Joel Cunningham [Mon, 30 Jan 2017 16:00:44 +0000 (10:00 -0600)]
HTTP: improve performance by reducing forward seeks
This commit optimizes HTTP performance by reducing forward seeks, instead
favoring a read-ahead and discard on the current connection (referred to
as a short seek) for seeks that are within a TCP window's worth of data.
This improves performance because with TCP flow control, a window's worth
of data will be in the local socket buffer already or in-flight from the
sender once congestion control on the sender is fully utilizing the window.
Note: this approach doesn't attempt to differentiate from a newly opened
connection which may not be fully utilizing the window due to congestion
control vs one that is. The receiver can't get at this information, so we
assume worst case; that full window is in use (we did advertise it after all)
and that data could be in-flight
The previous behavior of closing the connection, then opening a new
with a new HTTP range value results in a massive amounts of discarded
and re-sent data when large TCP windows are used. This has been observed
on MacOS/iOS which starts with an initial window of 256KB and grows up to
1MB depending on the bandwidth-product delay.
When seeking within a window's worth of data and we close the connection,
then open a new one within the same window's worth of data, we discard
from the current offset till the end of the window. Then on the new
connection the server ends up re-sending the previous data from new
offset till the end of old window.
For a real world test example, I have MP4 file of ~25MB, which ffplay
only reads ~16MB and performs 177 seeks. With current ffmpeg, this results
in 177 HTTP GETs and ~73MB worth of TCP data communication. With this
patch, ffmpeg issues 4 HTTP GETs and 3 seeks for a total of ~22MB of TCP data
communication.
To support this feature, the short seek logic in avio_seek() has been
extended to call a function to get the short seek threshold value. This
callback has been plumbed to the URLProtocol structure, which now has
infrastructure in HTTP and TCP to get the underlying receiver window size
via SO_RCVBUF. If the underlying URL and protocol don't support returning
a short seek threshold, the default s->short_seek_threshold is used
This feature has been tested on Windows 7 and MacOS/iOS. Windows support
is slightly complicated by the fact that when TCP window auto-tuning is
enabled, SO_RCVBUF doesn't report the real window size, but it does if
SO_RCVBUF was manually set (disabling auto-tuning). So we can only use
this optimization on Windows in the later case
Signed-off-by: Joel Cunningham <joel.cunningham@me.com> Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
This marks the first time anyone has written an Opus encoder without
using any libopus code. The aim of the encoder is to prove how far
the format can go by writing the craziest encoder for it.
Right now the encoder's basic, it only supports CBR encoding, however
internally every single feature the CELT layer has is implemented
(except the pitch pre-filter which needs to work well with the rest of
whatever gets implemented). Psychoacoustic and rate control systems are
under development.
The encoder takes in frames of 120 samples and depending on the value of
opus_delay the plan is to use the extra buffered frames as lookahead.
Right now the encoder will pick the nearest largest legal frame size and
won't use the lookahead, but that'll change once there's a
psychoacoustic system.
Even though its a pretty basic encoder its already outperforming
any other native encoder FFmpeg has by a huge amount.
The PVQ search algorithm is faster and more accurate than libopus's
algorithm so the encoder's performance is close to that of libopus
at zero complexity (libopus has more SIMD).
The algorithm might be ported to libopus or other codecs using PVQ in
the future.
The encoder still has a few minor bugs, like desyncs at ultra low
bitrates (below 9kbps with 20ms frames).
opus_celt: rename structures to better names and reorganize them
This is meant to be applied on top of my previous patch which
split PVQ into celt_pvq.c and made opus_celt.h
Essentially nothing has been changed other than renaming CeltFrame
to CeltBlock (CeltFrame had absolutely nothing at all to do with
a frame) and CeltContext to CeltFrame.
3 variables have been put in CeltFrame as they make more sense
there rather than being passed around as arguments.
The coefficients have been moved to the CeltBlock structure
(why the hell were they in CeltContext and not in CeltFrame??).
