On Fri, Jan 20, 2012 at 3:24 PM, Luca Barbato wrote:
> On 20/01/12 13:49, Martin Storsjö wrote:
>> On Fri, 20 Jan 2012, Luca Barbato wrote:
>>
>>> On 20/01/12 11:00, Martin Storsjö wrote:
Originally, sizeof(struct MOVIentry) was 48, after the reordering,
it is 40 in my build configuratio
On Tue, Jan 24, 2012 at 7:15 PM, Paul B Mahol wrote:
> On 1/24/12, Maans Rullgaard wrote:
>> Paul B Mahol writes:
>>
>>> On 1/24/12, Ronald S. Bultje wrote:
Hi,
On Tue, Jan 24, 2012 at 2:03 AM, Paul B Mahol wrote:
> Module: libav
> Branch: master
> Commit: 27ed027bcd
On 1/24/12, Maans Rullgaard wrote:
> Paul B Mahol writes:
>
>> On 1/24/12, Ronald S. Bultje wrote:
>>> Hi,
>>>
>>> On Tue, Jan 24, 2012 at 2:03 AM, Paul B Mahol wrote:
Module: libav
Branch: master
Commit: 27ed027bcdcb7445ab62d4870a8be3615ffc1e31
Author:Paul B Mahol
Overall speedup of HE-AAC decoding 2.3x on Cortex-A8, 1.2x on A9.
Signed-off-by: Mans Rullgard
---
Improved scheduling in a few functions, slightly faster overall.
---
libavcodec/arm/Makefile |4 +
libavcodec/arm/sbrdsp_init_arm.c | 70 +++
libavcodec/arm/sbrdsp_neon.S |
On 24/01/12 15:10, Justin Ruggles wrote:
> clear MDCT overlap buffer and reset previous window mode when seeking
Seems ok.
lu
--
Luca Barbato
Gentoo/linux
http://dev.gentoo.org/~lu_zero
___
libav-devel mailing list
libav-devel@libav.org
https://list
clear MDCT overlap buffer and reset previous window mode when seeking
---
libavcodec/vorbisdec.c | 12
1 files changed, 12 insertions(+), 0 deletions(-)
diff --git a/libavcodec/vorbisdec.c b/libavcodec/vorbisdec.c
index dcf25e6..541966e 100644
--- a/libavcodec/vorbisdec.c
+++ b/lib
On 24/01/12 13:25, Alex Converse wrote:
> I think these QMFs can be performed with substantially less
> computation (smaller transforms with similar amounts of pre/post
> processing). I can send interested parties the relevant paper (yes I
> know papers are a dime a dozen but this particular paper
2012/1/24 Måns Rullgård :
> Alex Converse writes:
>
>> I think these QMFs can be performed with substantially less
>> computation (smaller transforms with similar amounts of pre/post
>> processing). I can send interested parties the relevant paper (yes I
>> know papers are a dime a dozen but this
Jason Garrett-Glaser writes:
> I'll try my best to come up with some comments/questions despite
> incomplete NEON knowledge, hopefully these are at least useful.
>
>> + vld1.32 {q0}, [r0,:128]
>> + vld1.32 {q1}, [r1,:128]!
>> + vadd.f32 q0, q0,
Alex Converse writes:
> I think these QMFs can be performed with substantially less
> computation (smaller transforms with similar amounts of pre/post
> processing). I can send interested parties the relevant paper (yes I
> know papers are a dime a dozen but this particular paper seems
> promisin
I'll try my best to come up with some comments/questions despite
incomplete NEON knowledge, hopefully these are at least useful.
> + vld1.32 {q0}, [r0,:128]
> + vld1.32 {q1}, [r1,:128]!
> + vadd.f32 q0, q0, q1
> + vld1.32 {q2},
Paul B Mahol writes:
> On 1/24/12, Ronald S. Bultje wrote:
>> Hi,
>>
>> On Tue, Jan 24, 2012 at 2:03 AM, Paul B Mahol wrote:
>>> Module: libav
>>> Branch: master
>>> Commit: 27ed027bcdcb7445ab62d4870a8be3615ffc1e31
>>>
>>> Author:Paul B Mahol
>>> Committer: Kostya Shishkov
>>> Date:
On Tue, Jan 24, 2012 at 4:55 AM, Mans Rullgard wrote:
> This prepares for assembly optimisations by moving the most
> time-consuming loops to functions called through pointers
> in a new context.
