mirrors/gstreamer
mirrors/gstreamer
1
1
Fork 0
mirror of https://gitlab.freedesktop.org/gstreamer/gstreamer.git synced 2024-11-06 09:29:42 +00:00
Code Issues 1 Projects Releases Wiki Activity
gstreamer/subprojects/gst-plugins-base/gst-libs/gst/audio/audio-resampler-x86.h

72 lines
2.7 KiB
C
Raw Normal View History

audio-converter: add resampler Add a resampler to the processing chain when needed. port the audio resampler to the new audioconverter library
2015-11-13 14:32:29 +00:00
/* GStreamer
audio-resampler: reorder filter coefficients for more speed Reorder the filter coefficients to make it easier to use SIMD for interpolation. Fix orc flags a little. Add specialized nearest resampling function.
2016-02-19 15:39:43 +00:00
* Copyright (C) <2016> Wim Taymans <wim.taymans@gmail.com>
audio-converter: add resampler Add a resampler to the processing chain when needed. port the audio resampler to the new audioconverter library
2015-11-13 14:32:29 +00:00
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Library General Public
* License as published by the Free Software Foundation; either
* version 2 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Library General Public License for more details.
*
* You should have received a copy of the GNU Library General Public
* License along with this library; if not, write to the
* Free Software Foundation, Inc., 51 Franklin St, Fifth Floor,
* Boston, MA 02110-1301, USA.
*/
audioresample: Separate out CFLAGS used for SSE* code This makes sure that we only build files that need explicit SIMD support with the relevant CFLAGS. This allows the rest of the code to be built without, and specific SSE* code is only called after runtime checks for CPU features. https://bugzilla.gnome.org/show_bug.cgi?id=729276
2016-09-28 12:07:38 +00:00
#include "audio-resampler-macros.h"
#include "audio-resampler-x86-sse.h"
#include "audio-resampler-x86-sse2.h"
#include "audio-resampler-x86-sse41.h"
audio-resampler: Fix compilation of intrinsics Only compile intrinsics when we are building for the selected architecture. Add sse4.1 optimized int32 resampler code.
2016-01-15 11:45:47 +00:00
audio-resampler: make pluggable optimized functions Add support for x86 specialized functions and select them at runtime.
2016-01-12 17:55:19 +00:00
static void
audio-resampler: reorder filter coefficients for more speed Reorder the filter coefficients to make it easier to use SIMD for interpolation. Fix orc flags a little. Add specialized nearest resampling function.
2016-02-19 15:39:43 +00:00
audio_resampler_check_x86 (const gchar *option)
audio-resampler: make pluggable optimized functions Add support for x86 specialized functions and select them at runtime.
2016-01-12 17:55:19 +00:00
{
audio-resampler: reorder filter coefficients for more speed Reorder the filter coefficients to make it easier to use SIMD for interpolation. Fix orc flags a little. Add specialized nearest resampling function.
2016-02-19 15:39:43 +00:00
if (!strcmp (option, "sse")) {
audioresample: Separate out CFLAGS used for SSE* code This makes sure that we only build files that need explicit SIMD support with the relevant CFLAGS. This allows the rest of the code to be built without, and specific SSE* code is only called after runtime checks for CPU features. https://bugzilla.gnome.org/show_bug.cgi?id=729276
2016-09-28 12:07:38 +00:00
#if defined (HAVE_XMMINTRIN_H) && HAVE_SSE
audio-resampler: optimizations Improve int16 resampling by using pmaddwd Use intrinsics to scale and pack int16 samples Align the coefficients so that we can use aligned loads Add padding to taps and samples so that we don't have to use partial loads for the remainder of the loops. Remove copy_n, we can reuse the plain copy function with some new parameters. Align and pad the sample array.
2016-01-13 16:44:39 +00:00
GST_DEBUG ("enable SSE optimisations");
audio-resampler: remove stereo optimizations The stereo optimizations don't give enough benefit. Rename none to full to make it clear that we use a full filter instead of an interpolated one
2016-02-19 09:40:03 +00:00
resample_gfloat_full_1 = resample_gfloat_full_1_sse;
audio-resampler: add linear interpolation method Make more functions into macros. Add linear interpolation of filter coefficients.
2016-02-10 11:04:12 +00:00
resample_gfloat_linear_1 = resample_gfloat_linear_1_sse;
audio-resampler: add cubic interpolation
2016-02-10 16:28:24 +00:00
resample_gfloat_cubic_1 = resample_gfloat_cubic_1_sse;
audio-resampler: avoid some format conversion Store the filter in the desired sample format so that we can simply do a linear or cubic interpolation to get the new filter instead of having to go through gdouble and then convert.
2016-02-22 12:19:02 +00:00
interpolate_gfloat_linear = interpolate_gfloat_linear_sse;
interpolate_gfloat_cubic = interpolate_gfloat_cubic_sse;
audio-resampler: add more optimizations
2016-02-11 16:40:56 +00:00
#else
GST_DEBUG ("SSE optimisations not enabled");
audio-resampler: Fix compilation of intrinsics Only compile intrinsics when we are building for the selected architecture. Add sse4.1 optimized int32 resampler code.
2016-01-15 11:45:47 +00:00
#endif
audio-resampler: reorder filter coefficients for more speed Reorder the filter coefficients to make it easier to use SIMD for interpolation. Fix orc flags a little. Add specialized nearest resampling function.
