Commit graph

2292 commits

Author SHA1 Message Date
Sebastian Dröge
4815036708 vp8enc: Store configuration in the vpx_codec_enc_cfg_t struct instead of duplicating all variables
Also protect encoder with a mutex.
2012-09-17 09:29:26 +02:00
Sebastian Dröge
8ed85c7538 vp8enc: Update documentation to reflect new property names
...and also link to the WebM encoder parameters website.
2012-09-16 16:03:06 +02:00
Sebastian Dröge
acd0824cf6 vp8enc: Make some property names more readable 2012-09-16 15:58:51 +02:00
Sebastian Dröge
d4eaa932e5 vpx: Integrate into the build system 2012-09-16 15:43:39 +02:00
Sebastian Dröge
2a4e3d2110 vpx: Rename vp8 plugin to vpx
This is using libvpx, which can support more codecs than just VP8
and will likely support future codecs.
2012-09-16 15:33:57 +02:00
Sebastian Dröge
eed915b05e vp8: Apply remaining changes that got lost while moving the plugin via git am thanks to merges 2012-09-16 15:32:24 +02:00
Sebastian Dröge
c4e0aff12b [MOVED FROM BAD 134/134] vp8dec: Unref input/output states when stopping the decoder 2012-09-16 15:32:04 +02:00
Sebastian Dröge
9c0ff2f381 [MOVED FROM BAD 133/134] vp8enc: Update realtime profile to the new properties 2012-09-16 15:32:04 +02:00
Sebastian Dröge
b9ba9a28da [MOVED FROM BAD 132/134] vp8: Require latest libvpx release (1.1.0 from May 2012)
Fixes bug #684116 and simplifies configure checks.
2012-09-16 15:32:04 +02:00
Sebastian Dröge
711ae3c94b [MOVED FROM BAD 131/134] vp8enc: Use a string field for the profile in the caps
Just for consistency with all the other codecs.
2012-09-16 15:32:04 +02:00
Sebastian Dröge
1639d1e150 [MOVED FROM BAD 130/134] vp8enc: Correctly set profile in caps 2012-09-16 15:31:57 +02:00
Sebastian Dröge
35520df7c0 [MOVED FROM BAD 129/134] vp8: Update copyright and authors 2012-09-16 15:31:57 +02:00
Sebastian Dröge
392bd12a45 [MOVED FROM BAD 128/134] vp8enc: Rework encoder properties to be more in line with the libvpx tools and API
Also add all available properties.
2012-09-16 15:31:57 +02:00
Mark Nauwelaerts
c70dc304f4 [MOVED FROM BAD 127/134] replace gst_element_class_set_details_simple with gst_element_class_set_metadata 2012-09-16 15:31:48 +02:00
Edward Hervey
210742d934 [MOVED FROM BAD 126/134] vp8dec: Call gst_video_decoder_negotiate() 2012-09-16 15:31:38 +02:00
Sebastian Dröge
d61c1e9e02 [MOVED FROM BAD 125/134] vp8dec: Add support for multiple decoding threads 2012-09-16 15:31:38 +02:00
Sebastian Dröge
59d906b6c3 [MOVED FROM BAD 124/134] vp8dec: Add support for the MFQE postprocessing flag
Which is enabled by default if postprocessing is enabled.
2012-09-16 15:31:27 +02:00
Sebastian Dröge
58e8efb6f7 [MOVED FROM BAD 123/134] vp8: Use pkg-config file for getting the LIBS and CFLAGS 2012-09-16 15:31:27 +02:00
Sebastian Dröge
26f1269486 [MOVED FROM BAD 122/134] vp8enc: Update the per-component strides for every frame too
This is necessary because of GstVideoAlignment
2012-09-16 15:31:10 +02:00
Oleksij Rempel
4cb020932c [MOVED FROM BAD 121/134] vp8enc: initiate encoder to fix a crash.
Without this patch vp8enc send header before and after first
key frame. On second keyframe vp8dec will crash without getting
decoded frame. With this pipe it is easy to reproduce this issue:
gst-launch-1.0 videotestsrc ! vp8enc ! vp8dec ! fakesink

https://bugzilla.gnome.org/show_bug.cgi?id=680667
2012-09-16 15:31:10 +02:00
Sebastian Dröge
16db1b677a [MOVED FROM BAD 120/134] tag: Update for taglist/tag event API changes 2012-09-16 15:31:10 +02:00
Sebastian Dröge
1fdf044acf [MOVED FROM BAD 119/134] ext: Update for video base classes API changes 2012-09-16 15:31:10 +02:00
Oleksij Rempel
2d7cbc8113 [MOVED FROM BAD 118/134] vp8enc: fix memory leak
unref frame. i hope it is correct place to do it.