>
> Signed-off-by: Mans Rullgard
> ---
> libavcodec/Makefile | 3 +-
> libavcodec/aacsbr.c
On Tue, Jan 24, 2012 at 4:55 AM, Mans Rullgard wrote:
> Signed-off-by: Mans Rullgard
> ---
> libavcodec/aacsbr.c | 8 +---
> libavcodec/aacsbrdata.h | 4 +++-
> libavcodec/sbr.h | 12 ++--
> 3 files changed, 14 insertions(+), 10 deletions(-)
>
OK
_
On Tue, Jan 24, 2012 at 4:55 AM, Mans Rullgard wrote:
>
> Overall speedup of AAC-HE decoding 2.3x on Cortex-A8, 1.2x on A9.
>
HE-AAC
> Signed-off-by: Mans Rullgard
> ---
> libavcodec/arm/Makefile | 4 +
> libavcodec/arm/sbrdsp_init_arm.c | 70 +++
> libavcodec/arm/sbrdsp_neon
On 01/19/2012 03:01 PM, Justin Ruggles wrote:
> According to unofficial documentation, the video rate is locked to the audio
> sample rate. This results in proper synchronization of audio and video
> timestamps.
One problem with this patch is that it changes the time_base after
read_header(), wh
On 01/24/2012 01:40 PM, Paul B Mahol wrote:
> On 1/24/12, Ronald S. Bultje wrote:
>> Hi,
>>
>> On Tue, Jan 24, 2012 at 2:03 AM, Paul B Mahol wrote:
>>> Module: libav
>>> Branch: master
>>> Commit: 27ed027bcdcb7445ab62d4870a8be3615ffc1e31
>>>
>>> Author:Paul B Mahol
>>> Committer: Kostya Shi
On Tue, Jan 24, 2012 at 01:40:48PM -0500, Justin Ruggles wrote:
> According to Multimedia Wiki documentation, the frame rate is not always 15,
> and the actual value can be found in the header.
> ---
> libavformat/westwood_vqa.c |9 +++--
> 1 files changed, 7 insertions(+), 2 deletions(-)
On 1/24/12, Ronald S. Bultje wrote:
> Hi,
>
> On Tue, Jan 24, 2012 at 2:03 AM, Paul B Mahol wrote:
>> Module: libav
>> Branch: master
>> Commit: 27ed027bcdcb7445ab62d4870a8be3615ffc1e31
>>
>> Author:Paul B Mahol
>> Committer: Kostya Shishkov
>> Date: Sat Jan 21 17:45:52 2012 +
>>
>
According to Multimedia Wiki documentation, the frame rate is not always 15,
and the actual value can be found in the header.
---
libavformat/westwood_vqa.c |9 +++--
1 files changed, 7 insertions(+), 2 deletions(-)
diff --git a/libavformat/westwood_vqa.c b/libavformat/westwood_vqa.c
inde
On Tue, Jan 24, 2012 at 01:11:05PM -0500, Justin Ruggles wrote:
> On 01/24/2012 12:18 PM, Kostya Shishkov wrote:
>
> > On Tue, Jan 24, 2012 at 11:57:15AM -0500, Justin Ruggles wrote:
> >> The format has no coded timestamps and must use packet durations to
> >> generate
> >> them.
> >> ---
> >> l
On 01/24/2012 12:18 PM, Kostya Shishkov wrote:
> On Tue, Jan 24, 2012 at 11:57:15AM -0500, Justin Ruggles wrote:
>> The format has no coded timestamps and must use packet durations to generate
>> them.
>> ---
>> libavformat/westwood_vqa.c |2 ++
>> 1 files changed, 2 insertions(+), 0 deletion
On Tue, Jan 24, 2012 at 11:57:28AM -0500, Justin Ruggles wrote:
> There is only one stream, so it will always be index 0.
> This also allows for removal of the private context.
> ---
> libavformat/westwood_aud.c | 11 +--
> 1 files changed, 1 insertions(+), 10 deletions(-)
>
> diff --gi
On Tue, Jan 24, 2012 at 11:57:27AM -0500, Justin Ruggles wrote:
> also allows for removing some unused context fields
> ---
> libavformat/westwood_aud.c | 10 +-
> 1 files changed, 1 insertions(+), 9 deletions(-)
>
> diff --git a/libavformat/westwood_aud.c b/libavformat/westwood_aud.c
>
On Tue, Jan 24, 2012 at 11:57:26AM -0500, Justin Ruggles wrote:
> also allows for removing some unused context fields
> ---
> libavformat/westwood_aud.c | 51 ---
> 1 files changed, 28 insertions(+), 23 deletions(-)
lgtm
__
On Tue, Jan 24, 2012 at 11:57:25AM -0500, Justin Ruggles wrote:
> Time stamps are not read from the bitstream, so no wrapping is needed.
ever heard about cargo culting in programming?
