---
avconv.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/avconv.c b/avconv.c
index 13e6778..7ac175b 100644
--- a/avconv.c
+++ b/avconv.c
@@ -512,7 +512,9 @@ static void do_video_out(AVFormatContext *s,
in_picture-pts != AV_NOPTS_VALUE
in_picture-pts
On Wed, 19 Mar 2014 09:57:47 +0100, Luca Barbato lu_z...@gentoo.org wrote:
---
avconv.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/avconv.c b/avconv.c
index 13e6778..7ac175b 100644
--- a/avconv.c
+++ b/avconv.c
@@ -512,7 +512,9 @@ static void
Some files set the PreviousPartition field to point to its own offset.
If we are parsing forward the Previous partition is immediately known
and its value could be used, otherwise we can safely point to the
header.
Reported-By: Jean Baptiste Kempf j...@videolan.org
---
libavformat/mxfdec.c | 20
This fixes fate-float_dsp-test on RVCT 4.0.
---
libavutil/float_dsp.c | 48 +---
1 file changed, 25 insertions(+), 23 deletions(-)
diff --git a/libavutil/float_dsp.c b/libavutil/float_dsp.c
index 23468ad..03c8c11 100644
--- a/libavutil/float_dsp.c
+++
On 2014-03-19 14:51:27 +0200, Martin Storsjö wrote:
This fixes fate-float_dsp-test on RVCT 4.0.
---
libavutil/float_dsp.c | 48 +---
1 file changed, 25 insertions(+), 23 deletions(-)
ok
Janne
___
Although it is allowed it does not make much sense since the immediate
is already 8bits.
---
gas-preprocessor.pl | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl
index 914c41c..692e788 100755
--- a/gas-preprocessor.pl
+++
On 19/03/14 15:08, Janne Grunau wrote:
Although it is allowed it does not make much sense since the immediate
is already 8bits.
---
gas-preprocessor.pl | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl
index 914c41c..692e788
On Wed, 19 Mar 2014, Janne Grunau wrote:
Although it is allowed it does not make much sense since the immediate
is already 8bits.
---
gas-preprocessor.pl | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl
index 914c41c..692e788 100755
On Sat, Mar 15, 2014 at 11:56:24AM +0100, Anton Khirnov wrote:
On Fri, 14 Mar 2014 05:42:19 -0700, Diego Biurrun di...@biurrun.de wrote:
---
libavcodec/arm/dsputil_init_armv6.c | 11 +++---
libavcodec/bfin/dsputil_init.c | 23 +++--
libavcodec/dsputil.c| 67
Hey,
I want to discuss Project Adobe DNG Decoder(Basic Support) with
mentor Justin
Ruggles. But don't know How to?
Also, I am beginner in IRC. I don't know how to contact to mentor. Can
Somebody help me?.
With Regards,
Preeti Soni
--
*sayonara :)*
*Have a nice day *
Verified with profiling that this doesn't have a measurable effect upon
overall performance.
---
libavcodec/mlpdec.c | 40 +++-
libavcodec/mlpdsp.c | 36
libavcodec/mlpdsp.h | 22 ++
3 files
I present here a patch series aimed at improving the performance of Dolby
TrueHD audio decoding on ARM CPUs, with a particular focus on the ARM1176JZF-S
as featured in the Raspberry Pi. To date, only one function had been
optimised, and that was only for x86.
For each optimisation, I am including
Profiling on a Raspberry Pi revealed the best performance to correspond
with VLC_BITS = 5. Results for overall audio decode and the get_vlc2 function
in particular are as follows:
Before After
Mean StdDev Mean StdDev Confidence Change
6:2 total
Profiling results for overall decode and the output_data function in
particular are as follows:
Before After
Mean StdDev Mean StdDev Confidence Change
6:2 total 339.6 15.1 329.3 16.095.8% +3.1% (insignificant)
6:2 function 24.6
Profiling results for overall audio decode and the rematrix_channels function
in particular are as follows:
Before After
Mean StdDev Mean StdDev Confidence Change
6:2 total 370.8 17.0 348.8 20.199.9% +6.3%
6:2 function 46.4 8.4
Profiling results for overall audio decode and the mlp_filter_channel(_arm)
function in particular are as follows:
Before After
Mean StdDev Mean StdDev Confidence Change
6:2 total 380.4 22.0 370.8 17.087.4% +2.6% (insignificant)
Verified with profiling that this doesn't have a measurable effect upon
overall performance.
