On 12/3/17, Carl Eugen Hoyos wrote:
> 2017-12-01 17:26 GMT+01:00 Paul B Mahol :
>> Signed-off-by: Paul B Mahol
>> ---
>> libavformat/Makefile | 1 +
>> libavformat/allformats.c | 1 +
>> libavformat/nspdec.c | 101
>> +++
>> 3 files changed
On 5 November 2017 at 23:35, Aurelien Jacobs wrote:
> Hello everyone,
>
> Long time no see !
>
> I'm glad to see ffmpeg still strong.
>
> I'm curently playing with bluetooth audio (A2DP) and I wanted to
> use lavc to do the encoding/decoding, so I added some codecs that
> I need to ffmpeg.
>
> He
Signed-off-by: Paul B Mahol
---
doc/filters.texi | 31 ++
libavfilter/Makefile | 1 +
libavfilter/allfilters.c | 1 +
libavfilter/vf_setparams.c | 82 ++
4 files changed, 115 insertions(+)
create mode 100644 libavfil
On 02/12/17 23:14, Aman Gupta wrote:
> From: Aman Gupta
>
> ---
> libavutil/Makefile | 2 ++
> libavutil/hwcontext.c| 4
> libavutil/hwcontext.h| 1 +
> libavutil/hwcontext_mediacodec.c | 50
>
> libavutil/hw
On 02/12/17 23:14, Aman Gupta wrote:
> From: Aman Gupta
>
> ---
> libavcodec/mediacodecdec_common.c | 7 ++-
> 1 file changed, 6 insertions(+), 1 deletion(-)
>
> diff --git a/libavcodec/mediacodecdec_common.c
> b/libavcodec/mediacodecdec_common.c
> index cb2f6ae5e5..e524ffe0e4 100644
> ---
I tried to implement what you requested
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
---
doc/demuxers.texi | 6
libavformat/hls.c | 106 ++
2 files changed, 98 insertions(+), 14 deletions(-)
diff --git a/doc/demuxers.texi b/doc/demuxers.texi
index 73dc0feec1..33643f966a 100644
--- a/doc/demuxers.texi
+++ b/doc/demuxers.te
fixed mem leak poined out by Steven
---
doc/demuxers.texi | 5 +
libavformat/hls.c | 304 --
2 files changed, 209 insertions(+), 100 deletions(-)
diff --git a/doc/demuxers.texi b/doc/demuxers.texi
index 73dc0feec1..634b122e10 100644
--- a/do
---
libavformat/hls.c | 209 +++---
1 file changed, 120 insertions(+), 89 deletions(-)
diff --git a/libavformat/hls.c b/libavformat/hls.c
index ab6ff187a6..3f83707c1f 100644
--- a/libavformat/hls.c
+++ b/libavformat/hls.c
@@ -314,6 +314,8 @@ static
v4: fixed the memleak properly
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
---
libavformat/hls.c | 208 +++---
1 file changed, 119 insertions(+), 89 deletions(-)
diff --git a/libavformat/hls.c b/libavformat/hls.c
index ab6ff187a6..3c2c720abe 100644
--- a/libavformat/hls.c
+++ b/libavformat/hls.c
@@ -314,6 +314,8 @@ static
---
doc/demuxers.texi | 6
libavformat/hls.c | 106 ++
2 files changed, 98 insertions(+), 14 deletions(-)
diff --git a/doc/demuxers.texi b/doc/demuxers.texi
index 73dc0feec1..33643f966a 100644
--- a/doc/demuxers.texi
+++ b/doc/demuxers.te
On Sat, Dec 2, 2017 at 11:01 AM, Michael Niedermayer
wrote:
> On Fri, Dec 01, 2017 at 04:09:15PM -0500, Tristan Matthews wrote:
>> Fix suggested by Luca Barbato.
>>
>> This was causing spurious EOFs when using -rtsp_transport udp, as
>> reported in https://bugzilla.libav.org/show_bug.cgi?id=1103
>
Hello,
In attach patch to add a checkasm test for the recently added threshold8
SIMD
also add AVX2 version
Checkasm result :
threshold8_c: 584.8
threshold8_sse4: 65.0
threshold8_avx2: 43.5
001 : create ff_threshold_init_func in order to simplify checkasm write
(more like previous checkasm test
Hello,
Maybe you can use a macro for byte and short version,
only few lines are different in each version
Martin
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
On 12/3/17, Martin Vignali wrote:
> Hello,
>
> In attach patch to add a checkasm test for the recently added threshold8
> SIMD
>
> also add AVX2 version
>
> Checkasm result :
> threshold8_c: 584.8
> threshold8_sse4: 65.0
> threshold8_avx2: 43.5
>
>
> 001 : create ff_threshold_init_func in order to
On 12/3/17, Martin Vignali wrote:
> Hello,
>
> Maybe you can use a macro for byte and short version,
> only few lines are different in each version
Sure, feel free to send patches.