Signed-off-by: Oleksij Rempel <bug-track@fisher-privat.net>
2012-09-16 15:31:10 +02:00
Wim Taymans
c45fcd0f01 [MOVED FROM BAD 117/134] update for query api changes 2012-09-16 15:31:10 +02:00
Wim Taymans
365557edc1 [MOVED FROM BAD 116/134] update for query api changes 2012-09-16 15:31:10 +02:00
Wim Taymans
af3ddea42d [MOVED FROM BAD 115/134] update for allocation query changes 2012-09-16 15:31:10 +02:00
Vincent Penquerc'h
c76f4fe3ef [MOVED FROM BAD 114/134] vp8: fix codec state leaks
I only tested that vp8enc ! vp8dec does not crash, as valgrind does not grok
at least one of the instructions used by vp8enc, preventing me from checking
a leak, and the lack of one after the patch.
2012-09-16 15:31:10 +02:00
Wim Taymans
e644f68e26 [MOVED FROM BAD 113/134] update for tag event change 2012-09-16 15:31:10 +02:00
Sebastian Dröge
dcb3fc514e [MOVED FROM BAD 112/134] vp8: Port to 0.11 again 2012-09-16 15:31:10 +02:00
Vincent Penquerc'h
01b86b6b2a [MOVED FROM BAD 111/134] vp8enc: fix target bitrate config with libvpx 1.1.0
libvpx 1.1.0 disallows a bitrate of 0, which was used by
vp8enc as a default value.

Instead, we use the default libvpx bitrate, scaled to our
video size, if no bitrate was specified.

This fixes encoding VP8 video with libvpx 1.1.0.

https://bugzilla.gnome.org/show_bug.cgi?id=676245
2012-09-16 15:30:22 +02:00
Sebastian Dröge
fd598a8e3d [MOVED FROM BAD 110/134] vp8enc: Update for GstVideoCodecFrame API changes 2012-09-16 15:30:22 +02:00
Thiago Santos
1f65326f67 [MOVED FROM BAD 109/134] vp8dec: Improve output_state handling
Avoid getting output_state for every buffer as that requires
getting the objectlock and doing reference counting. Store it locally
when it is created and use it.
2012-09-16 15:30:22 +02:00
Thiago Santos
4484130cbf [MOVED FROM BAD 108/134] vp8dec: Use outputstate when copying output buffer data
Using the input state was causing a crash because the strides/offsets
would be wrong. Fix it by using the output as we are dealing with
the decoded frame.
2012-09-16 15:30:05 +02:00
Edward Hervey
74aa2e7c6a [MOVED FROM BAD 107/134] vp8: Port to -base video base classes
Conflicts:

	ext/vp8/Makefile.am
	ext/vp8/gstvp8dec.c
	ext/vp8/gstvp8enc.c

Back to 0.10 state for now, need to be ported again.
2012-09-16 15:29:52 +02:00
Vincent Penquerc'h
dccf3dc0b6 [MOVED FROM BAD 106/134] vp8enc: fix target bitrate config with libvpx 1.1.0
libvpx 1.1.0 disallows a bitrate of 0, which was used by
vp8enc as a default value.

Instead, we use the default libvpx bitrate, scaled to our
video size, if no bitrate was specified.