> ---
> libavformat/westwood_aud.c |2 +-
> 1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git
On Tue, Jan 24, 2012 at 11:57:24AM -0500, Justin Ruggles wrote:
> ---
> libavformat/westwood_aud.c | 16
> 1 files changed, 8 insertions(+), 8 deletions(-)
>
> diff --git a/libavformat/westwood_aud.c b/libavformat/westwood_aud.c
> index 652045f..21d16f7 100644
> --- a/libavform
On Tue, Jan 24, 2012 at 11:57:23AM -0500, Justin Ruggles wrote:
> Fixes decoding of the handful of "death scream" samples from
> "Command and Conquer" that use SND1. e.g.
> http://samples.libav.org/game-formats/westwood-aud/nuyell1.aud
> http://samples.libav.org/game-formats/westwood-aud/nuyell7.au
On Tue, Jan 24, 2012 at 11:57:22AM -0500, Justin Ruggles wrote:
> Stereo ADPCM IMA WS is planar for VQA version 3 and 2-sample interleaved for
> VQA version 2.
> ---
> libavcodec/adpcm.c | 30 +-
> libavformat/westwood_vqa.c |6 ++
> 2 files changed, 3
On Tue, Jan 24, 2012 at 11:57:21AM -0500, Justin Ruggles wrote:
> The stereo layout and extradata is significantly different from that in
> Westwood IMA ADPCM, so a separate codec_id is warranted.
> ---
> libavcodec/Makefile|1 +
> libavcodec/adpcm.c |7 +--
> libavcodec/allcod
On Tue, Jan 24, 2012 at 11:57:20AM -0500, Justin Ruggles wrote:
> ---
> libavformat/westwood_vqa.c |4
> 1 files changed, 0 insertions(+), 4 deletions(-)
>
> diff --git a/libavformat/westwood_vqa.c b/libavformat/westwood_vqa.c
> index c74090b..b2ffd17 100644
> --- a/libavformat/westwood_
On Tue, Jan 24, 2012 at 11:57:19AM -0500, Justin Ruggles wrote:
> Fix setting of bit_rate and bits_per_coded_sample.
> Ask for sample for unsupported VQA versions.
> ---
> libavformat/westwood_vqa.c | 46 ---
> 1 files changed, 30 insertions(+), 16 deletio
On Tue, Jan 24, 2012 at 11:57:18AM -0500, Justin Ruggles wrote:
> According to Multimedia Wiki documentation, the frame rate is not always 15,
> and the actual value can be found in the header.
> ---
> libavformat/westwood_vqa.c |3 +--
> 1 files changed, 1 insertions(+), 2 deletions(-)
>
> d
On Tue, Jan 24, 2012 at 11:57:17AM -0500, Justin Ruggles wrote:
> Also remove unused context field, audio_frame_counter.
> ---
> libavformat/westwood_vqa.c |8 +++-
> 1 files changed, 3 insertions(+), 5 deletions(-)
lgtm
___
libav-devel mailing
On Tue, Jan 24, 2012 at 11:57:16AM -0500, Justin Ruggles wrote:
> ---
> libavformat/westwood_vqa.c |3 ++-
> 1 files changed, 2 insertions(+), 1 deletions(-)
>
> diff --git a/libavformat/westwood_vqa.c b/libavformat/westwood_vqa.c
> index 20462d9..fef6877 100644
> --- a/libavformat/westwood_v
On Tue, Jan 24, 2012 at 11:57:15AM -0500, Justin Ruggles wrote:
> The format has no coded timestamps and must use packet durations to generate
> them.
> ---
> libavformat/westwood_vqa.c |2 ++
> 1 files changed, 2 insertions(+), 0 deletions(-)
>
> diff --git a/libavformat/westwood_vqa.c b/lib
There is only one stream, so it will always be index 0.
This also allows for removal of the private context.