2016-02-19 15:39:43 +00:00
} else if (!strcmp (option, "sse2")) {
audioresample: Separate out CFLAGS used for SSE* code This makes sure that we only build files that need explicit SIMD support with the relevant CFLAGS. This allows the rest of the code to be built without, and specific SSE* code is only called after runtime checks for CPU features. https://bugzilla.gnome.org/show_bug.cgi?id=729276
2016-09-28 12:07:38 +00:00
#if defined (HAVE_EMMINTRIN_H) && HAVE_SSE2
audio-resampler: optimizations Improve int16 resampling by using pmaddwd Use intrinsics to scale and pack int16 samples Align the coefficients so that we can use aligned loads Add padding to taps and samples so that we don't have to use partial loads for the remainder of the loops. Remove copy_n, we can reuse the plain copy function with some new parameters. Align and pad the sample array.
2016-01-13 16:44:39 +00:00
GST_DEBUG ("enable SSE2 optimisations");
audio-resampler: remove stereo optimizations The stereo optimizations don't give enough benefit. Rename none to full to make it clear that we use a full filter instead of an interpolated one
2016-02-19 09:40:03 +00:00
resample_gint16_full_1 = resample_gint16_full_1_sse2;
audio-resampler: Improve taps memory layout Rearrange the oversampled taps in memory to make it easier to use SIMD instructions on them. this simplifies some sse code. Add some more optimizations
2016-02-11 10:57:26 +00:00
resample_gint16_linear_1 = resample_gint16_linear_1_sse2;
resample_gint16_cubic_1 = resample_gint16_cubic_1_sse2;
audio-resampler: avoid some format conversion Store the filter in the desired sample format so that we can simply do a linear or cubic interpolation to get the new filter instead of having to go through gdouble and then convert.
2016-02-22 12:19:02 +00:00
interpolate_gint16_linear = interpolate_gint16_linear_sse2;
interpolate_gint16_cubic = interpolate_gint16_cubic_sse2;
audio-resampler: remove stereo optimizations The stereo optimizations don't give enough benefit. Rename none to full to make it clear that we use a full filter instead of an interpolated one
2016-02-19 09:40:03 +00:00
resample_gdouble_full_1 = resample_gdouble_full_1_sse2;
audio-resampler: Improve taps memory layout Rearrange the oversampled taps in memory to make it easier to use SIMD instructions on them. this simplifies some sse code. Add some more optimizations
2016-02-11 10:57:26 +00:00
resample_gdouble_linear_1 = resample_gdouble_linear_1_sse2;
resample_gdouble_cubic_1 = resample_gdouble_cubic_1_sse2;
audio-resampler: improve filter construction Remove some unused variables from the inner product functions. Make filter coefficients by interpolating if required. Rename some fields. Try hard to not recalculate filters when just chaging the rate. Add more proprties to audioresample.
2016-02-15 17:06:19 +00:00
interpolate_gdouble_linear = interpolate_gdouble_linear_sse2;
interpolate_gdouble_cubic = interpolate_gdouble_cubic_sse2;
audio-resampler: add more optimizations
2016-02-11 16:40:56 +00:00
#else
GST_DEBUG ("SSE2 optimisations not enabled");
audio-resampler: optimizations Improve int16 resampling by using pmaddwd Use intrinsics to scale and pack int16 samples Align the coefficients so that we can use aligned loads Add padding to taps and samples so that we don't have to use partial loads for the remainder of the loops. Remove copy_n, we can reuse the plain copy function with some new parameters. Align and pad the sample array.
2016-01-13 16:44:39 +00:00
#endif
audio-resampler: Fix compilation of intrinsics Only compile intrinsics when we are building for the selected architecture. Add sse4.1 optimized int32 resampler code.
2016-01-15 11:45:47 +00:00
} else if (!strcmp (option, "sse41")) {
audiorsample: Fix build on 32-bit x86 Turns out _mm_cvtsi128_si64() isn't available on 32-bit, so only build SSE 4.1 optimisations on x86-64 for now.
2016-09-29 14:15:16 +00:00
#if defined (__x86_64__) && \
defined (HAVE_SMMINTRIN_H) && defined (HAVE_EMMINTRIN_H) && \
HAVE_SSE41
audio-resampler: Fix compilation of intrinsics Only compile intrinsics when we are building for the selected architecture. Add sse4.1 optimized int32 resampler code.
2016-01-15 11:45:47 +00:00
GST_DEBUG ("enable SSE41 optimisations");
audio-resampler: remove stereo optimizations The stereo optimizations don't give enough benefit. Rename none to full to make it clear that we use a full filter instead of an interpolated one
2016-02-19 09:40:03 +00:00
resample_gint32_full_1 = resample_gint32_full_1_sse41;
audio-resampler: add more optimizations
2016-02-11 16:40:56 +00:00
resample_gint32_linear_1 = resample_gint32_linear_1_sse41;
resample_gint32_cubic_1 = resample_gint32_cubic_1_sse41;
#else
GST_DEBUG ("SSE41 optimisations not enabled");
audio-resampler: Fix compilation of intrinsics Only compile intrinsics when we are building for the selected architecture. Add sse4.1 optimized int32 resampler code.
2016-01-15 11:45:47 +00:00
#endif
}
audio-converter: add resampler Add a resampler to the processing chain when needed. port the audio resampler to the new audioconverter library
2015-11-13 14:32:29 +00:00
}
Reference in a new issue Copy permalink