Hi Christophe,
the fix looks weird to me. There is something else underlying.
Mickaël
2015-09-23 16:53 GMT+02:00 Ronald S. Bultje :
> Hi,
>
> On Wed, Sep 23, 2015 at 10:33 AM, Christophe Gisquet <
> christophe.gisq...@gmail.com> wrote:
>
> > Hi,
> >
> > under highly-threaded loads, parallel dec
Looks better to me.
Mickaël
Le mardi 17 février 2015, Christophe Gisquet
a écrit :
> 2015-02-17 8:28 GMT+01:00 Mickaël Raulet >:
> > It seems to me that you are affecting 8 when it is avx2 instead of 11.
> > Shouldn't it be the opposite? At least this what the commit
It seems to me that you are affecting 8 when it is avx2 instead of 11.
Shouldn't it be the opposite? At least this what the commit message says.
Mickaël
Le mardi 17 février 2015, Christophe Gisquet
a écrit :
> Hi,
>
> here's an updated patch, which takes into account the fact avx2
> sometimes u
ht be wrong)
nb_output <= s->sps->temporal_layer[s->sps->max_sub_layers - 1].
num_reorder_pics + s->interlaced +1
2015-02-08 20:09 GMT+01:00 Carl Eugen Hoyos :
> Kacper Michajłow gmail.com> writes:
>
> > 2015-02-08 10:48 GMT+01:00 Carl Eugen Hoyos:
> >
That's why we are delaying SAO.
Le jeudi 5 février 2015, James Almer a écrit :
> On 05/02/15 2:34 PM, Christophe Gisquet wrote:
> > 2015-02-05 18:28 GMT+01:00 James Almer
> >:
> >> On 05/02/15 10:22 AM, Mickaël Raulet wrote:
> >>> More coming soon for e
As we can consider, we won't have 4k interlaced content, copying a field
into a frame should be ok. This is what has been done in this
implementation.
Commit hash from ffmpeg/openhevc:
6b93a7a175fb500d1f5d4d671b2fab73798ca7b6
Comments welcome!
Mickaël
___
Michael,
Please find some commits that can be cherry picked from
https://github.com/OpenHEVC/FFmpeg/commits/ffmpeg_patch
Optimized deblocking filter (8bits only)
1b9ee47d2f43b0a029a9468233626102eb1473b8
Optimzed transform functions (4x4, 8x8, transform add 8bits only)
b153f55935969c794de4640f8d3
PPS_A_qualcomm_7, yes I have sometimes an issue with this sequence.
Patch LGTM otherwise, it does not change the behaviour of the previous
implementation.
Mickaël
2015-02-05 13:36 GMT+01:00 Christophe Gisquet
:
> Hi,
>
> 2015-02-05 13:17 GMT+01:00 Mickaël Raulet :
> >
on one sequence? PPS7_xxx ? Right?
Mickaël
2015-02-05 11:13 GMT+01:00 Christophe Gisquet
:
> 2015-02-05 10:13 GMT+01:00 Christophe Gisquet <
> christophe.gisq...@gmail.com>:
> > The patch breaks make fate-hevc THREADS=3, so needs more thought.
>
> Compilation issue, running make clean first pa
2015-02-05 8:17 GMT+01:00 Christophe Gisquet :
> Hi,
>
> 2015-02-05 7:29 GMT+01:00 Christophe Gisquet >:
> > We were previously reference-counting the sao-buffer. Should we do
> > that for sao_pixel_buffer_[hv], then?
>
> Something like the attached patch.
>
> Note: I'm probably overallocating co
LGTM.
Mickael
2015-02-04 13:51 GMT+01:00 Christophe Gisquet
:
> Hi,
>
> 2015-02-04 4:55 GMT+01:00 James Almer :
>
> > -DECLARE_ALIGNED(16, const xmm_reg, ff_pw_1)= {
> 0x0001000100010001ULL, 0x0001000100010001ULL };
> > -DECLARE_ALIGNED(16, const xmm_reg, ff_pw_2)= {
> 0x0002000200020
LGTM
Mickael
2015-02-04 13:39 GMT+01:00 Christophe Gisquet
:
> Hi,
>
> 2015-02-04 4:55 GMT+01:00 James Almer :
> > Original x86 intrinsics code and initial yasm port by Pierre-Edouard
> Lepere.