This fixes encoding VP8 video with libvpx 1.1.0.

https://bugzilla.gnome.org/show_bug.cgi?id=676245
2012-09-16 15:28:18 +02:00
Sebastian Dröge
4fa72a5c23 [MOVED FROM BAD 105/134] gst: Update for GST_PLUGIN_DEFINE() API changes 2012-09-16 15:28:18 +02:00
Sebastian Dröge
9894fc2abe [MOVED FROM BAD 104/134] gst: Update versioning 2012-09-16 15:28:18 +02:00
Sebastian Dröge
3cfb400de1 [MOVED FROM BAD 103/134] vp8enc: Fix 'argument to 'sizeof' in 'memset' call is the same expression as the destination' compiler warning 2012-09-16 15:28:18 +02:00
Wim Taymans
0eeb65f989 [MOVED FROM BAD 102/134] update for HEADER flag 2012-09-16 15:28:18 +02:00
Mark Nauwelaerts
85c73e93f3 [MOVED FROM BAD 101/134] port some more to new memory API
Fixes #668677.
2012-09-16 15:28:18 +02:00
Oleksij Rempel (Alexey Fisher)
ec7f8034af [MOVED FROM BAD 100/134] vp8enc: trace outgoing timestamps
add info level prints for outgoing timestamps.

Signed-off-by: Oleksij Rempel (Alexey Fisher) <bug-track@fisher-privat.net>
2012-09-16 15:28:18 +02:00
Oleksij Rempel (Alexey Fisher)
03d9022587 [MOVED FROM BAD 099/134] vp8dec: use is_alt_data option to prevent timestamp collisions
altref/invisible frames usually stored in container with same timestamp as
dependet frame. This make basevideodecoder to update timestamp for dependet
frame and couse TS colision on next frame:

                 ^- here is altref
time     : 1 2 3 4 5 6 7 8 9
webm ts  : 1   3 5 5   7   9
vp8dec ts: 1   3   7   7   9

Fix bug: https://bugzilla.gnome.org/show_bug.cgi?id=655245

Signed-off-by: Oleksij Rempel (Alexey Fisher) <bug-track@fisher-privat.net>
2012-09-16 15:28:18 +02:00
Oleksij Rempel (Alexey Fisher)
772f7b6605 [MOVED FROM BAD 098/134] vp8: add initial preset file
This is initial preset file, currently with only one profile
for realtime encoding.

Signed-off-by: Oleksij Rempel (Alexey Fisher) <bug-track@fisher-privat.net>
2012-09-16 15:28:18 +02:00
Vincent Penquerc'h
d286b04ee0 [MOVED FROM BAD 097/134] various: fix pad template ref leaks
https://bugzilla.gnome.org/show_bug.cgi?id=662664
2012-09-16 15:28:18 +02:00
Tim-Philipp Müller
020b7c902d [MOVED FROM BAD 096/134] vp8dec: use new basevideodecoder API to drop frames and get QoS messages posted 2012-09-16 15:27:15 +02:00
Mart Raudsepp
8fe80026a5 [MOVED FROM BAD 095/134] mimic, opencv, vp8, acmmp3dec, linsys: Don't build static plugins
Pass --tag=disable-static to libtool everywhere where it's been forgotten

https://bugzilla.gnome.org/show_bug.cgi?id=663768
2012-09-16 15:27:15 +02:00
Edward Hervey
c03ae2f3c2 [MOVED FROM BAD 094/134] vp8: Port to 0.11 2012-09-16 15:27:15 +02:00
David Schleef
653ad8b286 [MOVED FROM BAD 093/134] vp8enc: fix drop-frame property
Fixes #656929.
2012-09-16 15:27:15 +02:00
Vincent Penquerc'h
b20d432046 [MOVED FROM BAD 092/134] vp8: probe for the new tuning API to keep building with older libvpx
https://bugzilla.gnome.org/show_bug.cgi?id=656928
2012-09-16 15:27:15 +02:00
Sebastian Dröge
8e143b1c7d [MOVED FROM BAD 091/134] vp8enc: Remove unused and useless variable in tags handling 2012-09-16 15:27:15 +02:00
Sebastian Dröge
c58dd70675 [MOVED FROM BAD 090/134] vp8enc: Update for basevideoencoder ::get_caps() removal 2012-09-16 15:27:15 +02:00
David Schleef
73a6edbfe2 [MOVED FROM BAD 089/134] vp8enc: Add more properties 2012-09-16 15:27:15 +02:00
Alexey Fisher
1764be1296 [MOVED FROM BAD 088/134] vp8enc: add min/maxsection-pct option
This options should be good to redeuce decode CPU load.
for lowend hardware:
minsection-pct=15 maxsection-pct=400

for hiend hw:
minsection-pct=5 maxsection-pct=800

see example:
http://www.webmproject.org/tools/encoder-parameters/#2-pass_vbr_encoding_for_smooth_playback_on_low-end_hardware

