On Thu, Nov 21, 2013 at 9:00 PM, Anton Khirnov an...@khirnov.net wrote:
On Mon, 18 Nov 2013 15:19:51 +0100, Kostya Shishkov
kostya.shish...@gmail.com wrote:
On 18 November 2013 12:25, Kostya Shishkov kostya.shish...@gmail.com wrote:
Feel free to encourage Anton to implement MetaVoice.
As
On Mon, Nov 11, 2013 at 10:47 AM, Kostya Shishkov
kostya.shish...@gmail.com wrote:
On 8 November 2013 18:07, Vitor Sessak vitor1...@gmail.com wrote:
Hi
On Fri, Nov 8, 2013 at 10:17 AM, Kostya Shishkov
kostya.shish...@gmail.com wrote:
needed for 22kHz@20kbps modes in MetaSound (and nothing
Hi
On Fri, Nov 8, 2013 at 10:17 AM, Kostya Shishkov
kostya.shish...@gmail.com wrote:
needed for 22kHz@20kbps modes in MetaSound (and nothing else so far)
/**
* Parameters and tables that are different for each frame type
*/
@@ -162,7 +164,8 @@ typedef struct TwinVQContext {
//
On Sat, Oct 19, 2013 at 12:37 AM, Maxim Polijakowski max_p...@gmx.dewrote:
Hello guys,
due to a bunch of comments and fixes I've decided to start a new thread.
Please find attached an updated patch.
Below a detailed overview of what has been changed:
-- init_get_bits has been replaced
On Thu, Oct 10, 2013 at 9:14 PM, Maxim Polijakowski max_p...@gmx.de wrote:
Hi crews,
the attached patch adds an open-source decoder for Sony's ATRAC3+ format.
There is a partial description of its internals located here:
On Mon, Sep 16, 2013 at 9:06 PM, Martin Storsjö mar...@martin.st wrote:
n_div*2 must be less than 1024 (in read_cb_data in twinvqdec.c).
Reported-by: Mateusz j00ru Jurczyk and Gynvael Coldwind
CC: libav-sta...@libav.org
---
libavcodec/twinvq.c | 10 --
1 file changed, 8
On Aug 7, 2013 9:28 AM, Kostya Shishkov kostya.shish...@gmail.com wrote:
---
Now with proper formatting of LSP tables - thanks to Vitor Sessak for
that.
LGTM.
-Vitor
___
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailman
On Mon, Aug 5, 2013 at 8:59 PM, Kostya Shishkov
kostya.shish...@gmail.comwrote:
On Mon, Aug 05, 2013 at 08:40:31PM +0200, Vitor Sessak wrote:
On Mon, Aug 5, 2013 at 12:55 PM, Diego Biurrun di...@biurrun.de wrote:
From: Kostya Shishkov kostya.shish...@gmail.com
Signed-off-by: Diego
On Mon, Aug 5, 2013 at 1:00 PM, Kostya Shishkov
kostya.shish...@gmail.comwrote:
On Mon, Aug 05, 2013 at 12:57:55PM +0200, Diego Biurrun wrote:
On Mon, Aug 05, 2013 at 12:55:29PM +0200, Diego Biurrun wrote:
From: Kostya Shishkov kostya.shish...@gmail.com
Signed-off-by: Diego Biurrun
On Mon, Aug 5, 2013 at 12:55 PM, Diego Biurrun di...@biurrun.de wrote:
From: Kostya Shishkov kostya.shish...@gmail.com
Signed-off-by: Diego Biurrun di...@biurrun.de
---
Rebased on top of master.
--- /dev/null
+++ b/libavcodec/metasound_data_lsp.c
@@ -0,0 +1,475 @@
+/*
+ *
On Mon, Aug 5, 2013 at 12:55 PM, Diego Biurrun di...@biurrun.de wrote:
From: Kostya Shishkov kostya.shish...@gmail.com
This is required for the future addition of VoxWare MetaSound decoder, for
its
functions are mostly the same but bitstream reader is completely different
and bitstream
On Thu, Aug 1, 2013 at 12:29 PM, Kostya Shishkov
kostya.shish...@gmail.comwrote:
On Thu, Aug 01, 2013 at 12:26:30PM +0200, Diego Biurrun wrote:
---
libavcodec/twinvq.c | 189
++-
1 file changed, 95 insertions(+), 94 deletions(-)
on a
Hi
On Jul 30, 2013 12:36 PM, Kostya Shishkov kostya.shish...@gmail.com
wrote:
+#include twinvq.h
+#include metasound_data.h
/**
* Inverse quantization. Read CB coefficients for cb1 and cb2 from the
@@ -135,22 +58,21 @@ static void dequant(TwinContext *tctx, GetBitContext
*gb, float
On Tue, Jul 23, 2013 at 2:28 PM, Diego Biurrun di...@biurrun.de wrote:
---
Now taking the comments from Vitor and Luca into account.
libavcodec/twinvq.c | 584
--
libavcodec/twinvq_data.h | 325 +-
2 files
On Jul 23, 2013 11:37 PM, Diego Biurrun di...@biurrun.de wrote:
---
Now with Vitor's comments addressed.