> > Refactoring and optimizations by James Almer.
>
> Add your own copyright to this file then.
>
> >
lgtm.
Mickael
2015-02-04 4:55 GMT+01:00 James Almer :
> As with sao_band_filter, pass instead the two variables from the struct
> needed in the function.
> This simplifies writing asm optimized versions.
>
> Signed-off-by: James Almer
> ---
> libavcodec/hevc_filter.c | 4 +++-
> libavcode
ok.
2015-02-04 8:07 GMT+01:00 Christophe Gisquet :
> Hi,
>
> 2015-02-04 4:55 GMT+01:00 James Almer :
> [...]
>
> Ok, no need to resend a refreshed patch if patch 2/6 changes.
>
> --
> Christophe
> ___
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.or
OK too.
2015-02-04 8:04 GMT+01:00 Christophe Gisquet :
> Hi,
>
> 2015-02-04 4:55 GMT+01:00 James Almer :
> > +int a_stride, b_stride;
> > +int src_offset = 0;
> > +int dst_offset = 0;
>
> Could maybe use ptrdiff_t type, like the other strides?
>
> With or without, ok.
>
> --
> Christo
Ok too.
2015-02-04 8:09 GMT+01:00 Christophe Gisquet :
> Hi,
>
> 2015-02-04 6:13 GMT+01:00 Michael Niedermayer :
> > On Wed, Feb 04, 2015 at 12:55:48AM -0300, James Almer wrote:
> >> From: Seppo Tomperi
> >>
> >> ---
> >> libavcodec/hevc_filter.c | 3 ++-
> >> libavcodec/hevcdsp.c
it works now for me.
Thanks,
Mickaël
Le 3 févr. 2015 à 15:28, Mickaël Raulet a écrit :
> I will check it this evening.
>
> Mickaël
>
> 2015-02-03 15:15 GMT+01:00 Christophe Gisquet :
> 2015-02-03 12:57 GMT+01:00 Christophe Gisquet :
> > Actually, 940300945 does nee
I will check it this evening.
Mickaël
2015-02-03 15:15 GMT+01:00 Christophe Gisquet
:
> 2015-02-03 12:57 GMT+01:00 Christophe Gisquet <
> christophe.gisq...@gmail.com>:
> > Actually, 940300945 does need to be reverted for the patch to work, as
> > Mickael stated. It miscompiles hevc_mc.asm, mor
PL Lepere is the original author and I did some improvements on top of it.
Mickael
2015-02-02 18:11 GMT+01:00 Christophe Gisquet
:
> Hi,
>
> 2015-02-02 17:16 GMT+01:00 Mickaël Raulet :
> >
> https://github.com/OpenHEVC/FFmpeg/commit/940300945995c20f7583394ebe6907e72829b4a
&
Hi christophe,
I was willing to send avx2 mc patch, but I need to test it on top of ffmpeg
head. Last time I tried I had to revert
940300945995c20f7583394ebe6907e72829b4a.
https://github.com/OpenHEVC/FFmpeg/commit/940300945995c20f7583394ebe6907e72829b4a
.
Here is the avx2 implementation if someo
I did it first :)
2015-02-02 16:47 GMT+01:00 Michael Niedermayer :
> On Mon, Feb 02, 2015 at 04:11:33PM +0100, Mickaël Raulet wrote:
> > Hi here is a commit that support monochrome sequences!
> >
> >
> https://github.com/OpenHEVC/FFmpeg/commit/8e50557707d2ec11
Hi here is a commit that support monochrome sequences!
https://github.com/OpenHEVC/FFmpeg/commit/8e50557707d2ec11ccad657470b2e140f314348e
Commit hash: 8e50557707d2ec11ccad657470b2e140f314348e
Mickael
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.or
something like this.
https://github.com/OpenHEVC/openHEVC/commit/5d9f79edef2c11b915bdac3a025b59a32082f409
Mickaël
2015-02-01 20:37 GMT+01:00 Christophe Gisquet
:
> Hi,
>
> 2015-02-01 20:00 GMT+01:00 James Almer :
> > From: Christophe Gisquet
> >
> > For band filter, source and destination are
LGTM.