---
libavformat/westwood_aud.c | 11 +--
1 files changed, 1 insertions(+), 10 deletions(-)
diff --git a/libavformat/westwood_aud.c b/libavformat/westwood_aud.c
index 2195acc..f08
also allows for removing some unused context fields
---
libavformat/westwood_aud.c | 10 +-
1 files changed, 1 insertions(+), 9 deletions(-)
diff --git a/libavformat/westwood_aud.c b/libavformat/westwood_aud.c
index 79f2198..2195acc 100644
--- a/libavformat/westwood_aud.c
+++ b/libavfor
---
libavformat/westwood_aud.c | 16
1 files changed, 8 insertions(+), 8 deletions(-)
diff --git a/libavformat/westwood_aud.c b/libavformat/westwood_aud.c
index 652045f..21d16f7 100644
--- a/libavformat/westwood_aud.c
+++ b/libavformat/westwood_aud.c
@@ -169,14 +169,14 @@ stati
also allows for removing some unused context fields
---
libavformat/westwood_aud.c | 51 ---
1 files changed, 28 insertions(+), 23 deletions(-)
diff --git a/libavformat/westwood_aud.c b/libavformat/westwood_aud.c
index 27dc3fa..79f2198 100644
--- a/libavf
Time stamps are not read from the bitstream, so no wrapping is needed.
---
libavformat/westwood_aud.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/libavformat/westwood_aud.c b/libavformat/westwood_aud.c
index 21d16f7..27dc3fa 100644
--- a/libavformat/westwood_aud.c
+++
The format has no coded timestamps and must use packet durations to generate
them.
---
libavformat/westwood_vqa.c |2 ++
1 files changed, 2 insertions(+), 0 deletions(-)
diff --git a/libavformat/westwood_vqa.c b/libavformat/westwood_vqa.c
index 2eff4cc..20462d9 100644
--- a/libavformat/westwo
The stereo layout and extradata is significantly different from that in
Westwood IMA ADPCM, so a separate codec_id is warranted.
---
libavcodec/Makefile|1 +
libavcodec/adpcm.c |7 +--
libavcodec/allcodecs.c |1 +
libavcodec/avcodec.h |1 +
libavcodec/utils.c |
Stereo ADPCM IMA WS is planar for VQA version 3 and 2-sample interleaved for
VQA version 2.
---
libavcodec/adpcm.c | 30 +-
libavformat/westwood_vqa.c |6 ++
2 files changed, 35 insertions(+), 1 deletions(-)
diff --git a/libavcodec/adpcm.c b/libavcode
Fixes decoding of the handful of "death scream" samples from
"Command and Conquer" that use SND1. e.g.
http://samples.libav.org/game-formats/westwood-aud/nuyell1.aud
http://samples.libav.org/game-formats/westwood-aud/nuyell7.aud
---
libavformat/westwood_aud.c | 32 ++-
---
libavformat/westwood_vqa.c |4
1 files changed, 0 insertions(+), 4 deletions(-)
diff --git a/libavformat/westwood_vqa.c b/libavformat/westwood_vqa.c
index c74090b..b2ffd17 100644
--- a/libavformat/westwood_vqa.c
+++ b/libavformat/westwood_vqa.c
@@ -54,10 +54,7 @@
#define VQA_PREAMBL
Fix setting of bit_rate and bits_per_coded_sample.
Ask for sample for unsupported VQA versions.
---
libavformat/westwood_vqa.c | 46 ---
1 files changed, 30 insertions(+), 16 deletions(-)
diff --git a/libavformat/westwood_vqa.c b/libavformat/westwood_vqa.
According to Multimedia Wiki documentation, the frame rate is not always 15,
and the actual value can be found in the header.
---
libavformat/westwood_vqa.c |3 +--
1 files changed, 1 insertions(+), 2 deletions(-)
diff --git a/libavformat/westwood_vqa.c b/libavformat/westwood_vqa.c
index 9a48
Also remove unused context field, audio_frame_counter.
---
libavformat/westwood_vqa.c |8 +++-
1 files changed, 3 insertions(+), 5 deletions(-)
diff --git a/libavformat/westwood_vqa.c b/libavformat/westwood_vqa.c
index fef6877..9a48263 100644
--- a/libavformat/westwood_vqa.c
+++ b/libavfo
---
libavformat/westwood_vqa.c |3 ++-
1 files changed, 2 insertions(+), 1 deletions(-)
diff --git a/libavformat/westwood_vqa.c b/libavformat/westwood_vqa.c
index 20462d9..fef6877 100644
--- a/libavformat/westwood_vqa.c
+++ b/libavformat/westwood_vqa.c
@@ -122,7 +122,6 @@ static int wsvqa_rea
This patch set is for Westwood AUD and VQA demuxers. My basic motivation
was to fix timestamps, but I also ran into a few other things that needed
fixing or implementing.