---
libavcodec/mlpdec.c | 37 -
libavcodec/mlpdsp.c | 33 +
libavcodec/mlpdsp.h | 23 +++
3 files changed, 68
There is no point in populating NuvContext with another DSPContext.
Also split static and dynamic initialization bits to avoid running the
static initialization parts over and over.
---
I tried separating the init function split into a separate commit, but
it was just too annoying, so it remains
---
Amended to not break some long lines for enhanced readability.
libavcodec/arm/hpeldsp_init_arm.c | 2 +-
libavcodec/dsputil.h | 2 --
libavcodec/dsputil_template.c | 2 ++
libavcodec/hpel_template.c | 2 ++
libavcodec/hpeldsp_template.c | 2 ++
On 19/03/14 16:38, preeti soni wrote:
Hey,
Hi, please do not cross post.
I want to discuss Project Adobe DNG Decoder(Basic Support) with
mentor Justin
Ruggles. But don't know How to?
IRC is an option, email is another.
Also, I am beginner in IRC. I don't know how to contact to mentor. Can
On Wed, Mar 19, 2014 at 05:24:27PM +, Ben Avison wrote:
--- a/libavcodec/arm/mlpdsp_init_arm.c
+++ b/libavcodec/arm/mlpdsp_init_arm.c
@@ -41,8 +41,72 @@ void ff_mlp_rematrix_channel_arm(int32_t *samples,
+
+switch (max_matrix_channel) {
+case 1: ch_index = 0; break;
+case 5:
On 19/03/14 18:24, Ben Avison wrote:
Verified with profiling that this doesn't have a measurable effect upon
overall performance.
---
libavcodec/mlpdec.c | 37 -
libavcodec/mlpdsp.c | 33 +
libavcodec/mlpdsp.h | 23
On Wed, Mar 19, 2014 at 05:24:26PM +, Ben Avison wrote:
--- a/libavcodec/mlpdsp.c
+++ b/libavcodec/mlpdsp.c
@@ -89,10 +89,46 @@ void ff_mlp_rematrix_channel(int32_t *samples,
+
+int32_t ff_mlp_pack_output(int32_t lossless_check_data,
This function is not used outside of the file, so it
On Wed, Mar 19, 2014 at 05:24:22PM +, Ben Avison wrote:
--- a/libavcodec/arm/Makefile
+++ b/libavcodec/arm/Makefile
@@ -21,9 +21,13 @@ OBJS-$(CONFIG_H264PRED)+=
arm/h264pred_init_arm.o
OBJS-$(CONFIG_H264QPEL)+= arm/h264qpel_init_arm.o
On Wed, 19 Mar 2014, Luca Barbato wrote:
On 19/03/14 18:24, Ben Avison wrote:
Verified with profiling that this doesn't have a measurable effect upon
overall performance.
---
libavcodec/mlpdec.c | 37 -
libavcodec/mlpdsp.c | 33
On Wed, Mar 19, 2014 at 06:54:44PM +0100, Luca Barbato wrote:
On 19/03/14 18:34, Diego Biurrun wrote:
---
Amended to not break some long lines for enhanced readability.
Readability remains impaired.
+#define HPELDSP_AVG_PIXELS16(CPUEXT) \
+
On Wed, Mar 19, 2014 at 05:24:23PM +, Ben Avison wrote:
--- a/libavcodec/mlpdec.c
+++ b/libavcodec/mlpdec.c
--- a/libavcodec/mlpdsp.c
+++ b/libavcodec/mlpdsp.c
@@ -57,9 +57,42 @@ static void mlp_filter_channel(int32_t *state, const
int32_t *coeff,
+void
---
libavcodec/x86/dsputil_x86.h| 14 --
libavcodec/x86/{hpeldsp_mmx.c = hpeldsp.h} | 43 ++---
libavcodec/x86/hpeldsp_init.c | 2 +-
libavcodec/x86/hpeldsp_mmx.c| 2 +-
libavcodec/x86/rnd_mmx.c|
---
libavcodec/x86/cavsdsp.c | 1 +
libavcodec/x86/dsputil_init.c | 1 +
libavcodec/x86/dsputil_x86.h | 17 -
libavcodec/x86/fpel.h | 43 +++
libavcodec/x86/fpel_mmx.c | 2 +-
libavcodec/x86/h264_qpel.c| 2 +-
This allows detemplatizing the bits that are not instantiated twice.
