On Wed, Jul 8, 2020 at 10:34 PM Carl Eugen Hoyos wrote:
> Am Mi., 8. Juli 2020 um 07:28 Uhr schrieb Josef Zlomek :
> >
> > Fixes: 4907
>
> It seems surprising that two commits should fix a ticket.
>
My understanding of ffmpeg internals is limited:
The parser is used for input from a pipe, the de
On Wed, Jul 8, 2020 at 10:04 AM Steven Liu wrote:
>
> (base) liuqi05:dash liuqi$ ./ffplay_g test.webp
> [...]
> Segmentation fault: 11
> (base) liuqi05:dash liuqi$
>
The segfault was caused by the usage of unaligned buffers for frames.
When using aligned buffers, it does not crash anymore.
Thank
Josef Zlomek (12020-07-08):
> Fixes: 4907
>
> Adds support for demuxing of animated WebP.
>
> The WebP demuxer splits the input stream into packets containing one frame.
> It also sets the timing information properly.
Thanks for the patch. A few comments.
>
> Signed-off-by: Josef Zlomek
> ---
Hi,
Am Donnerstag, 9. Juli 2020, 02:56:28 CEST schrieb Manolis Stamatogiannakis:
> On Tue, 7 Jul 2020 at 16:27, Nicolas George wrote:
> > > Is tree threading that important? A PR is essentially a single thread of
> > > discussion.
> >
> > It is a single thread of discussion until the discussion b
Hi,
On Thu, Jul 9, 2020 at 11:44 AM Nicolas George wrote:
> Josef Zlomek (12020-07-08):
> > Fixes: 4907
> >
> > Adds support for demuxing of animated WebP.
> >
> > The WebP demuxer splits the input stream into packets containing one
> frame.
> > It also sets the timing information properly.
>
>
Signed-off-by: Andreas Rheinhardt
---
libavformat/sdp.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavformat/sdp.c b/libavformat/sdp.c
index 34e9839b67..2ce1a62262 100644
--- a/libavformat/sdp.c
+++ b/libavformat/sdp.c
@@ -212,7 +212,7 @@ static char *extradata2psets(AV
No audio stream is created unconditionally and if none has been created,
no packet with stream_index 1 may be returned. This fixes an assert in
ff_read_packet() in libavformat/utils reported in ticket #8782.
Signed-off-by: Andreas Rheinhardt
---
libavformat/mm.c | 2 ++
1 file changed, 2 inserti
Signed-off-by: Andreas Rheinhardt
---
libavformat/sdp.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavformat/sdp.c b/libavformat/sdp.c
index 3acbf5d197..aa0569cd0d 100644
--- a/libavformat/sdp.c
+++ b/libavformat/sdp.c
@@ -231,7 +231,7 @@ static char *extradata2psets_he
In this case, extradata2psets_hevc would have used avio_open_dyn_buf() +
avio_close_dyn_buf() to convert the annex B extradata to the hvcc format
(which is easier parseable); the temporary buffer would then be freed.
avio_close_dyn_buf() + av_free() can be replaced by avio_get_dyn_buf() +
ffio_free
avc_find_startcode_internal() would subtract 6 from a pointer
(representing the end of a buffer) without checking whether the buffer
was actually large enough; but pointer arithmetic is undefined except
when one stays in the buffer.
Signed-off-by: Andreas Rheinhardt
---
libavformat/avc.c | 13 ++
Signed-off-by: Andreas Rheinhardt
---
libavformat/sdp.c | 10 +-
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/libavformat/sdp.c b/libavformat/sdp.c
index aa0569cd0d..023c88a583 100644
--- a/libavformat/sdp.c
+++ b/libavformat/sdp.c
@@ -24,6 +24,7 @@
#include "libavutil/d
On Thu, Jul 9, 2020 at 11:44 AM Nicolas George wrote:
> Josef Zlomek (12020-07-08):
>
> > +@item min_delay
> > +Set the minimum valid delay between frames in milliseconds.
> > +Range is 0 to 6. Default value is 10.
> > +
> > +@item max_webp_delay
> > +Set the maximum valid delay between frame
Hi,
Could someone please take a look at this?
It's a pretty small fix and it has been sitting in the queue since March.
Best Regards,
Jacek
pon., 16 mar 2020 o 11:54 Jacek Tomasiak
napisał(a):
> HLS input streams ignored CLI parameters which control the probing
> process. Each stream was initi
Thanks Gerion.
This sounds more or less like what I had in mind (but I was trying to keep
it short).
– The present thread is a non-technical discussion, and branching indeed
helps. The mailing list is perfect for this kind of discussions.