I'm not very macro proficient.
___
ffmpeg-devel mailing list
ffmpeg-dev
>
> > 002 : Add checkasm test for vf_threshold
>
> Why is this GPL?
>
> Because i copy paste the header of vf_blend checkasm
Will change to LGPL.
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
2017-12-03 17:46 GMT+01:00 Paul B Mahol :
> On 12/3/17, Martin Vignali wrote:
> > Hello,
> >
> > Maybe you can use a macro for byte and short version,
> > only few lines are different in each version
>
> Sure, feel free to send patches.
>
> I'm not very macro proficient.
>
Ok, i will take a look
> >
>> > 002 : Add checkasm test for vf_threshold
>>
>> Why is this GPL?
>>
>> Because i copy paste the header of vf_blend checkasm
> Will change to LGPL.
>
> in fact most(all ?) checkasm test are under GPL licence.
and the checkasm exe, is under GPL too.
So i will let the current GPL licence
Mar
On 12/3/17, Martin Vignali wrote:
>> >
>>> > 002 : Add checkasm test for vf_threshold
>>>
>>> Why is this GPL?
>>>
>>> Because i copy paste the header of vf_blend checkasm
>> Will change to LGPL.
>>
>> in fact most(all ?) checkasm test are under GPL licence.
> and the checkasm exe, is under GPL to
On 12/02/2017 06:32 PM, James Almer wrote:
> On 12/2/2017 10:48 PM, Michael Niedermayer wrote:
>> On Sat, Dec 02, 2017 at 03:26:34PM -0300, James Almer wrote:
>>> On 12/2/2017 3:00 PM, Carl Eugen Hoyos wrote:
2017-12-02 18:51 GMT+01:00 James Almer :
> On 12/2/2017 2:40 PM, Carl Eugen Hoyos
> 2017-12-03 17:46 GMT+01:00 Paul B Mahol :
>
>> On 12/3/17, Martin Vignali wrote:
>> > Hello,
>> >
>> > Maybe you can use a macro for byte and short version,
>> > only few lines are different in each version
>>
>> Sure, feel free to send patches.
>>
>> I'm not very macro proficient.
>>
>
> Ok, i
2017-12-03 18:01 GMT+01:00 Paul B Mahol :
> On 12/3/17, Martin Vignali wrote:
> >> >
> >>> > 002 : Add checkasm test for vf_threshold
> >>>
> >>> Why is this GPL?
> >>>
> >>> Because i copy paste the header of vf_blend checkasm
> >> Will change to LGPL.
> >>
> >> in fact most(all ?) checkasm test
Hello,
Patch in attach add SIMD for threshold16
Checkasm result :
threshold16_c: 304.5
threshold16_sse4: 60.5
threshold16_avx2: 45.0
001 : modify threshold macro, and add threshold 16
002 : add checkasm test for threshold16
Martin
0001-avfilter-x86-vf_threshold-add-threshold16-SIMD-SSE4-.p
On 12/3/2017 3:09 PM, Martin Vignali wrote:
>> 2017-12-03 17:46 GMT+01:00 Paul B Mahol :
>>
>>> On 12/3/17, Martin Vignali wrote:
Hello,
Maybe you can use a macro for byte and short version,
only few lines are different in each version
>>>
>>> Sure, feel free to send patches.
>
> Can you post a disassembly of hflip_byte_c?
>
>
> in O1 : clang -S -O1 test_asm_gen.c
.section__TEXT,__text,regular,pure_instructions
.macosx_version_min 10, 12
.globl_hflip_byte_c
.p2align4, 0x90
_hflip_byte_c: ## @hflip_byte_c
.cfi_start
On 12/3/2017 3:55 PM, Martin Vignali wrote:
> in O2 or O3 : clang -S -O3 test_asm_gen.c
>
> If i correctly understand, same idea than paul's patch
> but processing two xmm in the main loop
>
> .section__TEXT,__text,regular,pure_instructions
> .macosx_version_min 10, 12
> .section
Michael Niedermayer (2017-12-01):
> Do you think this comment will fix any issue or improve anything ?