---
libavcodec/dsputil.c | 3 ++-
libavcodec/dsputil_template.c| 27 -
libavcodec/dsputilenc_template.c | 51
3 files changed, 53 insertions(+), 28
---
libavcodec/ac3enc_template.c | 4
libavcodec/h264_mb_template.c | 2 ++
libavcodec/h264_mc_template.c | 2 ++
libavcodec/h264idct_template.c | 1 +
libavcodec/hpel_template.c | 4
libavcodec/motion_est_template.c | 2 ++
---
libavcodec/hpeldsp_template.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/libavcodec/hpeldsp_template.c b/libavcodec/hpeldsp_template.c
index 6869fe2..246421e 100644
--- a/libavcodec/hpeldsp_template.c
+++ b/libavcodec/hpeldsp_template.c
@@ -309,9 +309,9 @@
There is no point in having this separate; it is not used as a template.
---
libavcodec/hpeldsp.c | 301 +-
libavcodec/hpeldsp_template.c | 331 --
2 files changed, 300 insertions(+), 332 deletions(-)
delete
There is no point in having a separate file just for the instantiation
that provides the public functions.
---
libavcodec/x86/Makefile | 4 +---
libavcodec/x86/hpeldsp_init.c | 9 +
libavcodec/x86/rnd_mmx.c | 37 -
3 files changed, 10
On 19/03/14 2:24 PM, Ben Avison wrote:
diff --git a/libavcodec/mlpdsp.h b/libavcodec/mlpdsp.h
index bd864d9..7b7640e 100644
--- a/libavcodec/mlpdsp.h
+++ b/libavcodec/mlpdsp.h
@@ -23,6 +23,7 @@
#define AVCODEC_MLPDSP_H
#include stdint.h
+#include mlp.h
void
---
libavcodec/ppc/dsputil_altivec.h | 5 -
libavcodec/ppc/h264qpel.c | 2 +-
libavcodec/ppc/hpeldsp_altivec.c | 2 +-
.../ppc/{dsputil_altivec.h = hpeldsp_altivec.h} | 19 ---
Finally, I start to put dsputil on a diet, like the doctor ordered.
There are still preparatory patches that pave the way for later changes,
but some big chunks are broken out of dsputil and off into separate
structures.
[PATCH 070/132] x86: dsputil: Move fpel declarations to a separate
[PATCH
---
configure | 3 +-
doc/optimization.txt | 3 -
libavcodec/Makefile| 1 +
libavcodec/dsputil.c | 299 +---
libavcodec/dsputil.h | 16 --
libavcodec/h264qpel_template.c | 1 +
Multiple inclusion makes no sense as it is only used in the 8-bit case.
---
libavcodec/dsputil.c | 1 +
libavcodec/dsputil_template.c | 4
2 files changed, 1 insertion(+), 4 deletions(-)
diff --git a/libavcodec/dsputil.c b/libavcodec/dsputil.c
index b81ba47..e41ea86 100644
---
The indirection makes no sense without multiple instantiation.
---
libavcodec/hpeldsp.c | 2 +
libavcodec/hpeldsp_template.c | 214 --
2 files changed, 104 insertions(+), 112 deletions(-)
diff --git a/libavcodec/hpeldsp.c b/libavcodec/hpeldsp.c
The functions are not used templatized.
---
libavcodec/dsputil.c | 38 ++
libavcodec/dsputil_template.c | 40
2 files changed, 38 insertions(+), 40 deletions(-)
diff --git a/libavcodec/dsputil.c
---
libavcodec/dsputil.c | 22 ---
libavcodec/dsputil.h | 6 ---
libavcodec/dsputil_template.c | 52 -
libavcodec/rv40dsp.c | 90 ---
4 files changed, 85 insertions(+), 85 deletions(-)
diff --git
The indirection makes no sense without multiple instantiation.