– Patches with narrow scope are focused, technical discuss
The current code for parsing unsigned exponential-golomb codes is not
suitable for long codes: If there are enough leading zeroes, it
left-shifts 1 by 32 places and uses get_bitsz to read 32 bits, although
it is only suitable to read 0-25 bits. There is an av_assert2 to
ensure this when using the u
On 7/9/2020 3:01 AM, Jeyapal, Karthick wrote:
> Pushed the patch, with minor change to commit message.
>
> Thanks,
> Karthick
> On 6/30/20 11:14 AM, Siyuan Huang wrote:
>> Update with no file mode changes :
>>
>>
>>
>> From 818095d4f0aa50dfee3cb0622146a2180801c5fe Mon Sep 17 00:00:00 2001
>>
>>
Signed-off-by: James Almer
---
libavfilter/x86/vf_blend.asm | 90 ++--
1 file changed, 45 insertions(+), 45 deletions(-)
diff --git a/libavfilter/x86/vf_blend.asm b/libavfilter/x86/vf_blend.asm
index 251bbb5a12..77a6ed04c7 100644
--- a/libavfilter/x86/vf_blend.asm
Signed-off-by: James Almer
---
libavcodec/x86/h264_deblock.asm | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavcodec/x86/h264_deblock.asm b/libavcodec/x86/h264_deblock.asm
index 6702ae98d4..a2e745cd8e 100644
--- a/libavcodec/x86/h264_deblock.asm
+++ b/libavcodec/x86/h264_
From: Henrik Gramner
Some new warnings regarding use of empty macro parameters has been
added, so adjust some x86inc code to silence those
Signed-off-by: James Almer
---
libavutil/x86/x86inc.asm | 46 ++--
1 file changed, 30 insertions(+), 16 deletions(-)
d
Version bump would be added when upstreamed.
Signed-off-by: Linjie Fu
---
libavcodec/avcodec.h | 1 +
libavcodec/hevc_ps.c | 2 ++
libavcodec/profiles.c | 1 +
3 files changed, 4 insertions(+)
diff --git a/libavcodec/avcodec.h b/libavcodec/avcodec.h
index c91b2fd..32e4770 100644
--- a/libavco
As a part of the support for VA-API HEVC SCC decoding (gen12, Tiger Lake+).
The full support could be accessed in:
https://github.com/intel-media-ci/ffmpeg/pull/231
The VAAPI part would be provided later once the implementations of native
parsing
and reference management are all decent.
[Reques
Signed-off-by: Linjie Fu
---
libavcodec/hevcdec.c | 6 ++
libavcodec/hevcdec.h | 4
2 files changed, 10 insertions(+)
diff --git a/libavcodec/hevcdec.c b/libavcodec/hevcdec.c
index e363e68..8194f18 100644
--- a/libavcodec/hevcdec.c
+++ b/libavcodec/hevcdec.c
@@ -800,6 +800,12 @@ static
According to 7.3.2.2.3 in T-REC-H.265-201911.
Signed-off-by: Linjie Fu
---
libavcodec/hevc.h| 3 +++
libavcodec/hevc_ps.c | 32 +---
libavcodec/hevc_ps.h | 15 +++
3 files changed, 47 insertions(+), 3 deletions(-)
diff --git a/libavcodec/hevc.h b/lib
Screen Content Coding allows P slice in an IDR frame, and would add
frames themselves to the short term rps in RefList0 (8-10 in spec),
hence some previous restricts are not suitable any more.
[Request for comments].
Signed-off-by: Linjie Fu
---
libavcodec/hevc_refs.c | 21 +++--
uint8_t is big enough and keep consistent with the definition in
cbs_h265.h.
Signed-off-by: Linjie Fu
---
libavcodec/hevcdec.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavcodec/hevcdec.h b/libavcodec/hevcdec.h
index 1164af2..464eb7c 100644
--- a/libavcodec/hevcdec.h
Signed-off-by: Linjie Fu
---
libavcodec/hevc_ps.c | 53 ++--
libavcodec/hevc_ps.h | 17 +
2 files changed, 68 insertions(+), 2 deletions(-)
diff --git a/libavcodec/hevc_ps.c b/libavcodec/hevc_ps.c
index 7610d3b..887b960 100644
--- a
According to 7.3.6.1, use_integer_mv_flag should be parsed if
motion_vector_resolution_control_idc equals to 2.
Otherwise wrong parameters in the subsequent parsing procedures
would be got.
Signed-off-by: Linjie Fu
---
libavcodec/hevcdec.c | 8
libavcodec/hevcdec.h | 1 +
2 files chang
On 7/9/2020 1:17 PM, Linjie Fu wrote:
> According to 7.3.2.2.3 in T-REC-H.265-201911.