This comment had one goal: prevent somebody from pushing incorrect
changes before I had time to make my point.
> First i do not even know what you speak of exactly and as this is
I posted the e
On 12/3/17, Paul B Mahol wrote:
> Signed-off-by: Paul B Mahol
> ---
> libavfilter/hflip.h | 38
> libavfilter/vf_hflip.c | 133
> ++--
> libavfilter/x86/Makefile| 2 +
> libavfilter/x86/vf_hflip.asm| 98 +++
On 12/3/17, Paul B Mahol wrote:
> On 12/3/17, Paul B Mahol wrote:
>> Signed-off-by: Paul B Mahol
>> ---
>> libavfilter/hflip.h | 38
>> libavfilter/vf_hflip.c | 133
>> ++--
>> libavfilter/x86/Makefile| 2 +
>> lib
Signed-off-by: Paul B Mahol
---
libavfilter/hflip.h | 38
libavfilter/vf_hflip.c | 133 ++--
libavfilter/x86/Makefile| 2 +
libavfilter/x86/vf_hflip.asm| 98 +
libavfilter/x86/vf_hfli
>
> In any case, if clang or gcc can generate better code, then the hand
> written version needs to be optimized to be as fast or faster.
>
>
>
Quick test : pass checkasm (but probably only because width = 256)
hflip_byte_c: 26.4
hflip_byte_ssse3: 20.4
INIT_XMM ssse3
cglobal hflip_byte, 3, 5, 2,
On 12/3/17, Martin Vignali wrote:
>>
>> In any case, if clang or gcc can generate better code, then the hand
>> written version needs to be optimized to be as fast or faster.
>>
>>
>>
> Quick test : pass checkasm (but probably only because width = 256)
> hflip_byte_c: 26.4
> hflip_byte_ssse3: 20.4
2017-12-03 20:36 GMT+01:00 Paul B Mahol :
> On 12/3/17, Martin Vignali wrote:
> >>
> >> In any case, if clang or gcc can generate better code, then the hand
> >> written version needs to be optimized to be as fast or faster.
> >>
> >>
> >>
> > Quick test : pass checkasm (but probably only because
Maybe the problem come from the skip part :
+INIT_XMM ssse3
> +cglobal hflip_byte, 3, 5, 3, src, dst, w, x, v
> +movam0, [pb_flip_byte]
> +mov xq, 0
> +mov wd, dword wm
> +sub wq, 2 * mmsize
> +cmp wq, mmsize
> +jl .skip
> +
> +.loop0:
> +neg
On 12/3/17, Martin Vignali wrote:
> 2017-12-03 20:36 GMT+01:00 Paul B Mahol :
>
>> On 12/3/17, Martin Vignali wrote:
>> >>
>> >> In any case, if clang or gcc can generate better code, then the hand
>> >> written version needs to be optimized to be as fast or faster.
>> >>
>> >>
>> >>
>> > Quick t
On 12/3/17, Martin Vignali wrote:
> Maybe the problem come from the skip part :
>
> +INIT_XMM ssse3
>> +cglobal hflip_byte, 3, 5, 3, src, dst, w, x, v
>> +movam0, [pb_flip_byte]
>> +mov xq, 0
>> +mov wd, dword wm
>> +sub wq, 2 * mmsize
>> +cmp wq, mmsize
>>
On 01.12.2017 at 17:45 Michael Niedermayer wrote:
> AVDISCARD_NONREF is about frames which are not referenced by any other
> doesnt ff_hevc_frame_nb_refs() produce the number of references curently
> aka the other end of the reference arrows ?
Yes, that sounds right.
