Add docs for the internal audioconvert object before moving it to the
audio library.
Remove get_sizes and implement the trivial logic in the element.
Remove some unused orc functions
Move the audio quantize code from audioconvert to the audio library.
work on making an audio converter helper function similar to the video
converter.
Fold fastrandom directly into the quantizer, add some ORC code to
optimize this later.
Rewrite audioconvert to try to make it more clear what steps are
executed during conversion.
Add passthrough step that just does a memcpy when possible.
Add ORC optimized dither and quantization functions.
Implement noise-shaping on S32 samples only and allow for arbitrary
noise shaping coefficients if we want this later.
Rework the converter to use the pack/unpack functions
Because the unpack functions can only unpack to 1 format, add a separate
conversion step for doubles when the unpack function produces int.
Do conversion to S32 in the quantize function directly.
Tweak the conversion factor for doing float->int conversion slightly to
get the full range of negative samples, use clamp to make sure we don't
exceed our int range on the positive axis (see also #755301)
The patch calculates a second channel mixing matrix from the current one. The
matrix contains the original values * (2^10) as integers. This matrix is used
when integer-formatted channels are mixed.
On a ARM Cortex-A8, single core, 800MHz this improves performance in a
testcase from 29s to 9s for downmixing 6 channels to stereo.
https://bugzilla.gnome.org/show_bug.cgi?id=747005
Rework the audio caps similar to the video caps. Remove
width/depth/endianness/signed fields and replace with a simple string
format and media type audio/x-raw.
Create a GstAudioInfo and some helper methods to parse caps.
Remove duplicate code from the ringbuffer and replace with audio info.
Use AudioInfo in the base audio filter class.
Port elements to new API.
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:
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/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/audioconvert.c: (float), (double), (float_hq),
(double_hq), (audio_convert_get_func_index),
(audio_convert_prepare_context), (audio_convert_convert):
* gst/audioconvert/audioconvert.h:
* gst/audioconvert/gstchannelmix.c: (gst_channel_mix_setup_matrix),
(gst_channel_mix_mix_int), (gst_channel_mix_mix_float):
* gst/audioconvert/gstchannelmix.h:
* tests/check/elements/audioconvert.c: (GST_START_TEST):
Add float as an intermediate format, as well as float mixing. Enable
test that was failing before. Fixes#339837
Original commit message from CVS:
* gst/audioconvert/audioconvert.h:
* gst/audioconvert/gstchannelmix.c: (gst_channel_mix_unset_matrix),
(gst_channel_mix_setup_matrix), (gst_channel_mix_mix):
Alloc temp storage somewhere else where we can do it more
portable.
Original commit message from CVS:
* gst/audioconvert/audioconvert.c: (float),
(audio_convert_prepare_context), (audio_convert_convert):
* gst/audioconvert/audioconvert.h:
Cleanups, speedups, simplifications, added back support
for 24 bits.