>
> Signed-off-by: Linjie Fu
> ---
> libavcodec/hevc.h| 3 +++
> libavcodec/hevc_ps.c | 32 +---
> libavcodec/hevc_ps.h | 15 +++
> 3 files changed, 47 insertions(+)
On 7/9/2020 1:51 PM, James Almer wrote:
> On 7/9/2020 1:17 PM, Linjie Fu wrote:
>> According to 7.3.2.2.3 in T-REC-H.265-201911.
>>
>> Signed-off-by: Linjie Fu
>> ---
>> libavcodec/hevc.h| 3 +++
>> libavcodec/hevc_ps.c | 32 +---
>> libavcodec/hevc_ps.h | 15
On Fri, 2020-06-19 at 10:33 +0200, Sergio Acereda wrote:
> +/**
> + * Fetch titlebar height from handle.
> + *
> + * @param hwnd Handle of the window.
> + * @return titlebar height
> + */
> +static int
> +calc_titlebar_height(HWND hwnd) {
> +TITLEBARINFOEX tinfo;
> +tinfo.cbSize = sizeof(ti
[...]
On Wed, Jul 8, 2020 at 7:42 PM James Almer wrote:
> The sum of all sizes should be size_t if each size is a size_t. But if
> you do that you can't return error values, so i'm not sure what to
> suggest other than just stick to ints for both (ptrdiff_t linesizes
> should be ok).
> I'd like t
On Fri, Jul 10, 2020 at 12:54 AM James Almer wrote:
>
> On 7/9/2020 1:51 PM, James Almer wrote:
> > On 7/9/2020 1:17 PM, Linjie Fu wrote:
> >> According to 7.3.2.2.3 in T-REC-H.265-201911.
> >>
> >> Signed-off-by: Linjie Fu
> >> ---
> >> libavcodec/hevc.h| 3 +++
> >> libavcodec/hevc_ps.c |
On 7/9/2020 3:20 PM, Brian Kim wrote:
> [...]
>
> On Wed, Jul 8, 2020 at 7:42 PM James Almer wrote:
>> The sum of all sizes should be size_t if each size is a size_t. But if
>> you do that you can't return error values, so i'm not sure what to
>> suggest other than just stick to ints for both (pt
Up until now, a caller had no function to easily parse a single annex B
NAL unit. There was a "low-level" function that returned the position of
the beginning of the next start code, which could be used to get the
beginning and the end of NAL units. Yet using this function still
required calling th
When an empty annex B NAL unit (i.e. 0x01 immediately followed by a
three or four-byte start code) is encountered during annex B->mp4
conversion, a NAL unit of size zero is created; this is invalid.
Furthermore, several functions simply presumed all NAL units to be
nonempty and treated the firs
by using the new ff_avc_parse_nalu function.
Signed-off-by: Andreas Rheinhardt
---
libavformat/avc.c| 11 ++-
libavformat/movenccenc.c | 12 ++--
2 files changed, 4 insertions(+), 19 deletions(-)
diff --git a/libavformat/avc.c b/libavformat/avc.c
index 39078aa5be..d08905
by using ff_avc_parse_nalu() which means that one no longer has to
take care of finding both the start as well as the end of a NAL unit
separately.
Signed-off-by: Andreas Rheinhardt
---
libavformat/sdp.c | 34 +++---
1 file changed, 15 insertions(+), 19 deletions(-)
When creating H.264 or HEVC extradata from annex B extradata or when
transforming annex B into HEVC while also filtering parameter sets away,
the whole input has first been transformed into mp4-style H.264/HEVC
in order to simplify parsing at the next step. By using ff_avc_parse_nalu,
one can avoid
and also keep the reserved_zero_2bits in the corresponding H264SPS member.
This simplifies parsing and will prove beneficial in ff_isom_write_avcc.
Signed-off-by: Andreas Rheinhardt
---
libavformat/avc.c| 8 +---
libavformat/avc.h| 2 +-
libavformat/mxfenc.c | 2 +-
3 files changed,
The annex B format of H.264 and HEVC allows there to be trailing zero
bytes in byte stream NAL units; this means that there can be trailing
zero bytes after the actual NAL units (besides start codes
prepended to them).
Samples in mp4 contain a sequence of length-prefixed NAL units and not
byte str
On 7/9/2020 7:35 AM, Andreas Rheinhardt wrote:
> The current code for parsing unsigned exponential-golomb codes is not
> suitable for long codes: If there are enough leading zeroes, it
> left-shifts 1 by 32 places and uses get_bitsz to read 32 bits, although
> it is only suitable to read 0-25 bits.
According to the specifications, both the SPS as well as the PPS in
an AVCDecoderConfigurationRecord (the mp4-style extradata for H.264) are
supposed to be ordered according to increasing SPS/PPS id.
ff_isom_write_avcc did not ensure this; in fact, it did not even ensure
that there are no two param
For some profiles, the AVCDecoderConfigurationRecord contains an array
containing all the SPS Extension NAL units; for the other profiles,
these NAL units should be put into the SPS array. Yet the latter hasn't
been done. This commit implements it.
