32_add_neon: 11988.9 8088.3
vp9_inv_wht_wht_4x4_add_neon: 63.0 57.7
The speedup vs C code (2-4x) is smaller than in the 32 bit case,
mostly because the C code ends up significantly faster (around
1.6x faster, with GCC 5.4) when built for aarch64.
Examples of runtimes vs C on a Cortex A57 (for a slightly older version
of the patch):
A57 gcc-5.3 neon
vp9_inv_adst_adst_4x4_add_neon: 152.2 60.0
vp9_inv_adst_adst_8x8_add_neon: 948.2 288.0
vp9_inv_adst_adst_16x16_add_neon: 4830.4 1380.5
vp9_inv_dct_dct_4x4_add_neon: 153.0 58.6
vp9_inv_dct_dct_8x8_add_neon: 789.2 180.2
vp9_inv_dct_dct_16x16_add_neon: 3639.6 917.1
vp9_inv_dct_dct_32x32_add_neon: 20462.1 4985.0
vp9_inv_wht_wht_4x4_add_neon: 91.0 49.8
The asm is around factor 3-4 faster than C on the cortex-a57 and the asm
is around 30-50% faster on the a57 compared to the a53.
Diego Biurrun [Thu, 12 Mar 2015 23:36:41 +0000 (00:36 +0100)]
ratecontrol: Use correct function pointer casts instead of void*
libavcodec/ratecontrol.c:120:9: warning: ISO C forbids initialization between function pointer and ‘void *’ [-Wpedantic]
libavcodec/ratecontrol.c:121:9: warning: ISO C forbids initialization between function pointer and ‘void *’ [-Wpedantic]
Martin Storsjö [Mon, 10 Oct 2016 06:48:03 +0000 (09:48 +0300)]
arm: vp9: Add NEON loop filters
This work is sponsored by, and copyright, Google.
The implementation tries to have smart handling of cases
where no pixels need the full filtering for the 8/16 width
filters, skipping both calculation and writeback of the
unmodified pixels in those cases. The actual effect of this
is hard to test with checkasm though, since it tests the
full filtering, and the benefit depends on how many filtered
blocks use the shortcut.
The speedup vs C code is around 2-6x. The numbers are quite
inconclusive though, since the checkasm test runs multiple filterings
on top of each other, so later rounds might end up with different
codepaths (different decisions on which filter to apply, based
on input pixel differences). Disabling the early-exit in the asm
doesn't give a fair comparison either though, since the C code
only does the necessary calcuations for each row.
Based on START_TIMER/STOP_TIMER wrapping around a few individual
functions, the speedup vs C code is around 4-9x.
This is pretty similar in runtime to the corresponding routines
in libvpx. (This is comparing vpx_lpf_vertical_16_neon,
vpx_lpf_horizontal_edge_8_neon and vpx_lpf_horizontal_edge_16_neon
to vp9_loop_filter_h_16_8_neon, vp9_loop_filter_v_16_8_neon
and vp9_loop_filter_v_16_16_neon - note that the naming of horizonal
and vertical is flipped between the libraries.)
In order to have stable, comparable numbers, the early exits in both
asm versions were disabled, forcing the full filtering codepath.
Our version is consistently faster on on A7 and A53, marginally slower on
A8, and sometimes faster, sometimes slower on A9 (marginally slower in all
three tests in this particular test run).
Martin Storsjö [Sat, 8 Oct 2016 19:36:18 +0000 (22:36 +0300)]
arm: vp9: Add NEON itxfm routines
This work is sponsored by, and copyright, Google.
For the transforms up to 8x8, we can fit all the data (including
temporaries) in registers and just do a straightforward transform
of all the data. For 16x16, we do a transform of 4x16 pixels in
4 slices, using a temporary buffer. For 32x32, we transform 4x32
pixels at a time, in two steps of 4x16 pixels each.
This is mostly marginally faster than the corresponding routines
in libvpx on most cores, tested with their 32x32 idct (compared to
vpx_idct32x32_1024_add_neon). These numbers are slightly in libvpx's
favour since their version doesn't clear the input buffer like ours
do (although the effect of that on the total runtime probably is
negligible.)
Only on the Cortex A8, the libvpx function is faster. On the other cores,
ours is slightly faster even though ours has got source block clearing
integrated.