Mickaël
Le samedi 31 janvier 2015, Christophe Gisquet
a écrit :
> Hi,
>
> 2015-01-30 19:50 GMT+01:00 James Almer >:
> > +%macro HEVC_SAO_BAND_FILTER_COMPUTE 3
> > +psraw %2, %3, %1-5
> > +pcmpeqw m10, %2, m0
> > +pcmpeqw m11, %2, m1
> > +pcmpe
Looks good to me. Looks like an evil bitstream.
Mickaël
> Le 11 nov. 2014 à 03:54, Timothy Gu a écrit :
>
> Hi,
>
> On Monday, November 10, 2014, WeiChangjiang
> wrote:
>
>> Hi guys,
>> The attached is a very minor fix for hevc decoder. For big negative
>> motion vector, namely, mv_y, the
The best place is ffmpeg, it has a lot of testing platform we don t have.
Mickael
Le mardi 23 septembre 2014, James Almer a écrit :
> On 22/09/14 11:31 PM, Zhenan Lin wrote:
> > Hi, all! Our team is working on SIMD Optimization on PowerPC for HEVC
> > Decoder. However, we have no idea about whe
this commit might help to solve the issue with SSE4
https://github.com/OpenHEVC/FFmpeg/commit/df8ebe304df453f26c28ff8f11d607f49b90a4c2
Mickaël
Le 24 août 2014 à 11:52, Michael Niedermayer a écrit :
> On Sun, Aug 24, 2014 at 08:46:31AM +, Christophe Gisquet wrote:
>> ---
>> libavcodec/x86/he
For avx2 I have some to push to the trunk, I did merge it yesterday with
all recent changes. But I don t remember what those tables looks like.
For 10 and 12bits, ssse3 should slow down the decoding since it uses 4 more
instructions in the loop.
Le samedi 23 août 2014, Christophe Gisquet a
écrit
For 10bits and 12bits, they should stay sse4 as well because of packusdw. You
need some instructions to convert it to ssse3 see below
static av_always_inline __m128i _MM_PACKUS_EPI32( __m128i a, __m128i b )
{
a = _mm_slli_epi32 (a, 16);
a = _mm_srai_epi32 (a, 16);
b = _mm_slli_epi
for the whole patchset.
Mickaël
Le 22 août 2014 à 13:25, Michael Niedermayer a écrit :
> On Fri, Aug 22, 2014 at 11:40:17AM +0200, Mickaël Raulet wrote:
>> Patch okay.
>
> patch applied
>
> just to make sure i dont misunderstand, that "okay" was just for thi
Patch okay.
Mickaël
Le 4 août 2014 à 10:31, Christophe Gisquet a
écrit :
> Hi,
>
> 2014-08-02 14:48 GMT+02:00 Michael Niedermayer :
>> seems to fail with
>> libavcodec/x86/hevc_mc.asm:1258: error: (add:2) cannot reference symbol
>> `MAX_PB_SIZE' in preprocessor
>
> I forgot the initial patch
Ok.
Mickael
Le mercredi 20 août 2014, Timothy Gu a écrit :
> On Tue, Aug 19, 2014 at 6:49 PM, Michael Niedermayer > wrote:
> > Fixes memleak
> > Fixes CID1231989
> >
> > Signed-off-by: Michael Niedermayer >
> > ---
> > libavcodec/hevc_ps.c |3 ++-
> > 1 file changed, 2 insertions(+), 1 de
Patch ok
Mickael
Le mercredi 20 août 2014, James Almer a écrit :
> ~15% faster than sse2
>
> Signed-off-by: James Almer >
> ---
> libavcodec/x86/hevc_res_add.asm | 15 +++
> libavcodec/x86/hevcdsp.h| 4
> libavcodec/x86/hevcdsp_init.c | 4
> 3 files changed, 1
Hi
Le 10 août 2014 à 15:48, Michael Niedermayer a écrit :
> On Sun, Aug 10, 2014 at 03:16:23PM +0200, Christophe Gisquet wrote:
>> Hi,
>>
>> 2014-08-10 14:42 GMT+02:00 Ronald S. Bultje :
>>> Are we using the checked bitstream reader? If we are, we're fine already...