One last comment.
= (bit_size + 13) / 14;
-rounded_up = (bit_size + tctx-n_div[i] - 1)/tctx-n_div[i];
-rounded_down = (bit_size )/tctx-n_div[i];
-
Hello again,
On Fri, Jul 19, 2013 at 9:22 AM, Kostya Shishkov
kostya.shish...@gmail.comwrote:
---
It's good enough for the only file I care about but maybe it's useful for
the
others too.
The source can benefit from heavy diegoing (because it's mostly TwinVQ
decoder
bits that were
Hello,
On Jul 20, 2013 12:26 AM, Diego Biurrun di...@biurrun.de wrote:
---
I removed parts of the changes to twinvq_data.h, as adding 4 spaces of
indentation to all those tables makes for a 1.5MB patch. I kid you not ..
If what you didn't like about twinvq_data.h is that the huge tables
Hello!
On Fri, Jul 19, 2013 at 9:22 AM, Kostya Shishkov
kostya.shish...@gmail.comwrote:
---
It's good enough for the only file I care about but maybe it's useful for
the
others too.
The source can benefit from heavy diegoing (because it's mostly TwinVQ
decoder
bits that were tweaked to
On Sat, Jun 15, 2013 at 2:16 PM, Diego Biurrun di...@biurrun.de wrote:
On 2013-06-14 22:55, Vitor Sessak wrote:
On Fri, Jun 14, 2013 at 7:09 PM, Vitor Sessak vitor1...@gmail.com
wrote:
--- /dev/null
+++ b/libavcodec/on2avcdata.c
@@ -0,0 +1,7104 @@
+#include on2avcdata.h
+
+const int
On Sat, Jun 15, 2013 at 9:19 AM, Kostya Shishkov
kostya.shish...@gmail.comwrote:
On Fri, Jun 14, 2013 at 10:55:55PM +0200, Vitor Sessak wrote:
On Fri, Jun 14, 2013 at 7:09 PM, Vitor Sessak vitor1...@gmail.com
wrote:
Hi!
Weird, I'have no idea where the rest of my email is gone
On Sat, Jun 15, 2013 at 8:44 PM, Diego Biurrun di...@biurrun.de wrote:
On 2013-06-15 16:25, Vitor Sessak wrote:
On Sat, Jun 15, 2013 at 2:16 PM, Diego Biurrun di...@biurrun.de wrote:
On 2013-06-14 22:55, Vitor Sessak wrote:
On Fri, Jun 14, 2013 at 7:09 PM, Vitor Sessak vitor1...@gmail.com
On Fri, Jun 14, 2013 at 7:09 PM, Vitor Sessak vitor1...@gmail.com wrote:
Hi!
Weird, I'have no idea where the rest of my email is gone. Posting it again
:-(
---
aka The Most Original Audio Codec in the World
Reconstruction is not perfect but at least it's recognisable (there are
some
On Wed, Jan 16, 2013 at 1:58 AM, Ronald S. Bultje rsbul...@gmail.com wrote:
Hi!
+INIT_XMM sse
+cglobal vorbis_inverse_coupling, 3, 3, 6, mag, ang, block_size
+movsxdifnidnblock_sizeq, block_sized
+mova m5, [pdw_8000]
+leamagq,
On 09/25/2012 04:13 PM, Justin Ruggles wrote:
Use the new function in the volume filter.
---
C: 22249
SSE: 1721
AVX: 1219
libavfilter/af_volume.c|6 +++---
libavutil/float_dsp.c |9 +
libavutil/float_dsp.h | 15 +++
On 09/21/2012 05:23 PM, Justin Ruggles wrote:
---
libavcodec/twinvq.c | 31 ---
1 files changed, 12 insertions(+), 19 deletions(-)
diff --git a/libavcodec/twinvq.c b/libavcodec/twinvq.c
index ab22f64..d009196 100644
--- a/libavcodec/twinvq.c
+++
On 05/11/2012 10:31 PM, Vitor Sessak wrote:
---
libavcodec/x86/fft.c | 12 +++---
libavcodec/x86/fft.h | 12 +++---
libavcodec/x86/fft_3dn2.c | 26 ++--
libavcodec/x86/fft_mmx.asm | 101 ++-
4 files changed, 77 insertions
On 05/11/2012 10:32 PM, Vitor Sessak wrote:
---
libavutil/x86/x86util.asm | 13 ++---
1 files changed, 6 insertions(+), 7 deletions(-)
Ping.