Mark Thompson [Thu, 3 Nov 2016 00:13:35 +0000 (00:13 +0000)]
pthread_frame: Unreference hw_frames_ctx on per-thread codec contexts
When decoding with threads enabled, the get_format callback will be
called with one of the per-thread codec contexts rather than with the
outer context. If a hwaccel is in use too, this will add a reference
to the hardware frames context on that codec context, which will then
propagate to all of the other per-thread contexts for decoding. Once
the decoder finishes, however, the per-thread contexts are not freed
normally, so these references leak.
Martin Storsjö [Thu, 10 Nov 2016 09:07:39 +0000 (11:07 +0200)]
arm: vp9mc: Minor adjustments from review of the aarch64 version
This work is sponsored by, and copyright, Google.
The speedup for the large horizontal filters is surprisingly
big on A7 and A53, while there's a minor slowdown (almost within
measurement noise) on A8 and A9.
Martin Storsjö [Thu, 10 Nov 2016 09:06:26 +0000 (11:06 +0200)]
aarch64: vp9: Add NEON optimizations of VP9 MC functions
This work is sponsored by, and copyright, Google.
These are ported from the ARM version; it is essentially a 1:1
port with no extra added features, but with some hand tuning
(especially for the plain copy/avg functions). The ARM version
isn't very register starved to begin with, so there's not much
to be gained from having more spare registers here - we only
avoid having to clobber callee-saved registers.
These are generally about as fast as the corresponding ARM
routines on the same CPU (at least on the A53), in most cases
marginally faster.
The speedup vs C code is pretty much the same as for the 32 bit
case; on the A53 it's around 6-13x for ther larger 8tap filters.
The exact speedup varies a little, since the C versions generally
don't end up exactly as slow/fast as on 32 bit.
James Almer [Wed, 9 Nov 2016 21:38:24 +0000 (18:38 -0300)]
matroskaenc: write updated STREAMINFO metadata for FLAC streams if available
FLAC streams originating from the FLAC encoder send updated and more
complete STREAMINFO metadata as part of the last packet, so write that
to CodecPrivate instead of the incomplete one available in extradata
during init.
Signed-off-by: James Almer <jamrial@gmail.com> Signed-off-by: Anton Khirnov <anton@khirnov.net>
James Almer [Wed, 9 Nov 2016 21:38:23 +0000 (18:38 -0300)]
matroskaenc: fix muxing AAC streams when using aac_adtstoasc bsf
aac_adtstoasc makes the aac extradata available only after the first packet
is filtered, and as packet side data.
Assume extradata will be available as part of the first packet if
avpriv_mpeg4audio_get_config() fails the first time due to missing extradata
and reserve space for the OutputSampleRate element in the Tracks master.
Signed-off-by: James Almer <jamrial@gmail.com> Signed-off-by: Anton Khirnov <anton@khirnov.net>
Anton Khirnov [Wed, 10 Aug 2016 12:58:08 +0000 (14:58 +0200)]
hwcontext_qsv: transfer data through the child context when VPP fails
Uploading/downloading data through VPP may not work for some formats, in
that case we can still try to call av_hwframe_transfer_data() on the
child context.
Anton Khirnov [Wed, 10 Aug 2016 11:01:55 +0000 (13:01 +0200)]
hwcontext_qsv: add support for the P8 format
When using GPU surfaces with QSV, one needs to supply a frame allocator,
which will be invoked to pass surface pools to libmfx.
For encoding, this allocator gets invoked not only for the pool of input
frames, but also for a separate pool of (apparently) reconstructed frames
and another pool of MFX_FOURCC_P8, which on Windows needs to return
D3DFMT_P8 D3D surfaces. Those are probably used to store the encoded
bitstream on the GPU.
Diego Biurrun [Fri, 30 Oct 2015 13:04:08 +0000 (14:04 +0100)]
x86: Drop stray semicolons after function definitions
libavcodec/x86/rv40dsp_init.c:97:2: warning: ISO C does not allow extra ‘;’ outside of a function [-Wpedantic]
libavcodec/x86/vp9dsp_init.c:94:40: warning: ISO C does not allow extra ‘;’ outside of a function [-Wpedantic]
Martin Storsjö [Thu, 3 Nov 2016 12:14:12 +0000 (14:14 +0200)]
configure: Silence lld-link when getting the version number
In recent lld-link versions, this command prints the version to
stdout, but also prints an error to stderr:
$ lld-link -flavor gnu --version
LLD 4.0.0 (trunk 285641)
lld-link: error: no input files
lld-link: error: target emulation unknown: -m or at least one .o file required