Signed-off-by: Alexey Fisher <bug-track@fisher-privat.net>
Signed-off-by: David Schleef <ds@schleef.org>
2012-09-16 15:27:15 +02:00
Alexey Fisher
01ec7efbe4 [MOVED FROM BAD 087/134] vp8enc: add lag-in-frames option.
This option set maximum of frames codec should remember,
to make better prediktion for alt-ref frames.

See example:
http://www.webmproject.org/tools/encoder-parameters/#2-pass_best_quality_vbr_encoding

Signed-off-by: Alexey Fisher <bug-track@fisher-privat.net>
Signed-off-by: David Schleef <ds@schleef.org>
2012-09-16 15:27:15 +02:00
Alexey Fisher
24d21d5e5f [MOVED FROM BAD 086/134] vp8enc: use multipass.cache file name as default for multipass mode.
Signed-off-by: Alexey Fisher <bug-track@fisher-privat.net>
Signed-off-by: David Schleef <ds@schleef.org>
2012-09-16 15:27:15 +02:00
Sebastian Dröge
6af5e9ce38 [MOVED FROM BAD 085/134] vp8enc: Update for GstBaseVideoEncoder::finish() signature change 2012-09-16 15:27:15 +02:00
Olivier Crête
140bd5c092 [MOVED FROM BAD 084/134] vp8: Fix set-but-unused warnings 2012-09-16 15:27:15 +02:00
Sebastian Dröge
c498e188b9 [MOVED FROM BAD 083/134] vp8enc: Use destroy notify to free the coder hook 2012-09-16 15:27:15 +02:00
David Schleef
ec126d9589 [MOVED FROM BAD 082/134] vp8enc: update for new libvpx api 2012-09-16 15:27:15 +02:00
Alexey Fisher
2d8a989c95 [MOVED FROM BAD 081/134] vp8enc: generate a timestamp for alt-ref frames.
It will fix handling of altref/invisible frames since matroska-mux
drop any fram with no timestamp.

see also:
http://www.webmproject.org/code/specs/container/
The encoder will currently set the AR's timestamp as close as possible
to the previous frame while attempting to provide a timestamp that is
strictly increasing. In cases where the time base given to the encoder
at configure time is not granular enough to allow for this the AR
will share the same timestamp as D, but should be
treated as having no duration.

Fixes bug #652951

Signed-off-by: Alexey Fisher <bug-track@fisher-privat.net>
2012-09-16 15:27:15 +02:00
Alexey Fisher
293d378289 [MOVED FROM BAD 080/134] vp8dec: add check if we have legal aspect-ratio before reset it.
the commit f9b552f049 (vp8dec: set par to 1/1)
will fix situation where no aspect-ratio is set, but it brake
stream with available aspect-ratio. This patch fix it.

Fixes: #652902.

Signed-off-by: Alexey Fisher <bug-track@fisher-privat.net>
2012-09-16 15:27:15 +02:00
David Schleef
e46f4318f7 [MOVED FROM BAD 079/134] vp8dec: set par to 1/1 2012-09-16 15:27:15 +02:00
Sebastian Dröge
24519a0ad3 [MOVED FROM BAD 078/134] vp8enc: Name max/min quantizer properties {max,min}-quantizer
Also improve quality property description.
2012-09-16 15:27:15 +02:00
Alexey Fisher
fbf253c632 [MOVED FROM BAD 077/134] vp8enc: Add properties to select a maximum and minimum quantizer
Fixes bug #641405.
2012-09-16 15:27:15 +02:00
Alexey Fisher
b6a52550b4 [MOVED FROM BAD 076/134] vp8enc: Fix quality to (constant) quantizer mapping
This now allows to select all possible quantizers between
0 and 63.