-Vitor
___
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailman
patch attached.
-Vitor
From 9903a9685de6687517abeb3263e29ae5810ce3db Mon Sep 17 00:00:00 2001
From: Vitor Sessak vitor1...@gmail.com
Date: Thu, 10 May 2012 20:53:08 +0200
Subject: [PATCH] x86: use new schema for ASM macros
---
libavcodec/x86/fft.c | 12 ++--
libavcodec/x86/fft.h
On 05/15/2012 02:23 PM, Mans Rullgard wrote:
While these codecs are covered elsewhere, the container is different
from the other psx-str file.
Again, there are two types of mdec files. This is the test for the v2
type...
-Vitor
___
libav-devel
On 05/15/2012 06:51 PM, Måns Rullgård wrote:
Vitor Sessakvitor1...@gmail.com writes:
On 05/15/2012 02:23 PM, Mans Rullgard wrote:
While these codecs are covered elsewhere, the container is different
from the other psx-str file.
Again, there are two types of mdec files. This is the test for
On 05/14/2012 05:59 PM, Måns Rullgård wrote:
Vitor Sessakvitor1...@gmail.com writes:
On 05/12/2012 02:17 PM, Mans Rullgard wrote:
The codecs in this file are tested elsewhere. The container differs
from psx-str-demux in that it has a RIFF header, which the new name
reflects.
This test is
On 05/13/2012 11:40 PM, Diego Elio Pettenò wrote:
Il 12/05/2012 01:15, Vitor Sessak ha scritto:
Here is a version with 128-bit regs (again, completely untested).
The patch seems to require a new series that factors x86: use more
standard construct for setting ASM funcitons in FFT code
---
libavcodec/x86/fft.c | 45 +++--
1 files changed, 23 insertions(+), 22 deletions(-)
diff --git a/libavcodec/x86/fft.c b/libavcodec/x86/fft.c
index 3e0c42f..5495821 100644
--- a/libavcodec/x86/fft.c
+++ b/libavcodec/x86/fft.c
@@ -25,30 +25,31 @@
On 05/12/2012 02:17 PM, Mans Rullgard wrote:
The codecs in this file are tested elsewhere. The container differs
from psx-str-demux in that it has a RIFF header, which the new name
reflects.
This test is AFAIK the only one who tests the decode_dc() path in the
MDEC decoder. Maybe it should
---
libavcodec/x86/fft.c | 12 +++---
libavcodec/x86/fft.h | 12 +++---
libavcodec/x86/fft_3dn2.c | 26 ++--
libavcodec/x86/fft_mmx.asm | 101 ++-
4 files changed, 77 insertions(+), 74 deletions(-)
diff --git
---
libavutil/x86/x86util.asm | 13 ++---
1 files changed, 6 insertions(+), 7 deletions(-)
diff --git a/libavutil/x86/x86util.asm b/libavutil/x86/x86util.asm
index 508f24e..066384b 100644
--- a/libavutil/x86/x86util.asm
+++ b/libavutil/x86/x86util.asm
@@ -84,13 +84,12 @@
%macro
On 05/11/2012 10:31 PM, Vitor Sessak wrote:
---
libavcodec/x86/fft.c |9 +++-
libavcodec/x86/fft.h |2 +
libavcodec/x86/fft_mmx.asm | 108 +---
libavcodec/x86/fft_sse.c |7 +++
libavutil/x86/x86inc.asm |4 +-
5 files
On 05/09/2012 03:49 PM, Justin Ruggles wrote:
The current SSE version is slower than the MMX version on Athlon64 and Sandy
Bridge, but the SSE4 and AVX versions are faster on Sandy Bridge.
---
libavresample/x86/audio_convert.asm| 30 --
On 05/08/2012 09:59 PM, Justin Ruggles wrote:
On 05/07/2012 11:40 AM, Vitor Sessak wrote:
On 05/07/2012 04:29 PM, Justin Ruggles wrote:
On 05/07/2012 08:18 AM, Vitor Sessak wrote:
On 05/03/2012 09:39 PM, Justin Ruggles wrote:
The SSE version is slower than the MMX version on Athlon64
On 05/03/2012 09:39 PM, Justin Ruggles wrote:
The SSE version is slower than the MMX version on Athlon64 and Sandy Bridge,
but the 3-arg AVX version is faster on Sandy Bridge.
---
Athlon64
C - 26322
MMX - 7997
SSE - 8662
Sandy Bridge
C - 13011
MMX - 4262
SSE - 4680
AVX - 3960
Note
On 05/07/2012 04:29 PM, Justin Ruggles wrote:
On 05/07/2012 08:18 AM, Vitor Sessak wrote:
On 05/03/2012 09:39 PM, Justin Ruggles wrote:
The SSE version is slower than the MMX version on Athlon64 and Sandy
Bridge,
but the 3-arg AVX version is faster on Sandy Bridge.