Signed-off-by: Andreas Rheinhardt
---
Honestly,
This commit adds some basic checks for invalid values when parsing an
SPS; e.g. the earlier code did not check whether the values read from
the bitstream get truncated when returned via the H264SPS struct (whose
members are uint8_t), so that a caller could not even check for this
error itself.
Sig
On Thu, Jul 09, 2020 at 02:01:45AM +0200, Manolis Stamatogiannakis wrote:
> On Tue, 7 Jul 2020 at 16:00, Nicolas George wrote:
>
> > Manolis Stamatogiannakis (12020-07-05):
[...]
> > > +Additionally, it is recommended to register for a
> > > +@uref{https://patchwork.ffmpeg.org, patchwork account}
On Thu, 9 Jul 2020, lance.lmw...@gmail.com wrote:
On Wed, Jul 08, 2020 at 11:59:17PM +0200, Marton Balint wrote:
On Wed, 8 Jul 2020, lance.lmw...@gmail.com wrote:
> On Wed, Jul 01, 2020 at 09:39:58PM +0800, lance.lmw...@gmail.com wrote:
> > From: Limin Wang
> >
> > Signed-off-by: Limin W
On Mon, Jul 6, 2020 at 10:56 AM Yongle Lin wrote:
>
>
> On Thu, Jun 25, 2020 at 12:09 PM Yongle Lin
> wrote:
>
>> Add qp visualization in codecview filter which supports H264 and VP9
>> codecs. Add options for luma/chroma qp and AC/DC qp as well. There is a old
>> way to visualize it but it's de
On Mon, Jul 6, 2020 at 11:31 AM Yongle Lin wrote:
> it makes sense to export block structure like src_x, src_y, width and
> height when segmentation isn't enable so we could visualize and see the
> structure of the block.
> ---
> libavcodec/vp9.c | 8 +++-
> 1 file changed, 3 insertions(+),
On Thu, Jul 2, 2020 at 8:09 AM Michael Niedermayer
wrote:
> On Wed, Jul 01, 2020 at 05:42:48PM +, Yongle Lin wrote:
> > example command line to visualize block decomposition:
> > ./ffmpeg -export_side_data +venc_params -i input.webm -vf
> > codecview=bs=true output.webm
> > ---
>
> > doc/fil
Fixes: 86987846-429c8d80-c197-11ea-916b-bb4738e09687.jpg
Fixes: Regression since ec3d8a0e6945fe015d16cd98a1e7dbb4be815c15
Signed-off-by: Michael Niedermayer
---
libavcodec/mjpeg_parser.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavcodec/mjpeg_parser.c b/libavcodec/mj
Given the questions regarding the format used for the timecodes, this
patch from my personal repo might help:
commit e4c9ae661fb55e59eff1461140a8cd039b7c3f04
Author: Devin Heitmueller
Date: Thu Jul 5 17:18:36 2018 -0400
Clarify where the binary representation comes from for ffmpeg's SMPTE s
On Mon, Mar 16, 2020 at 7:00 PM Jacek Tomasiak wrote:
>
> HLS input streams ignored CLI parameters which control the probing
> process. Each stream was initialized with fresh context and used
> the default values for probing.
> ---
> libavformat/hls.c | 8
> 1 file changed, 8 insertions(
On Thu, Jul 09, 2020 at 07:16:28PM -0400, Devin Heitmueller wrote:
> Given the questions regarding the format used for the timecodes, this
> patch from my personal repo might help:
>
> commit e4c9ae661fb55e59eff1461140a8cd039b7c3f04
> Author: Devin Heitmueller
> Date: Thu Jul 5 17:18:36 2018 -0
On Thu, Jul 09, 2020 at 11:15:51PM +0200, Marton Balint wrote:
>
>
> On Thu, 9 Jul 2020, lance.lmw...@gmail.com wrote:
>
> > On Wed, Jul 08, 2020 at 11:59:17PM +0200, Marton Balint wrote:
> > >
> > >
> > > On Wed, 8 Jul 2020, lance.lmw...@gmail.com wrote:
> > >
> > > > On Wed, Jul 01, 2020 at
On Thu, Jul 9, 2020 at 11:30 AM wrote:
>
> From: Gautam Ramakrishnan
>
> This patch fixes a check in the RPCL progression order,
> tested on p1_07.j2k.
> ---
> libavcodec/jpeg2000dec.c | 19 ---
> 1 file changed, 12 insertions(+), 7 deletions(-)
>
> diff --git a/libavcodec/jpeg20
55 matches
Mail list logo