---
libavcodec/dsputil.c | 18 +---
libavcodec/dsputil_template.c | 223 +-
2 files changed, 114 insertions(+), 127 deletions(-)
diff --git a/libavcodec/dsputil.c
On 2014-03-19 16:25:10 +0200, Martin Storsjö wrote:
On Wed, 19 Mar 2014, Janne Grunau wrote:
Although it is allowed it does not make much sense since the immediate
is already 8bits.
---
gas-preprocessor.pl | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git
On 19/03/14 19:02, Diego Biurrun wrote:
I don't get your point. This patch is about unifying two duplicated
macros: CALL_2X_PIXELS and PIXELS16. The former name is more
descriptive, so I settled for it.
The end result is not improving readability at all.
Do as you prefer.
lu
On 19/03/14 19:13, Diego Biurrun wrote:
---
libavcodec/x86/cavsdsp.c | 1 +
libavcodec/x86/dsputil_init.c | 1 +
libavcodec/x86/dsputil_x86.h | 17 -
libavcodec/x86/fpel.h | 43
+++
libavcodec/x86/fpel_mmx.c | 2
On 19/03/14 19:13, Diego Biurrun wrote:
---
libavcodec/ac3enc_template.c | 4
libavcodec/h264_mb_template.c | 2 ++
libavcodec/h264_mc_template.c | 2 ++
libavcodec/h264idct_template.c | 1 +
libavcodec/hpel_template.c | 4
On 19/03/14 19:13, Diego Biurrun wrote:
---
libavcodec/ppc/dsputil_altivec.h | 5 -
libavcodec/ppc/h264qpel.c | 2 +-
libavcodec/ppc/hpeldsp_altivec.c | 2 +-
.../ppc/{dsputil_altivec.h = hpeldsp_altivec.h} |
On Wed, 19 Mar 2014, Ben Avison wrote:
Ben Avison (6):
truehd: add hand-scheduled ARM asm version of mlp_filter_channel.
truehd: break out part of rematrix_channels into platform-specific
callback.
truehd: add hand-scheduled ARM asm version of
ff_mlp_rematrix_channel.
truehd: tune VLC
On 19/03/14 18:56, Martin Storsjö wrote:
On Wed, 19 Mar 2014, Luca Barbato wrote:
On 19/03/14 18:24, Ben Avison wrote:
Verified with profiling that this doesn't have a measurable effect upon
overall performance.
---
libavcodec/mlpdec.c | 37 -
[Belatedly changing out of digest mode - hope this doesn't screw up
people's threading too much...]
--- a/libavcodec/mlpdec.c
+++ b/libavcodec/mlpdec.c
--- a/libavcodec/mlpdsp.c
+++ b/libavcodec/mlpdsp.c
@@ -57,9 +57,42 @@ static void mlp_filter_channel(int32_t *state, const int32_t
*coeff,
--- a/libavcodec/mlpdsp.c
+++ b/libavcodec/mlpdsp.c
@@ -89,10 +89,46 @@ void ff_mlp_rematrix_channel(int32_t *samples,
+
+int32_t ff_mlp_pack_output(int32_t lossless_check_data,
This function is not used outside of the file, so it can be made
static and the ff_ prefix can be removed.
It's used
On 19/03/14 18:34, Diego Biurrun wrote:
---
Amended to not break some long lines for enhanced readability.
Readability remains impaired.
+#define HPELDSP_AVG_PIXELS16(CPUEXT) \
+CALL_2X_PIXELS(put_no_rnd_pixels16_x2 ## CPUEXT,
ff_put_no_rnd_pixels8_x2 ## CPUEXT,
---
libavformat/mpeg.c | 6 +++---
libavformat/srtp.c | 10 +-
2 files changed, 8 insertions(+), 8 deletions(-)
diff --git a/libavformat/mpeg.c b/libavformat/mpeg.c
index 7430bb0..7996d84 100644
--- a/libavformat/mpeg.c
+++ b/libavformat/mpeg.c
@@ -336,7 +336,7 @@ static int
On 19/03/14 22:23, Tanja Batchelor wrote:
---
libavformat/mpeg.c | 6 +++---
libavformat/srtp.c | 10 +-
2 files changed, 8 insertions(+), 8 deletions(-)
I already gave you a review on irc.