---
Athlon64
C - 26322
On 05/02/2012 11:18 PM, Justin Ruggles wrote:
On 05/02/2012 05:01 PM, Vitor Sessak wrote:
On 05/02/2012 08:53 PM, Justin Ruggles wrote:
On 05/02/2012 02:46 PM, Vitor Sessak wrote:
On 05/01/2012 10:49 PM, Justin Ruggles wrote:
---
Athlon64
C - 8176
SSE2 - 1784
SandyBridge
C
On 05/01/2012 10:49 PM, Justin Ruggles wrote:
---
Athlon64
C - 8176
SSE2 - 1784
SandyBridge
C - 5578
SSE2 - 418
Is there any reason why not to do an AVX (YMM) version?
-Vitor
___
libav-devel mailing list
libav-devel@libav.org
On 05/02/2012 08:53 PM, Justin Ruggles wrote:
On 05/02/2012 02:46 PM, Vitor Sessak wrote:
On 05/01/2012 10:49 PM, Justin Ruggles wrote:
---
Athlon64
C - 8176
SSE2 - 1784
SandyBridge
C - 5578
SSE2 - 418
Is there any reason why not to do an AVX (YMM) version?
It was actually
On 05/01/2012 07:27 PM, Mans Rullgard wrote:
ModeTab.fmode has only 3 elements, so indexing it with ftype
in the initialier for 'size' is invalid when ftype == FT_PPC.
This fixes crashes with gcc 4.8.
Signed-off-by: Mans Rullgardm...@mansr.com
---
libavcodec/twinvq.c |6 --
1 file
On 04/13/2012 10:30 PM, Kieran Kunhya wrote:
vshufps with ymm is a four operand instruction so a mova is still
necessary. I'll benchmark it anyway.
Oops, indeed. But the following works for me:
vperm2f128 ymm0, ymm0, [bufq], 0x23
vshufps ymm0, ymm0, ymm0, q0123
-Vitor
On Fri, Apr 13, 2012 at 6:53 AM, Loren Merritt lor...@u.washington.edu wrote:
On Thu, 12 Apr 2012, Justin Ruggles wrote:
On 04/12/2012 08:39 PM, Kieran Kunhya wrote:
If anyone knows of a way to reverse floats in a YMM then I could add
AVX to the reverse fmul as well.
Might be slower
On 03/24/2012 07:35 AM, Kostya Shishkov wrote:
Those descriptions are stored in nibbles, so they are easy to extract.
And this way we don't need to pad tables for possible bit reader overreads.
---
libavcodec/ralf.c |8
1 files changed, 4 insertions(+), 4 deletions(-)
LGTM
On 03/23/2012 07:17 AM, Kostya Shishkov wrote:
On Thu, Mar 22, 2012 at 01:46:13PM -0700, Ronald S. Bultje wrote:
Hi,
On Thu, Mar 22, 2012 at 1:08 PM, Kostya Shishkov
kostya.shish...@gmail.com wrote:
On Thu, Mar 22, 2012 at 12:55:38PM -0700, Ronald S. Bultje wrote:
From: Ronald S.
On 03/18/2012 08:43 AM, Kostya Shishkov wrote:
On Sat, Mar 17, 2012 at 10:16:01PM +0100, Vitor Sessak wrote:
On 03/17/2012 08:54 AM, Kostya Shishkov wrote:
$subj
ralfdata.h is attached separately because it's less interesting and it's more
than 600k in plain form (yet it's still smaller
On 03/17/2012 08:54 AM, Kostya Shishkov wrote:
$subj
ralfdata.h is attached separately because it's less interesting and it's more
than 600k in plain form (yet it's still smaller that libavcodec/twinvq_data.h).
Just two small things:
#define FILTERPARAM_ELEMENTS 643
#define BIAS_ELEMENTS
On 03/17/2012 08:54 AM, Kostya Shishkov wrote:
$subj
ralfdata.h is attached separately because it's less interesting and it's more
than 600k in plain form (yet it's still smaller that libavcodec/twinvq_data.h).
A couple more comments:
+}
+
+static int decode_block(AVCodecContext *avctx,
On 02/24/2012 10:03 AM, Christophe Gisquet wrote:
Hi,
I noticed this appearing in some profile logs, and discovered an
_interleave implementation was already there, which had most of the
code already done.
