This should speed up standard Vorbis encoding and decoding pipelines a bit.
Thanks to David Schleef for the assistance to get the ORC code right
and explaining everything.
Use foo_LDADD instead of foo_LDFLAGS to specify the libraries to link to.
This should make sure arguments are passed to the linker in the right
order, and makes LDFLAGS usable again.
Based on initial patch by Brian Cameron <brian.cameron@oracle.com>
Fixes#615697.
Original commit message from CVS:
* gst/audioconvert/audioconvert.c: (audio_convert_convert):
Prevent overflows with big buffer when calculating the size of
the intermediate buffer by using gst_util_uint64_scale() instead of
plain arithmetics. Fixes bug #552801.
Original commit message from CVS:
* gst/audioconvert/audioconvert.h:
* gst/audioconvert/gstaudioquantize.c:
(gst_audio_quantize_setup_dither),
(gst_audio_quantize_free_dither):
* gst/audioconvert/gstfastrandom.h:
Implement a linear congruential generator as pseudo random number
generator for the dither noise. This is about 2 times faster than
using GLib's mersenne twister. Also this uses only integer math for
generating integers while GLib internally uses floating point math.
Original commit message from CVS:
* gst/audioconvert/gstaudioconvert.c:
(structure_has_fixed_channel_positions),
(gst_audio_convert_transform_caps):
Allow up to 11 positioned channels now that audioconvert can handle
this but add no default positions for > 8 channels.
* tests/check/elements/audioconvert.c: (GST_START_TEST):
Add some unit tests for the above change: Test conversion of
11 positioned channels to stereo and the other way around, test
conversion of 15 unpositioned channels in different ways.
Original commit message from CVS:
* gst/audioconvert/gstchannelmix.c:
(gst_channel_mix_fill_one_other):
If mixing left or right to center (or the other way around) only take
the complete value if we don't already have the original position in
the source.
Original commit message from CVS:
* gst-libs/gst/audio/multichannel.c:
(gst_audio_check_channel_positions),
(gst_audio_set_structure_channel_positions_list),
(gst_audio_fixate_channel_positions):
Allow rear center together with rear left/right and other previously
conflicting channel positions. The reason why they weren't allowed
was the channel mixing implementation in audioconvert.
Also take this into account when fixing channel layouts.
Allow setting channel positions for 1/2 channels when using
gst_audio_set_structure_channel_position().
* gst/audioconvert/gstchannelmix.c:
(gst_channel_mix_fill_compatible), (gst_channel_mix_detect_pos),
(gst_channel_mix_fill_one_other), (gst_channel_mix_fill_others),
(gst_channel_mix_fill_special), (gst_channel_mix_fill_matrix):
Major rewrite of the channel mixing.
We now allow previously conflicting channel positions to appear
together (rear center and rear left/right for example).
Fixes bug #533817.
Rework the way channels are mixed together to take more possible
channel positions into account, properly mix from/to side channels
and don't assume that either center, left&right or nothing of a
specific position is available anymore.
* tests/check/elements/audioconvert.c: (GST_START_TEST):
Adjust unit tests with non-standard 1/2 channel layouts to the more
correct new behaviour.
Add a unit test for 5.1->Stereo downmixing.
Original commit message from CVS:
* gst/audioconvert/gstchannelmix.c: (gst_channel_mix_fill_others):
Add support for mixing GST_AUDIO_CHANNEL_POSITION_SIDE_* from/to
other channel positions when source has SIDE channels and dest doesn't
or the other way around.
Original commit message from CVS:
* gst/audioconvert/gstchannelmix.c:
(gst_channel_mix_fill_normalize):
Prevent division by zero if the channel mix matrix contains only
zeroes.
Original commit message from CVS:
* gst/audioconvert/gstaudioconvert.c:
(gst_audio_convert_fixate_channels):
Passthrough the channel positions if the number of output channels is
the same as the number of input channels, the input had a channel
layout and downstream requests no special one. We did this already for
> 2 channels but now it's also done for 1 channel. Fixes bug #533617.
Original commit message from CVS:
* gst/audioconvert/gstaudioconvert.c:
(gst_audio_convert_fixate_channels):
Correctly set the default channel positions when converting to 8
channels.
Original commit message from CVS:
Based on a patch by: Tim-Philipp Müller <tim.muller at collabora co uk>
* gst/audioconvert/audioconvert.c: (audio_convert_prepare_context):
* gst/audioconvert/audioconvert.h:
* gst/audioconvert/gstaudioconvert.c:
(gst_audio_convert_parse_caps),
(structure_has_fixed_channel_positions),
(gst_audio_convert_transform_caps):
* gst/audioconvert/gstchannelmix.c: (gst_channel_mix_fill_matrix):
Add support for more than 8 channels and NONE channel layouts. For
more than 8 channels no channel conversion is supported yet, only
format conversions are supported. Fixes bug #398033.