Thanks,
Justin
Justin Ruggles (14):
vqa: set stream start_time to 0.
vqa: use 1/sample_rate as the audio stream time base
Patch set looks good to me.
MvH
Benjamin Larsson
___
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailman/listinfo/libav-devel
On 2012-01-24 06:55:17 +0100, Anton Khirnov wrote:
> Error recognition flags can and should be set directly with AVOptions.
> ---
> avplay.c |3 ---
> 1 files changed, 0 insertions(+), 3 deletions(-)
>
> diff --git a/avplay.c b/avplay.c
> index 432afc1..27ac01a 100644
> --- a/avplay.c
> +++ b
On 2012-01-24 06:55:18 +0100, Anton Khirnov wrote:
> It has been deprecated only recently.
> ---
> libavcodec/version.h |4 ++--
> 1 files changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/libavcodec/version.h b/libavcodec/version.h
> index 77e1682..1de790e 100644
> --- a/libavcodec/v
On 2012-01-24 06:55:19 +0100, Anton Khirnov wrote:
> It has been deprecated only recently.
> ---
> libavformat/version.h |2 +-
> 1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git a/libavformat/version.h b/libavformat/version.h
> index cd774fb..d50b2a5 100644
> --- a/libavformat
On 2012-01-24 06:55:20 +0100, Anton Khirnov wrote:
> Although it has been deprecated for a long time, its intended
> replacement (request_channel_layout) is not actually used anywhere, so
> request_channels is currently the only way to access that functionality.
> ---
> libavcodec/version.h |2
On 2012-01-24 06:55:16 +0100, Anton Khirnov wrote:
> ---
> configure|2 -
> libavdevice/Makefile |1 -
> libavdevice/alldevices.c |3 -
> libavdevice/avdevice.h |4 -
> libavdevice/v4l.c| 366
> --
> 5 file
On 2012-01-24 11:18:34 +0100, Diego Biurrun wrote:
> On Mon, Jan 23, 2012 at 04:13:39PM +, Sean McGovern wrote:
> >
> > Looking at this failure on the valgrind FATE runs, should
> > mxfdec.c:1060 be an av_realloc() instead of an av_mallocz()?
>
> Have you tested the change?
already fixed and
Overall speedup of AAC-HE decoding 2.3x on Cortex-A8, 1.2x on A9.
Signed-off-by: Mans Rullgard
---
libavcodec/arm/Makefile |4 +
libavcodec/arm/sbrdsp_init_arm.c | 70 +++
libavcodec/arm/sbrdsp_neon.S | 406 ++
libavcodec/sbrdsp.c
This prepares for assembly optimisations by moving the most
time-consuming loops to functions called through pointers
in a new context.
Signed-off-by: Mans Rullgard
---
libavcodec/Makefile |3 +-
libavcodec/aacsbr.c | 183 ++---
libavcodec/aacsbrdata.
Signed-off-by: Mans Rullgard
---
libavcodec/aacsbr.c |8 +---
libavcodec/aacsbrdata.h |4 +++-
libavcodec/sbr.h| 12 ++--
3 files changed, 14 insertions(+), 10 deletions(-)
diff --git a/libavcodec/aacsbr.c b/libavcodec/aacsbr.c
index 01601b0..3a30fca 100644
---
Anton Khirnov writes:
> ---
> configure|2 -
> libavdevice/Makefile |1 -
> libavdevice/alldevices.c |3 -
> libavdevice/avdevice.h |4 -
> libavdevice/v4l.c| 366
> --
> 5 files changed, 0 insertions(+),
Hi,
On Tue, Jan 24, 2012 at 7:21 PM, Ronald S. Bultje wrote:
> Also implement sse2/ssse3/avx versions.
> ---
> libswscale/x86/input.asm | 299
> +
> libswscale/x86/swscale_mmx.c | 48 ---
> libswscale/x86/swscale_template.c | 159 +--
Also implement sse2/ssse3/avx versions.
---
libswscale/x86/input.asm | 299 +
libswscale/x86/swscale_mmx.c | 48 ---
libswscale/x86/swscale_template.c | 159 +---
3 files changed, 328 insertions(+), 178 deletions(-)
diff --
On Mon, Jan 23, 2012 at 04:13:39PM +, Sean McGovern wrote:
>
> Looking at this failure on the valgrind FATE runs, should
> mxfdec.c:1060 be an av_realloc() instead of an av_mallocz()?
Have you tested the change?
Diego
___
libav-devel mailing list
l
65 matches
Mail list logo