I can't evaluate the AVX implementation. In particular, I'm not sure
all buffers of all
On 03/03/2012 11:04 AM, Christophe Gisquet wrote:
Hi,
scalarproduct_int16 isn't entirely fit for wma lossless, because:
1) it is assumed that the number of loops is a multiple of 16
wmalossless needs 8 instead (cf. decoding of cdlms' order in decode_cdlms)
MMX2 implementation is fine, but SSE2
On 03/03/2012 11:29 AM, Christophe Gisquet wrote:
2012/3/3 Vitor Sessakvitor1...@gmail.com:
This function is at the moment not used in libav. It is used in a G729
decoder committed to FFmpeg, but there it is only used with shift=0.
It is also used in ff_acelp_decode_gain_code again with
On 03/03/2012 11:08 AM, Christophe Gisquet wrote:
Hi,
2012/3/3 Vitor Sessakvitor1...@gmail.com:
C: 12400
SSE: 3344
AVX: 2542
Yes, different lengths and cpus, but that sounds about right.
I haven't set myself up to using obe2 (ssh key issue, downtime etc),
so I'm glad you took the time to do
On 03/02/2012 07:30 PM, Ronald S. Bultje wrote:
Hi,
On Thu, Mar 1, 2012 at 3:44 PM, Ronald S. Bultjersbul...@gmail.com wrote:
From: Ronald S. Bultjersbul...@gmail.com
Prevents warnings because the dst and src overlap (are the same) in the
memcpy() inside the function.
---
On 02/29/2012 10:47 PM, Måns Rullgård wrote:
Vitor Sessakvitor1...@gmail.com writes:
On 02/29/2012 04:28 PM, Janne Grunau wrote:
On 2012-02-26 09:52:44 +0100, Vitor Sessak wrote:
---
libavcodec/ra144dec.c |2 ++
libavcodec/ra288.c|2 ++
libavcodec/sipr.c |2
On 02/29/2012 11:19 PM, Ronald S. Bultje wrote:
Hi,
On Wed, Feb 29, 2012 at 1:12 PM, Ronald S. Bultjersbul...@gmail.com wrote:
On Wed, Feb 29, 2012 at 1:09 PM, Vitor Sessakvitor1...@gmail.com wrote:
---
libavcodec/amrnbdec.c |8
1 files changed, 8 insertions(+), 0
On 02/29/2012 10:50 PM, Ronald S. Bultje wrote:
Hi,
2012/2/29 Måns Rullgårdm...@mansr.com:
Vitor Sessakvitor1...@gmail.com writes:
On 02/29/2012 04:28 PM, Janne Grunau wrote:
On 2012-02-26 09:52:44 +0100, Vitor Sessak wrote:
---
libavcodec/ra144dec.c |2 ++
libavcodec/ra288.c
---
libavcodec/nellymoserdec.c | 16
1 files changed, 8 insertions(+), 8 deletions(-)
diff --git a/libavcodec/nellymoserdec.c b/libavcodec/nellymoserdec.c
index 96fb053..1583c20 100644
--- a/libavcodec/nellymoserdec.c
+++ b/libavcodec/nellymoserdec.c
@@ -49,14 +49,15 @@
On 02/29/2012 04:28 PM, Janne Grunau wrote:
On 2012-02-26 09:52:44 +0100, Vitor Sessak wrote:
---
libavcodec/ra144dec.c |2 ++
libavcodec/ra288.c|2 ++
libavcodec/sipr.c |2 ++
libavcodec/twinvq.c |2 ++
4 files changed, 8 insertions(+), 0 deletions(-)
Why
---
libavcodec/amrnbdec.c |7 +++
1 files changed, 7 insertions(+), 0 deletions(-)
diff --git a/libavcodec/amrnbdec.c b/libavcodec/amrnbdec.c
index fff0e72..d88f5f0 100644
--- a/libavcodec/amrnbdec.c
+++ b/libavcodec/amrnbdec.c
@@ -946,6 +946,13 @@ static int
On 02/26/2012 03:21 PM, Diego Biurrun wrote:
On Sun, Feb 26, 2012 at 12:00:21PM +0100, Vitor Sessak wrote:
--- a/libavcodec/amrnbdec.c
+++ b/libavcodec/amrnbdec.c
@@ -946,6 +946,13 @@ static int amrnb_decode_frame(AVCodecContext *avctx, void
*data,
+if (buf_size frame_sizes_nb[p
---
libavcodec/mpegaudiodec.c |7 +++
1 files changed, 7 insertions(+), 0 deletions(-)
diff --git a/libavcodec/mpegaudiodec.c b/libavcodec/mpegaudiodec.c
index d82432c..d6a09c8 100644
--- a/libavcodec/mpegaudiodec.c
+++ b/libavcodec/mpegaudiodec.c
@@ -29,6 +29,7 @@
#include get_bits.h
On 02/26/2012 02:10 PM, Benjamin Larsson wrote:
OK
ping (and patch updated with better commit message).