I've revised the patch to che
On 12/3/17, Paul B Mahol wrote:
> On 12/3/17, Martin Vignali wrote:
>> Maybe the problem come from the skip part :
>>
>> +INIT_XMM ssse3
>>> +cglobal hflip_byte, 3, 5, 3, src, dst, w, x, v
>>> +movam0, [pb_flip_byte]
>>> +mov xq, 0
>>> +mov wd, dword wm
>>> +sub wq
I modify the checkasm test, to test various width
if (check_func(s.flip_line[0], "hflip_%s", report_name)) {
for (i = 1; i < w; i++) {
call_ref(src, dst_ref, i);
call_new(src, dst_new, i);
if (memcmp(dst_ref, dst_new, WIDTH)) {
printf("FA
On 2017-12-03 19:30, Martin Vignali wrote:
> libavfilter/x86/vf_threshold.asm| 19 ++-
> libavfilter/x86/vf_threshold_init.c | 34 --
> 2 files changed, 34 insertions(+), 19 deletions(-)
>
> diff --git a/libavfilter/x86/vf_threshold.asm
> b/lib
2017-12-03 21:15 GMT+01:00 James Darnley :
> On 2017-12-03 19:30, Martin Vignali wrote:
> > libavfilter/x86/vf_threshold.asm| 19 ++-
> > libavfilter/x86/vf_threshold_init.c | 34 --
>
> > 2 files changed, 34 insertions(+), 19 deletions(-)
> >
Signed-off-by: Paul B Mahol
---
libavfilter/hflip.h | 38
libavfilter/vf_hflip.c | 133 ++--
libavfilter/x86/Makefile| 2 +
libavfilter/x86/vf_hflip.asm| 102 ++
libavfilter/x86/vf_hfl
On Sat, 2 Dec 2017, Marton Balint wrote:
On Fri, 24 Nov 2017, Marton Balint wrote:
The current edit unit cannot be reliably determined for the last packet of
a
video stream, because we can't query the start offset of the next edit unit
from the index. This caused missing timestamps for th
On Thu, 30 Nov 2017, John Stebbins wrote:
---
fftools/ffplay.c | 21 -
1 file changed, 16 insertions(+), 5 deletions(-)
diff --git a/fftools/ffplay.c b/fftools/ffplay.c
index 10a917194d..152d220cdb 100644
--- a/fftools/ffplay.c
+++ b/fftools/ffplay.c
@@ -198,6 +198,8 @@ type
Checkasm result (osx) for your last patch :
hflip_byte_c: 28.5
hflip_byte_ssse3: 29.0
hflip_short_c: 277.7
hflip_short_ssse3: 65.0
if you add a "cmp xq, wq" after the simd loop
you can be faster than c (clang), if width is multiple of mmsize*2
hflip_byte_c: 28.5
hflip_byte_ssse3: 27.5
see below
On 12/3/2017 5:50 PM, Paul B Mahol wrote:
> Signed-off-by: Paul B Mahol
> ---
> libavfilter/hflip.h | 38
> libavfilter/vf_hflip.c | 133
> ++--
> libavfilter/x86/Makefile| 2 +
> libavfilter/x86/vf_hflip.asm| 1
On Sat, Dec 02, 2017 at 09:50:26PM +0100, Mateusz wrote:
> After commit 3701d49 'error_resilience: remove avpriv_atomic usage'
> we have included windows.h in much more files and we should
> avoid conflicts with defines/function declarations.
>
> Signed-off-by: Mateusz Brzostek
> ---
> libavcode
On Sat, Dec 02, 2017 at 11:46:56AM -0600, Andrew D'Addesio wrote:
> Add av_sat_sub32 and av_sat_dsub32 as the subtraction analogues to
> av_sat_add32/av_sat_dadd32.
>
> Also clarify the formulas for dadd32/dsub32.
>
> Signed-off-by: Andrew D'Addesio
> ---
> libavutil/arm/intmath.h | 16
On Sat, Dec 02, 2017 at 11:46:57AM -0600, Andrew D'Addesio wrote:
> The relevant sections from the RFC are:
>
> Sec.6. Integer Wrap-Around in Inverse Gain Computation
> 32-bit integer overflow in Levinson recursion. Affects
> silk_is_lpc_stable().
>
> Sec.8. Cap on Band Energy
> NaN d
On Sat, Dec 02, 2017 at 11:32:16PM -0300, James Almer wrote:
> On 12/2/2017 10:48 PM, Michael Niedermayer wrote:
> > On Sat, Dec 02, 2017 at 03:26:34PM -0300, James Almer wrote:
[...]
> I'm surprised for that matter that the libvpx wrappers in old FFMpeg
> versions don't work with >= 1.4.0. I don'
On Sun, Dec 03, 2017 at 10:57:59AM +, Rostislav Pehlivanov wrote:
> On 5 November 2017 at 23:35, Aurelien Jacobs wrote:
>
> > Hello everyone,
> >
> > Long time no see !
> >
> > I'm glad to see ffmpeg still strong.