diff --git a/libavformat/mpeg.c b/libavformat/mpeg.c
index 7430bb0..7996d84 100644
---
On Wednesday, March 19, 2014, Luca Barbato lu_z...@gentoo.org wrote:
On 19/03/14 22:23, Tanja Batchelor wrote:
---
libavformat/mpeg.c | 6 +++---
libavformat/srtp.c | 10 +-
2 files changed, 8 insertions(+), 8 deletions(-)
I already gave you a review on irc.
diff --git
Profiling results for overall audio decode and the mlp_filter_channel(_arm)
function in particular are as follows:
Before After
Mean StdDev Mean StdDev Confidence Change
6:2 total 380.4 22.0 370.8 17.087.4% +2.6% (insignificant)
Profiling results for overall decode and the output_data function in
particular are as follows:
Before After
Mean StdDev Mean StdDev Confidence Change
6:2 total 339.6 15.1 329.3 16.095.8% +3.1% (insignificant)
6:2 function 24.6
Verified with profiling that this doesn't have a measurable effect upon
overall performance.
---
libavcodec/mlpdec.c | 37 -
libavcodec/mlpdsp.c | 33 +
libavcodec/mlpdsp.h | 23 +++
3 files changed, 68
Profiling results for overall audio decode and the rematrix_channels function
in particular are as follows:
Before After
Mean StdDev Mean StdDev Confidence Change
6:2 total 370.8 17.0 348.8 20.199.9% +6.3%
6:2 function 46.4 8.4
Verified with profiling that this doesn't have a measurable effect upon
overall performance.
---
libavcodec/mlpdec.c | 40 +++-
libavcodec/mlpdsp.c | 38 ++
libavcodec/mlpdsp.h | 22 ++
3 files
An updated series taking into account comments to date.
Ben Avison (6):
truehd: add hand-scheduled ARM asm version of mlp_filter_channel.
truehd: break out part of rematrix_channels into platform-specific
callback.
truehd: add hand-scheduled ARM asm version of
Profiling on a Raspberry Pi revealed the best performance to correspond
with VLC_BITS = 5. Results for overall audio decode and the get_vlc2 function
in particular are as follows:
Before After
Mean StdDev Mean StdDev Confidence Change
6:2 total
---
My first encoder/decoder \o/
A sum up of the format is available here
http://www.mediatel.lu/workshop/graphic/2D_fileformat/h_aliaspix.html
GIMP is able to read and write PIX images, it's completely compatibile with the
ones
produced/read from this decoder/encoder.
Cheers,
Vittorio
---
tests/fate/image.mak| 3 +++
tests/ref/fate/aliaspix | 2 ++
2 files changed, 5 insertions(+)
create mode 100644 tests/ref/fate/aliaspix
diff --git a/tests/fate/image.mak b/tests/fate/image.mak
index 1357cbd..361b66d 100644
--- a/tests/fate/image.mak
+++ b/tests/fate/image.mak
@@ -1,3
On Wed, Mar 19, 2014 at 7:13 PM, Diego Biurrun di...@biurrun.de wrote:
---
libavcodec/x86/dsputil_x86.h| 14 --
libavcodec/x86/{hpeldsp_mmx.c = hpeldsp.h} | 43
++---
libavcodec/x86/hpeldsp_init.c | 2 +-
On Wed, Mar 19, 2014 at 7:13 PM, Diego Biurrun di...@biurrun.de wrote:
---
libavcodec/hpeldsp_template.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
OK
Vittorio
___
libav-devel mailing list
libav-devel@libav.org
On Wed, Mar 19, 2014 at 7:13 PM, Diego Biurrun di...@biurrun.de wrote:
Multiple inclusion makes no sense as it is only used in the 8-bit case.
---
libavcodec/dsputil.c | 1 +
libavcodec/dsputil_template.c | 4
2 files changed, 1 insertion(+), 4 deletions(-)
Ok
Vittorio
68 matches
Mail list logo