>>
>> I think we are. On the
Hi
Le 10 août 2014 à 15:16, Christophe Gisquet a
écrit :
> Hi,
>
> 2014-08-10 14:42 GMT+02:00 Ronald S. Bultje :
>> Are we using the checked bitstream reader? If we are, we're fine already...
>
> I think we are. On the other hand, it seems the top caller,
> ff_hevc_decode_nal_vps, is never ch
Patch ok
Mickael
Le mercredi 6 août 2014, Christophe Gisquet
a écrit :
> Hi,
>
> this patch is mostly cosmetical. I don't like seeing arrays passed to
> dsp functions being of a type whose length may not be fixed, though
> it's a small matter here.
>
> --
> Christophe
>
Patch ok.
Mickael
Le lundi 4 août 2014, James Almer a écrit :
> Signed-off-by: James Almer >
> ---
> libavcodec/x86/hevc_mc.asm | 3 +--
> 1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/libavcodec/x86/hevc_mc.asm b/libavcodec/x86/hevc_mc.asm
> index fc78062..a16b0ab 100644
> -
Le 30 juil. 2014 à 16:35, Ronald S. Bultje a écrit :
> Hi!
>
> On Wed, Jul 30, 2014 at 9:33 AM, Pierre Edouard Lepere <
> pierre-edouard.lep...@insa-rennes.fr> wrote:
>
>> Here's a patch adding ASM transform_add functions for HEVC.
>
>
> Yay! I'll try to review soon. Do you have rough perfor
Hi Christophe
>> hevc.c |9 +
>> 1 file changed, 5 insertions(+), 4 deletions(-)
>> 2445ba15d38b2472f8f1b24aa75e63c089971480
>> 0012-hevc_mc-reduce-stride-for-bidir-temp-buffers.patch
>> From 126adf820bc54c2d00f794629595ad6310fbfc37 Mon Sep 17 00:00:00 2001
>> From: Christophe Gisquet
Hi,
> Signed-off-by: James Almer
> ---
> Didn't realize that the newest HEVC fate tests were 12 bit, so i didn't
> bother adding what i thought i couldn't feasibly test.
> And as i mentioned in the previous patch, adding this is trivial.
>
> libavcodec/x86/hevc_idct.asm | 16
Hi,
> Only 8-bit and 10-bit idct_dc() functions are included (adding others should
> be trivial).
>
> Benchmarks on an Intel Core i5-4200U:
>
> idct8x8_dc
> SSE2 MMXEXT C
> cycles 22 26 57
>
> idct16x16_dc
> AVX2 SSE2C
> cycles 27 32 249
>
> idct32x32_dc
Le 23 juil. 2014 à 19:33, Christophe Gisquet a
écrit :
> This should be mostly cosmetics.
>
>
Okay. This would also need to clean all transform_dc_ since they are not
use anymore with Rext add-ons, but this will done/change in a near future.
patch okay as it is for now.
Mickaël
_
Patch ok
Le mercredi 23 juillet 2014, Christophe Gisquet <
christophe.gisq...@gmail.com> a écrit :
> Hi,
>
> this may have consequences on threading, so it is a bit more important
> than being able to avoid deriving the other candidates.
>
> --
> Christophe
>
_
Le mercredi 23 juillet 2014, Christophe Gisquet <
christophe.gisq...@gmail.com> a écrit :
> Hi,
>
> this change is actually from openhevc. I just changed the type, not
> the order used. The values are casted to int by the sao functions
> anyway.
>
> --
> Christophe
>
Ok
Mickael
_
hi
Le 23 juil. 2014 à 01:37, Michael Niedermayer a écrit :
> On Wed, Jul 23, 2014 at 12:52:48AM +0200, Carl Eugen Hoyos wrote:
>> On Wednesday 23 July 2014 12:43:06 am Ronald S. Bultje wrote:
>>> Hi,
>>>
>>> On Tue, Jul 22, 2014 at 5:45 PM, Carl Eugen Hoyos wrote:
Hi!
Attached p
45 matches
Mail list logo