* tests/check/elements/audioconvert.c: (verify_convert),
(GST_START_TEST), (audioconvert_suite):
Add some unit tests by Tim for checking the NONE channel layouts
and more than 8 channels and add some more unit tests for channel
conversions.
Original commit message from CVS:
* gst/audioconvert/gstaudioconvert.c: (gst_audio_convert_init),
(gst_audio_convert_create_silence_buffer),
(gst_audio_convert_transform):
Make audioconvert GAP-aware by outputting silence buffers when the
input has the GAP flag set. This is up to 8x faster.
Based on a patch by Stefan Kost. Fixes bug #517813.
Original commit message from CVS:
* gst/audioconvert/gstaudioconvert.c: (find_suitable_channel_layout),
(gst_audio_convert_fixate_channels), (gst_audio_convert_fixate_caps):
Preserve channel layout when fixating the number of channels in the
output caps, or make sure there's a suitable channel position layout
set on the caps if required. Fixes#430677.
Original commit message from CVS:
* ext/alsa/Makefile.am:
There is no GST_PLUGINS_BASE_LIBS defined.
* ext/alsa/gstalsa.c:
* ext/alsa/gstalsasink.c: (gst_alsasink_delay):
* ext/alsa/gstalsasrc.c: (gst_alsasrc_delay):
Add support for ALSA 24-bit formats.
snd_pcm_delay can return an error code, especially
during XRUNS. In that case, the best we can do is assume
delay = 0.
* gst/audioconvert/Makefile.am:
Add flags from -base before any more-remote dependencies.
Original commit message from CVS:
Patch by: Jens Granseuer <jensgr at gmx dot net>
* gst/audioconvert/gstaudioquantize.c:
Fix C89 incompatibilities and spelling of explanations. Fixes#463215.
Original commit message from CVS:
* gst/audioconvert/Makefile.am:
* gst/audioconvert/audioconvert.c: (audio_convert_get_func_index),
(check_default), (audio_convert_prepare_context),
(audio_convert_clean_context), (audio_convert_convert):
* gst/audioconvert/audioconvert.h:
* gst/audioconvert/gstaudioconvert.c:
(gst_audio_convert_dithering_get_type),
(gst_audio_convert_ns_get_type), (gst_audio_convert_class_init),
(gst_audio_convert_init), (gst_audio_convert_set_caps),
(gst_audio_convert_set_property), (gst_audio_convert_get_property):
* gst/audioconvert/gstaudioconvert.h:
* gst/audioconvert/gstaudioquantize.c:
(gst_audio_quantize_setup_noise_shaping),
(gst_audio_quantize_free_noise_shaping),
(gst_audio_quantize_setup_dither),
(gst_audio_quantize_free_dither),
(gst_audio_quantize_setup_quantize_func),
(gst_audio_quantize_setup), (gst_audio_quantize_free):
* gst/audioconvert/gstaudioquantize.h:
Implement dithering and noise shaping in audioconvert. By default now
TPDF dithering (and no noise shaping) will be used when converting
from a higher bit depth to 20 bit depth or smaller, otherwise
everything will be as it is now.
For the last audioconvert in a pipeline it would make sense to
use some kind of noise shaping, enabling it by default for all
conversions would give undesired results though. Fixes#360246.
* tests/check/elements/audioconvert.c: (setup_audioconvert),
(GST_START_TEST):
Adjust unit test for the new audioconvert.
Original commit message from CVS:
* gst/audioconvert/gstaudioconvert.c: (gst_audio_convert_init):
Initalize the AudioConvertCtx with zeroes, otherwise it will contain
pointers to random memory which are passed to g_free() when
audio_convert_prepare_context() is called the first time.
Original commit message from CVS:
Patch by: René Stadler <mail at renestadler dot de>
with some minor changes
* gst-libs/gst/floatcast/floatcast.h:
Use more efficient float endianness conversion functions that don't
involve 2 function calls per value.
* gst/audioconvert/audioconvert.c: (audio_convert_get_func_index),
(check_default), (audio_convert_prepare_context):
* gst/audioconvert/gstaudioconvert.c:
(gst_audio_convert_parse_caps), (make_lossless_changes):
Support non-native endianness floats as input and output.
Fixes#339838.
* tests/check/elements/audioconvert.c: (verify_convert),
(GST_START_TEST):
Add unit tests for the non-native endianness float conversions.
Original commit message from CVS:
* gst/audioconvert/audioconvert.c:
Add docs to the integer pack functions and implement proper
rounding. Before we had rounding towards negative infinity, i.e.
always the smaller number was taken. Now we use natural rounding,
i.e. rounding to the nearest integer and to the one with the largest
absolute value for X.5. The old rounding introduced some minor
distortions. Fixes#420079
* tests/check/elements/audioconvert.c: (GST_START_TEST):
Fix one unit test that assumed the old rounding and added unit tests
for checking signed/unsigned int16 <-> signed/unsigned int16 with
depth 8, one for signed int16 <-> unsigned int16 and one for the new
rounding from signed int32 to signed/unsigned int16.