Now the encoder would be able to use the exact context the decoder
uses (plus a couple of extra fields in there).
opus_celt: move quantization and band decoding to opus_pvq.c
A huge amount can be reused by the encoder, as the only thing
which needs to be done would be to add a 10 line celt_icwrsi,
a wrapper around it (celt_alg_quant) and templating the
ff_celt_decode_band to replace entropy decoding functions
with entropy encoding.
There is no performance loss but in fact a performance gain of
around 6% which is caused by the compiler being able to optimize
the decoding more efficiently.
avformat/http: Check for truncated buffers in http_connect()
Reported-by: SleepProgger <security@gnutp.com> Reviewed-by: Steven Liu <lingjiujianke@gmail.com> Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Alex Converse [Thu, 9 Feb 2017 16:58:47 +0000 (08:58 -0800)]
aac_latm: Align inband PCE to the start of the payload
A strict reading of the spec seems to imply that it should be aligned to
the start of the element instance tag, but that would break all of the
samples with PCEs.
It seems like a well formed LATM stream should have its PCE in the ASC
rather than inband.
wm4 [Fri, 10 Feb 2017 11:17:24 +0000 (12:17 +0100)]
hwcontext_dxva2: support D3D9Ex
D3D9Ex uses different driver paths. This helps with "headless"
configurations when no user logs in. Plain D3D9 device creation will
fail if no user is logged in, while it works with D3D9Ex.
Signed-off-by: Anton Khirnov <anton@khirnov.net>
Merges Libav commit c2f97f0508708.
wm4 [Thu, 2 Feb 2017 10:27:54 +0000 (11:27 +0100)]
AVFrame: add an opaque_ref field
This is an extended version of the AVFrame.opaque field, which can be
used to attach arbitrary user information to an AVFrame.
The usefulness of the opaque field is rather limited, because it can
store only up to 32 bits of information (or 64 bit on 64 bit systems).
It's not possible to set this field to a memory allocation, because
there is no way to deallocate it correctly.
The opaque_ref field circumvents this by letting the user set an
AVBuffer, which makes the user data refcounted.
Signed-off-by: Anton Khirnov <anton@khirnov.net>
Merges Libav commit 04f3bd349651.
Jan Ekström [Fri, 10 Feb 2017 23:21:14 +0000 (01:21 +0200)]
movenc: add support for track names in ISML manifests
This enables having multiple tracks of the same type which would
be treated as different things by the media server (as opposed to
different bit rate versions of the same track). According to the
smooth streaming specification, just setting the systemLanguage
tag is not enough to note that a track with the same attributes
differs from another one.
Reviewed-by: Martin Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Steven Liu [Sat, 11 Feb 2017 04:32:31 +0000 (12:32 +0800)]
avformat/hlsenc: deprecate hls_wrap option
When user use the hls_wrap, there have many problem:
1. some platform refersh the old but usefull segment
2. CDN(Content Delivery Network) Deliver HLS not friendly
The hls_wrap is used to wrap segments for use little space,
now user can use hls_list_size and hls_flags delete_segments
instead it.
Reviewed-by: Michael Niedermayer <michael@niedermayer.cc> Reviewed-by: Carl Eugen Hoyos <ceffmpeg@gmail.com> Signed-off-by: Steven Liu <lq@chinaffmpeg.org>
James Almer [Fri, 10 Feb 2017 03:53:39 +0000 (00:53 -0300)]
avformat/apetag: fix flag value to signal footer presence
According to the spec[1], a value of 0 means the footer is present and a value
of 1 means it's absent, the exact opposite of header presence flag where 1
means present and 0 absent.
The reason for this is compatibility with APEv1 tags, where there's no header,
footer presence was mandatory for all files, and the flags field was a zeroed
reserved field.
avcodec/mjpegenc: Remove non functional huffman reallocation and error handling
If this is wanted iam not against it but it must be designed to work with all cases
like slice threads, and a single growing buffer does not work very well with slices.