See bug #641405.
2012-09-16 15:27:15 +02:00
Mark Nauwelaerts
81d2ba6c9d [MOVED FROM BAD 075/134] vp8dec: debug code style fixes 2012-09-16 15:27:15 +02:00
Mark Nauwelaerts
28ec65b5a4 [MOVED FROM BAD 074/134] vp8dec: propagate downstream flow return to upstream 2012-09-16 15:27:15 +02:00
Mark Nauwelaerts
e8a200c6a2 [MOVED FROM BAD 073/134] basevideodecoder: really and only set src pad caps whenever requested
... since subclass is expected to be wise enough to know when to do so.
2012-09-16 15:27:15 +02:00
Mark Nauwelaerts
47067f7d1e [MOVED FROM BAD 072/134] basevideodecoder: invoke subclass start method at state change and use set_format
While this changes API slightly (e.g. actually uses set_format now), which is OK
for unstable API, it has following merits:
* symmetric w.r.t. stop at state change
* in line with other base class practice
* otherwise no subclass method at state change (global activation time)
Moreover, subclassese are either unaffected or trivially adjusted accordingly.
2012-09-16 15:27:15 +02:00
Mark Nauwelaerts
e03fefc740 [MOVED FROM BAD 071/134] basevideodecoder: subsume skip_frame into finish_frame 2012-09-16 15:27:15 +02:00
Mark Nauwelaerts
1d8a499ceb [MOVED FROM BAD 070/134] basevideoencoder: provide proper upstream flow return handling 2012-09-16 15:27:15 +02:00
Mark Nauwelaerts
ddd32f1b05 [MOVED FROM BAD 069/134] vp8enc: minor optimization in setting up image buffer 2012-09-16 15:27:15 +02:00
Mark Nauwelaerts
b6ca1c8171 [MOVED FROM BAD 068/134] vp8enc: refactor frame processing 2012-09-16 15:27:14 +02:00
Mark Nauwelaerts
a24df7d3c7 [MOVED FROM BAD 067/134] vp8enc: do init at set_format time 2012-09-16 15:27:14 +02:00
Mark Nauwelaerts
33ec5fd08f [MOVED FROM BAD 066/134] vp8enc: fix keyframe forcing 2012-09-16 15:27:14 +02:00
Mark Nauwelaerts
e68033782b [MOVED FROM BAD 065/134] basevideocodec: remove redundant caps field
... as it is already at hand as the src pad's negotiated caps.
2012-09-16 15:27:14 +02:00
Mark Nauwelaerts
402879a52e [MOVED FROM BAD 064/134] vp8enc: use baseclass event virtual handler 2012-09-16 15:27:14 +02:00
David Schleef
280354df3f [MOVED FROM BAD 063/134] basevideo: merge utils header into basevideocodec 2012-09-16 15:27:14 +02:00
Tim-Philipp Müller
32505ef105 [MOVED FROM BAD 062/134] vp8: fix LIBADD order in Makefile.am 2012-09-16 15:27:14 +02:00
Alexey Fisher
4d5ad6d0ca [MOVED FROM BAD 061/134] vp8enc: Add description for bitrate units. 2012-09-16 15:27:14 +02:00
David Schleef
7734336a97 [MOVED FROM BAD 060/134] vp8enc: Readd setting of granulepos
Revert parts of last patch that removed setting of granulepos.
oggmux still requires correct granulepos in incoming packet.
2012-09-16 15:27:14 +02:00
David Schleef
161b89d552 [MOVED FROM BAD 059/134] vp8enc: Don't override timestamps set by base class
Because the base class does it correctly.
Fixes: #635720, #625558.
2012-09-16 15:27:14 +02:00
Edward Hervey
785bf960dd [MOVED FROM BAD 058/134] vp8: Remove dead assignments 2012-09-16 15:27:14 +02:00
David Schleef
49d0bcce24 [MOVED FROM BAD 057/134] basevideo: Move common fields/functions to basecodec 2012-09-16 15:27:14 +02:00
David Schleef
d1d6fe1b8d [MOVED FROM BAD 056/134] basevideo: Move deadline to frame structure 2012-09-16 15:27:14 +02:00
Philip Jägenstedt
7af2375a46 [MOVED FROM BAD 055/134] vp8dec: Set GstBaseVideoDecoder::packetized to TRUE as soon as possible
This fixes an infinite loop if an EOS event is received before
GstBaseVideoDecoder::start() is called, e.g. immediately when the
pads are activated.