> >
> > I'm curently playing with bluetooth audio (A2DP) and I wanted to
> > us
2017-12-01 17:26 GMT+01:00 Paul B Mahol :
> +static int nsp_read_header(AVFormatContext *s)
> +{
> +int rate = 0, channels = 0;
> +uint32_t chunk, size;
> +AVStream *st;
> +int64_t pos;
> +
> +avio_skip(s->pb, 12);
I believe you could set the duration here for the
supported in
On Sun, Dec 3, 2017 at 10:42 PM, Michael Niedermayer
wrote:
> On Sat, Dec 02, 2017 at 11:32:16PM -0300, James Almer wrote:
>> On 12/2/2017 10:48 PM, Michael Niedermayer wrote:
>> > On Sat, Dec 02, 2017 at 03:26:34PM -0300, James Almer wrote:
> [...]
>
>> I'm surprised for that matter that the libv
On Sun, Dec 03, 2017 at 08:21:50PM +0100, Nicolas George wrote:
> Michael Niedermayer (2017-12-01):
> > Do you think this comment will fix any issue or improve anything ?
>
> This comment had one goal: prevent somebody from pushing incorrect
> changes before I had time to make my point.
>
> > Fir
W dniu 03.12.2017 o 22:26, Michael Niedermayer pisze:
> On Sat, Dec 02, 2017 at 09:50:26PM +0100, Mateusz wrote:
>> After commit 3701d49 'error_resilience: remove avpriv_atomic usage'
>> we have included windows.h in much more files and we should
>> avoid conflicts with defines/function declaration
Hi!
Attached patch fixes ticket #6717, files without sei can be produced
with remuxing and seeking, even if this is a (separate) bug, such
files exist in the wild.
Please comment, Carl Eugen
From c3ce0a75f13663958034a70fb4a982671532d269 Mon Sep 17 00:00:00 2001
From: Carl Eugen Hoyos
Date: Mon,
On Sun, Oct 29, 2017 at 01:39:06PM +0100, Michael Niedermayer wrote:
> On Sat, Oct 28, 2017 at 08:26:16PM +0200, Michael Niedermayer wrote:
> > On Thu, Oct 26, 2017 at 08:51:50PM -0700, Sasi Inguva wrote:
> > > On Tue, Oct 24, 2017 at 2:48 PM, Michael Niedermayer
> > > > > > wrote:
> > >
> > > >
On 04/12/17 00:09, Carl Eugen Hoyos wrote:
> Hi!
>
> Attached patch fixes ticket #6717, files without sei can be produced
> with remuxing and seeking, even if this is a (separate) bug, such
> files exist in the wild.
>
> Please comment, Carl Eugen
File are currently fixable by remuxing with:
-b
From: Aman Gupta
---
libavutil/Makefile | 1 +
libavutil/hwcontext.c| 4
libavutil/hwcontext.h| 1 +
libavutil/hwcontext_mediacodec.c | 50
libavutil/hwcontext_mediacodec.h | 36 +
From: Aman Gupta
---
libavcodec/mediacodecdec.c| 2 +-
libavcodec/mediacodecdec_common.c | 14 +-
2 files changed, 14 insertions(+), 2 deletions(-)
diff --git a/libavcodec/mediacodecdec.c b/libavcodec/mediacodecdec.c
index 39f5cbc045..eabf6d0648 100644
--- a/libavcodec/medi
2017-12-04 1:56 GMT+01:00 Mark Thompson :
> On 04/12/17 00:09, Carl Eugen Hoyos wrote:
>> Attached patch fixes ticket #6717, files without sei can be produced
>> with remuxing and seeking, even if this is a (separate) bug, such
>> files exist in the wild.
> File are currently fixable by remuxing
2017-12-04 1:22 GMT+01:00 Michael Niedermayer :
> On Sun, Oct 29, 2017 at 01:39:06PM +0100, Michael Niedermayer wrote:
>> On Sat, Oct 28, 2017 at 08:26:16PM +0200, Michael Niedermayer wrote:
>> > On Thu, Oct 26, 2017 at 08:51:50PM -0700, Sasi Inguva wrote:
>> > > On Tue, Oct 24, 2017 at 2:48 PM, Mi
2017-12-02 20:15 GMT+08:00 Sandeep Reddy :
> Hi,
>
> I applied the SCTE patch on hlsenc. But I am unable to find a way to
> intialize scte_interface of HLSContext .