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Mark Thompson [Tue, 29 Nov 2016 22:12:46 +0000 (22:12 +0000)]
vaapi_encode: Pass framerate parameters to driver
Only do this when building for a recent VAAPI version - initial
driver implementations were confused about the interpretation of the
framerate field, but hopefully this will be consistent everywhere
once 0.40.0 is released.
Mark Thompson [Sun, 29 Jan 2017 14:12:20 +0000 (14:12 +0000)]
vaapi_h264: Enable VBR mode
Default to using VBR when a target bitrate is set, unless the max rate
is also set and matches the target. Changes to the Intel driver mean
that min_qp is also respected in this case, so set a codec default to
unset the value rather than using the current default inherited from
the MPEG-4 part 2 encoder.
Mark Thompson [Sun, 29 Jan 2017 14:11:03 +0000 (14:11 +0000)]
vaapi_encode: Support VBR mode
This includes a backward-compatibility hack to choose CBR anyway on
old drivers which have no CBR support, so that existing programs will
continue to work their options now map to VBR.
Mark Thompson [Wed, 4 Jan 2017 23:05:10 +0000 (23:05 +0000)]
vaapi_h264: Scale log2_max_pic_order_cnt_lsb with max_b_frames
Before this change, it was possible to overflow pic_order_cnt_lsb and
generate a stream with invalid POC numbering. This makes sure that
the field is large enough that a single IDR B* P sequence uses fewer
than half the available POC lsb values.
Mark Thompson [Wed, 4 Jan 2017 23:17:23 +0000 (23:17 +0000)]
vaapi_encode: Fix GOP sizing
This change makes the configured GOP size be respected exactly -
previously the value could be exceeded slightly due to flaws in the
frame type selection logic.
Matt Wolenetz [Wed, 14 Dec 2016 23:26:19 +0000 (15:26 -0800)]
lavf/mov.c: Avoid heap allocation wrap in mov_read_uuid
Core of patch is from paul@paulmehta.com
Reference https://crbug.com/643951
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Check value reduced as the code does not support values beyond INT_MAX
Also the check is moved to a more common place and before integer truncation
Aman Gupta [Tue, 7 Feb 2017 03:58:43 +0000 (11:58 +0800)]
avformat/hlsenc: add hls_flag option to write segments to temporary file until complete
Adds a `-hls_flags +temp_file` which will write segment data to
filename.tmp, and then rename to filename when the segment is complete.
This patch is similar in spirit to one used in Plex's ffmpeg fork, and
allows a transcoding webserver to ensure incomplete segment files are
never served up accidentally.
Reviewed-by: Hendrik Leppkes <h.leppkes@gmail.com> Reviewed-by: Bodecs Bela <bodecsb@vivanet.hu> Signed-off-by: Aman Gupta <aman@tmm1.net> Signed-off-by: Steven Liu <lq@chinaffmpeg.org>
lavf/matroskadec: fix is_keyframe for early Blocks
Blocks are marked as key frames whenever the "reference" field is
zero. This breaks for non-keyframe Blocks with a reference timestamp
of zero.
The likelihood of reference timestamp being zero is increased by a
longstanding bug in muxing that encodes reference timestamp as the
absolute time of the referenced frame (rather than relative to the
current Block timestamp, as described in MKV spec).
Now using INT64_MIN to denote "no reference".
Reported to chromium at http://crbug.com/497889 (contains sample)
Mark Thompson [Sun, 5 Feb 2017 14:58:06 +0000 (14:58 +0000)]
Revert "avutil/hwcontext_vaapi: fix SEGV in vaTerminate when vaInitialize fails"
The original code is correctly following the API - vaTerminate() must
be called to free the resources of a VADisplay after it is created by
any of the vaGetDisplay*() calls; it is not necessary to have
successfully called vaInitialize() on it. The segfaults which
prompted this change must therefore be bugs in libva or the driver it
loads.