Fixes bug #626815.
2012-09-16 15:27:14 +02:00
Sebastian Dröge
3509034bc9 [MOVED FROM BAD 054/134] vp8enc: Add support for enabling automatic insertion of alt-ref frames by the encoder 2012-09-16 15:27:14 +02:00
Sebastian Dröge
084d1588a9 [MOVED FROM BAD 053/134] vp8enc: Fix handling of invisible/alt ref frames 2012-09-16 15:27:14 +02:00
Sebastian Dröge
6be4d8efe0 [MOVED FROM BAD 052/134] vp8: Add initial documentation, based on the theoradec/theoraenc documentation 2012-09-16 15:27:14 +02:00
Sebastian Dröge
2856d8a6fa [MOVED FROM BAD 051/134] vp8: Move structure definitions, etc to public header files for gtk-doc 2012-09-16 15:27:14 +02:00
Sebastian Dröge
83ffb384d0 [MOVED FROM BAD 050/134] vp8enc: Implement multipass encoding
Fixes bug #621348.
2012-09-16 15:27:14 +02:00
Sebastian Dröge
7f3e1a1484 [MOVED FROM BAD 049/134] vp8enc: Set VP8E_SET_CPUUSED to 0
This setting controls how much CPU can be used by the encoder, specified
in fractions of 16. Negative values mean strict enforcement of this
while positive values are adaptive.

The default value is -4, which means that we're not running as fast
as possible and probably are wasting some quality. 0 is the recommended
default by libvpx upstream.
2012-09-16 15:27:14 +02:00
Sebastian Dröge
ec2c2b5fcc [MOVED FROM BAD 048/134] vp8enc: Use VPX defines for REALTIME, GOOD/BEST quality deadlines instead of our own
These are the values used for the speed property.
2012-09-16 15:27:14 +02:00
Tim-Philipp Müller
87fee9151b [MOVED FROM BAD 047/134] vp8enc: fix printf format warning in log message
gstvp8enc.c:564: error: format ‘%d’ expects type ‘int’, but argument 8 has type ‘size_t’
gstvp8enc.c:744: error: format ‘%d’ expects type ‘int’, but argument 8 has type ‘size_t’
2012-09-16 15:27:14 +02:00
Tim-Philipp Müller
86db423828 [MOVED FROM BAD 046/134] basevideo, vp8: guard unstable API with GST_USE_UNSTABLE_API
Add some guards and fat warnings to the header files with still unstable
API, so people who just look at the installed headers know that it
actually is unstable API.

Merging previous commit into current codebase.
2012-09-16 15:27:14 +02:00
David Schleef
b8cd2ec0f6 [MOVED FROM BAD 045/134] basevideo: Move base video from vp8 to gst-libs 2012-09-16 15:27:14 +02:00
Sebastian Dröge
2014a7eee2 [MOVED FROM BAD 044/134] vp8: Use VPX_PLANE_* instead of PLANE_* 2012-09-16 15:27:14 +02:00
Sebastian Dröge
b9e31e22c6 [MOVED FROM BAD 043/134] vp8: Add compatilibity defines to work with older versions of libvpx too 2012-09-16 15:27:14 +02:00
Philip Jägenstedt
70305a8088 [MOVED FROM BAD 042/134] vp8dec: s/IMG_FMT_I420/VPX_IMG_FMT_I420/
This corresponds to upstream libvpx commit 6cd4a10e167203d1deb79abf60ee72599e97891b
2012-09-16 15:27:14 +02:00
Sebastian Dröge
0de97a2e85 [MOVED FROM BAD 041/134] vp8enc: Allow a maximum keyframe distance of 0, i.e. all frames are keyframes 2012-09-16 15:27:14 +02:00