-Vitor
___
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailman/listinfo/libav-devel
On 02/29/2012 09:08 PM, Ronald S. Bultje wrote:
Hi,
On Wed, Feb 29, 2012 at 12:03 PM, Vitor Sessakvitor1...@gmail.com wrote:
---
libavcodec/amrnbdec.c |7 +++
1 files changed, 7 insertions(+), 0 deletions(-)
diff --git a/libavcodec/amrnbdec.c b/libavcodec/amrnbdec.c
index
On 02/28/2012 11:23 PM, Ronald S. Bultje wrote:
From: Ronald S. Bultjersbul...@gmail.com
Check that the mode is valid before using it as an index into tables.
Also provide a size parameter to ff_amr_bit_reorder() to prevent
overreads.
Found-by: Mateusz j00ru Jurczyk and Gynvael Coldwind
CC:
On 02/29/2012 09:42 PM, Ronald S. Bultje wrote:
Hi,
On Wed, Feb 29, 2012 at 12:41 PM, Vitor Sessakvitor1...@gmail.com wrote:
On 02/29/2012 09:08 PM, Ronald S. Bultje wrote:
On Wed, Feb 29, 2012 at 12:03 PM, Vitor Sessakvitor1...@gmail.com
wrote:
---
libavcodec/amrnbdec.c |7 +++
---
libavcodec/amrnbdec.c |7 +++
1 files changed, 7 insertions(+), 0 deletions(-)
diff --git a/libavcodec/amrnbdec.c b/libavcodec/amrnbdec.c
index fff0e72..d88f5f0 100644
--- a/libavcodec/amrnbdec.c
+++ b/libavcodec/amrnbdec.c
@@ -946,6 +946,13 @@ static int
---
libavcodec/amrnbdec.c |8
1 files changed, 8 insertions(+), 0 deletions(-)
diff --git a/libavcodec/amrnbdec.c b/libavcodec/amrnbdec.c
index fff0e72..cdfcfc7 100644
--- a/libavcodec/amrnbdec.c
+++ b/libavcodec/amrnbdec.c
@@ -200,6 +200,10 @@ static enum Mode
On 02/29/2012 10:02 PM, Ronald S. Bultje wrote:
Hi,
On Wed, Feb 29, 2012 at 12:51 PM, Vitor Sessakvitor1...@gmail.com wrote:
---
libavcodec/amrnbdec.c |7 +++
1 files changed, 7 insertions(+), 0 deletions(-)
diff --git a/libavcodec/amrnbdec.c b/libavcodec/amrnbdec.c
index
---
libavcodec/ra144dec.c |2 ++
libavcodec/ra288.c|2 ++
libavcodec/sipr.c |2 ++
libavcodec/twinvq.c |2 ++
4 files changed, 8 insertions(+), 0 deletions(-)
diff --git a/libavcodec/ra144dec.c b/libavcodec/ra144dec.c
index dd8838c..7ec04ec 100644
---
On 02/26/2012 05:23 AM, Alex Converse wrote:
On Sat, Feb 25, 2012 at 11:50 AM, Vitor Sessakvitor1...@gmail.com wrote:
---
libavcodec/amrnbdec.c |1 +
libavcodec/amrwbdec.c |2 ++
Looking at AMR-WB, I think the best solution might be to not use
get_bits at all. The GetBitContext is
On 02/26/2012 02:32 AM, Ronald S. Bultje wrote:
Hi,
On Sat, Feb 25, 2012 at 12:05 PM, Ronald S. Bultjersbul...@gmail.com wrote:
On Sat, Feb 25, 2012 at 11:50 AM, Vitor Sessakvitor1...@gmail.com wrote:
---
libavcodec/amrnbdec.c |1 +
libavcodec/amrwbdec.c |2 ++
---
libavcodec/amrnbdec.c |7 +++
1 files changed, 7 insertions(+), 0 deletions(-)
diff --git a/libavcodec/amrnbdec.c b/libavcodec/amrnbdec.c
index fff0e72..5fe746a 100644
--- a/libavcodec/amrnbdec.c
+++ b/libavcodec/amrnbdec.c
@@ -946,6 +946,13 @@ static int
---
libavcodec/atrac1.c |2 ++
libavcodec/atrac3.c |2 ++
libavcodec/g722dec.c |2 ++
libavcodec/g726.c |3 +++
libavcodec/qcelpdec.c |2 ++
5 files changed, 11 insertions(+), 0 deletions(-)
diff --git a/libavcodec/atrac1.c b/libavcodec/atrac1.c
index
On 02/26/2012 03:21 PM, Diego Biurrun wrote:
On Sun, Feb 26, 2012 at 12:27:46PM +0100, Vitor Sessak wrote:
---
libavcodec/atrac1.c |2 ++
libavcodec/atrac3.c |2 ++
libavcodec/g722dec.c |2 ++
libavcodec/g726.c |3 +++
libavcodec/qcelpdec.c |2 ++
5 files
On 02/26/2012 12:27 PM, Vitor Sessak wrote:
---
libavcodec/atrac1.c |2 ++