>
> Please let me know ,how to intialize it
hlsenc initialize in hls_mux_init API, it is called by
hls_write_header, you can insert i
---
fftools/ffprobe.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/fftools/ffprobe.c b/fftools/ffprobe.c
index 0e7a771..6ddd81e 100644
--- a/fftools/ffprobe.c
+++ b/fftools/ffprobe.c
@@ -2280,6 +2280,7 @@ static av_always_inline int process_frame(WriterContext
*w,
break;
---
libavcodec/dvbsubdec.c | 10 +-
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/libavcodec/dvbsubdec.c b/libavcodec/dvbsubdec.c
index a657b1d..29997a2 100644
--- a/libavcodec/dvbsubdec.c
+++ b/libavcodec/dvbsubdec.c
@@ -1596,7 +1596,7 @@ static int dvbsub_display_end_segm
2017-12-02 10:03 GMT+08:00 Steven Liu :
> 2017-12-02 8:59 GMT+08:00 Michael Niedermayer :
>> On Fri, Dec 01, 2017 at 11:06:07AM +0100, Robert Nagy wrote:
>>> ---
>>> libavformat/hlsenc.c | 5 +
>>> 1 file changed, 1 insertion(+), 4 deletions(-)
>>>
>>> diff --git a/libavformat/hlsenc.c b/libav
On 11/30/2017 3:58 PM, Gyan Doshi wrote:
I added --disable-hwaccels during configure and noticed that all
autodetected hwaccel libs were still being linked - 6 in my case, which
makes for a more bloated binary than warranted.
Turns out the option only disables native components which make use
---
libavformat/dashdec.c | 112 --
1 file changed, 99 insertions(+), 13 deletions(-)
diff --git a/libavformat/dashdec.c b/libavformat/dashdec.c
index 3798649..d04bec0 100644
--- a/libavformat/dashdec.c
+++ b/libavformat/dashdec.c
@@ -148,6 +148,11
From 336be81bef86244a28b378b4c143ad8924c9e2d9 Mon Sep 17 00:00:00 2001
From: Jun Zhao
Date: Mon, 4 Dec 2017 12:50:34 +0800
Subject: [PATCH] avfilter/formats: fix wrong function name in error message
Use perdefined micro __FUNCTION__ rather than hard coding function name
to fix wrong function nam
> From: ffmpeg-devel [mailto:ffmpeg-devel-boun...@ffmpeg.org] On Behalf
> Of Mark Thompson
> Sent: Saturday, December 2, 2017 11:36 PM
> To: ffmpeg-devel@ffmpeg.org
> Subject: Re: [FFmpeg-devel] [PATCH 1/2] vaapi_h264: Add named options for
> setting profile and level
>
> On 01/12/17 05:24, Li, Zh
The Developer Documentation had instructions to
subscribe to the ffmpeg-cvslog email list. But that is
no longer accurate. For the purposes in this section --
review of patches, discussion of development issues --
ffmpeg_devel is the appropriate email list. Some developers
may want to monitor ffmpe
On 2 December 2017 at 17:36, Andrew D'Addesio wrote:
> Add av_sat_sub32 and av_sat_dsub32 as the subtraction analogues to
> av_sat_add32/av_sat_dadd32.
>
> Also clarify the formulas for dadd32/dsub32.
>
> Signed-off-by: Andrew D'Addesio
> ---
> libavutil/arm/intmath.h | 16
> l
On 2 December 2017 at 17:46, Andrew D'Addesio wrote:
> This decoder-side change, introduced in RFC 8251 (section 9), slightly
> improves the decoded quality of 16kbps speech in Hybrid Mode.
>
> Differences can be seen/heard in testvector05.bit, testvector06.bit,
> and testvector12.bit in the RFC
On 2 December 2017 at 17:46, Andrew D'Addesio wrote:
> When decoding to downmixed mono, don't put the channels out of phase,
> as they will cancel out and create audible artifacts. (See
> RFC 8251 sec. 10.)
>
> Signed-off-by: Andrew D'Addesio
> ---
> libavcodec/opus_pvq.c | 6 ++
> 1 file c
On 2 December 2017 at 17:36, Andrew D'Addesio wrote:
> The relevant sections from the RFC are:
>
> Sec.6. Integer Wrap-Around in Inverse Gain Computation
> 32-bit integer overflow in Levinson recursion. Affects
> silk_is_lpc_stable().
>
> Sec.8. Cap on Band Energy
> NaN due to large l
77 matches
Mail list logo