libavcodec/atrac3.c |2 ++
libavcodec/g722dec.c |2 ++
libavcodec/g726.c |3 +++
libavcodec/qcelpdec.c |2 ++
5 files changed, 11 insertions(+), 0 deletions(-)
Just a note to correct
---
libavcodec/mpegaudiodec.c |7 +++
1 files changed, 7 insertions(+), 0 deletions(-)
diff --git a/libavcodec/mpegaudiodec.c b/libavcodec/mpegaudiodec.c
index d82432c..d6a09c8 100644
--- a/libavcodec/mpegaudiodec.c
+++ b/libavcodec/mpegaudiodec.c
@@ -29,6 +29,7 @@
#include get_bits.h
---
libavcodec/amrnbdec.c |1 +
libavcodec/amrwbdec.c |2 ++
libavcodec/ra144dec.c |2 ++
libavcodec/ra288.c|2 ++
libavcodec/sipr.c |2 ++
libavcodec/twinvq.c |2 ++
6 files changed, 11 insertions(+), 0 deletions(-)
diff --git a/libavcodec/amrnbdec.c
On Sun, Feb 5, 2012 at 8:04 PM, Christophe Gisquet
christophe.gisq...@gmail.com wrote:
Hi,
2012/1/28 Christophe Gisquet christophe.gisq...@gmail.com:
Hi,
2012/1/21 Christophe Gisquet christophe.gisq...@gmail.com:
this patch has not been idling for that long, but as I'm pinging for
other
On Sat, Jan 21, 2012 at 2:32 PM, Christophe Gisquet
christophe.gisq...@gmail.com wrote:
Hi,
this patch has not been idling for that long, but as I'm pinging for
other x86 SIMD code...
Actually, the best solution would be to avoid this mempy() altogether,
but I'm not sure how hard would it be.
On Sat, Jan 7, 2012 at 5:39 PM, Ronald S. Bultje rsbul...@gmail.com wrote:
Hi,
On Thu, Jan 5, 2012 at 12:18 PM, Vitor Sessak vitor1...@gmail.com wrote:
[..]
Does this make non-SSE decoding noticeably slower, i.e. should we make
the interleaving be decided upon by the windowing
On Fri, Jan 6, 2012 at 11:50 AM, Loren Merritt lor...@u.washington.edu wrote:
On Thu, 5 Jan 2012, Vitor Sessak wrote:
+; input %1={x1,x2,x3,x4}, %2={y1,y2,y3,y4}
+; output %3={x4,y1,y2,y3}
+%macro ROTLEFT_SSE 3
+ BUILDINVHIGHLOW %1, %2, %3
+ shufps %3, %3, %2, 0x99
+%endmacro
(and other
On Sat, Jan 7, 2012 at 1:10 PM, Diego Biurrun di...@biurrun.de wrote:
On Sat, Jan 07, 2012 at 12:53:00PM +0100, Vitor Sessak wrote:
New patch attached.
--- a/libavcodec/x86/mpegaudiodec_mmx.c
+++ b/libavcodec/x86/mpegaudiodec_mmx.c
@@ -147,11 +159,79 @@ static void apply_window_mp3(float
From 57aa0e8757c9de303db7ad52f4af15b41c1de854 Mon Sep 17 00:00:00 2001
From: Vitor Sessak vitor1...@gmail.com
Date: Tue, 3 Jan 2012 21:25:59 +0100
Subject: [PATCH] mpegaudiodec: Use clearer pointer math
---
libavcodec/mpegaudiodec.c |2 +-
1 files changed, 1 insertions(+), 1 deletions
From 6d582bee4ddd2ada01ee777629df2aa9cb2a8726 Mon Sep 17 00:00:00 2001
From: Vitor Sessak vitor1...@gmail.com
Date: Tue, 3 Jan 2012 21:40:57 +0100
Subject: [PATCH] mpegaudiodec: interleave iMDCT buffer to simplify future SIMD
implementations
---
libavcodec/mpegaudiodec.c | 50
From 1b018fdd9dd483cdf4b37945b565889009d1860d Mon Sep 17 00:00:00 2001
From: Vitor Sessak vitor1...@gmail.com
Date: Wed, 4 Jan 2012 21:32:47 +0100
Subject: [PATCH] mpegaudiodec: move imdct and windowing function to mpegaudiodsp
---
libavcodec/mpegaudiodec.c | 193
From 98d1c9cfd2eeb18ef736044025fe62dffd0851ad Mon Sep 17 00:00:00 2001
From: Vitor Sessak vitor1...@gmail.com
Date: Thu, 5 Jan 2012 20:26:33 +0100
Subject: [PATCH] mpegaudiodec: optimized iMDCT transform
---
libavcodec/mpegaudiodec.c |2 +-
libavcodec/x86/Makefile |1
Without this, code like shufps m0, m1, m2, 0xaa would not work in
CPUs not supporting SSE2.
This is needed for my IMDCT36() patch (committed in ffmpeg, WIP for libav).
-Vitor
From 711295e4e3a82eb47c20b87d391220d533c154e5 Mon Sep 17 00:00:00 2001
From: Vitor Sessak vitor1...@gmail.com
Date: Sun
On Sat, Sep 3, 2011 at 12:50 PM, Vitor Sessak vitor1...@gmail.com wrote:
On Fri, Sep 2, 2011 at 3:04 PM, Loren Merritt lor...@u.washington.edu wrote:
On Fri, 2 Sep 2011, Vitor Sessak wrote:
; input %1={x1,x2,x3,x4}, %2={y1,y2,y3,y4}
; output %3={x4,y1,y2,y3}
%macro ROTLEFT 3
On Mon, Sep 5, 2011 at 10:52 PM, Luca Barbato lu_z...@gentoo.org wrote:
On 9/5/11 8:33 PM, Vitor Sessak wrote:
On Sat, Sep 3, 2011 at 12:50 PM, Vitor Sessakvitor1...@gmail.com wrote:
On Fri, Sep 2, 2011 at 3:04 PM, Loren Merrittlor...@u.washington.edu
wrote:
On Fri, 2 Sep 2011, Vitor
2011/9/2 Måns Rullgård m...@mansr.com:
Vitor Sessak vitor1...@gmail.com writes:
On Sun, Aug 21, 2011 at 4:51 PM, Vitor Sessak vitor1...@gmail.com wrote:
$subj, should allow SIMD optimizations.
ping (new version attached).
-Vitor
From 9102dae3eca565e86e5ed307d79a20952771e369 Mon Sep 17 00
On Sat, Sep 3, 2011 at 2:01 AM, Ronald S. Bultje rsbul...@gmail.com wrote:
Hi,
On Thu, Sep 1, 2011 at 9:54 PM, Vitor Sessak vitor1...@gmail.com wrote:
On Sun, Aug 21, 2011 at 4:53 PM, Vitor Sessak vitor1...@gmail.com wrote:
$subj. A lot faster on my Atom, much less impressive difference
On Fri, Sep 2, 2011 at 3:04 PM, Loren Merritt lor...@u.washington.edu wrote:
On Fri, 2 Sep 2011, Vitor Sessak wrote:
; input %1={x1,x2,x3,x4}, %2={y1,y2,y3,y4}
; output %3={x4,y1,y2,y3}
%macro ROTLEFT 3
BUILDINVHIGHLOW %1, %2, %3
shufps %3, %3, %2, 0x99
%endmacro
palignr
New
On Sun, Aug 21, 2011 at 4:51 PM, Vitor Sessak vitor1...@gmail.com wrote:
$subj, should allow SIMD optimizations.
ping (new version attached).
-Vitor
From 9102dae3eca565e86e5ed307d79a20952771e369 Mon Sep 17 00:00:00 2001
From: Vitor Sessak vitor1...@gmail.com
Date: Mon, 22 Aug 2011 07:58:56
On Sun, Aug 21, 2011 at 4:53 PM, Vitor Sessak vitor1...@gmail.com wrote:
$subj. A lot faster on my Atom, much less impressive difference for others
CPUs.
ping (new version attached).
-Vitor
From c49e56aeda477c499d523ea34ee65947bae6cc39 Mon Sep 17 00:00:00 2001
From: Vitor Sessak vitor1
$subj, should allow SIMD optimizations.
-Vitor
From 881453980ae49a52f884754e08fc31633d3b3a08 Mon Sep 17 00:00:00 2001
From: Vitor Sessak vitor1...@gmail.com
Date: Sun, 21 Aug 2011 16:41:25 +0200
Subject: [PATCH 1/2] mpegaudiodec: move imdct36() to MPADSPContext
---
libavcodec/mpegaudiodec.c
$subj. A lot faster on my Atom, much less impressive difference for others CPUs.
-Vitor
From a57855f575b0ef85d95dbe318ebe988314c770dc Mon Sep 17 00:00:00 2001
From: Vitor Sessak vitor1...@gmail.com
Date: Sun, 21 Aug 2011 16:44:20 +0200
Subject: [PATCH 2/2] mpegaudiodec: add SSE-optimized imdct36
1 - 100 of 146 matches
Mail list logo
