[FFmpeg-cvslog] swr: Add SOX resampler support
Rob Sykes
git at videolan.org
Tue Dec 11 22:17:28 CET 2012
ffmpeg | branch: master | Rob Sykes <aquegg at yahoo.co.uk> | Tue Dec 11 21:43:42 2012 +0100| [801b315729450b30d1ec3e8a9b7862b3c41146be] | committer: Michael Niedermayer
swr: Add SOX resampler support
Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=801b315729450b30d1ec3e8a9b7862b3c41146be
---
configure | 4 ++
libswresample/Makefile | 2 +
libswresample/resample.c | 3 +-
libswresample/soxr_resample.c | 89 +++++++++++++++++++++++++++++++++++
libswresample/swresample.c | 11 ++++-
libswresample/swresample.h | 1 +
libswresample/swresample_internal.h | 4 +-
7 files changed, 110 insertions(+), 4 deletions(-)
diff --git a/configure b/configure
index 7cf8d9d..7f3202e 100755
--- a/configure
+++ b/configure
@@ -213,6 +213,7 @@ External library support:
--enable-libpulse enable Pulseaudio input via libpulse [no]
--enable-librtmp enable RTMP[E] support via librtmp [no]
--enable-libschroedinger enable Dirac de/encoding via libschroedinger [no]
+ --enable-libsoxr enable Include libsoxr resampling [no]
--enable-libspeex enable Speex de/encoding via libspeex [no]
--enable-libstagefright-h264 enable H.264 decoding via libstagefright [no]
--enable-libtheora enable Theora encoding via libtheora [no]
@@ -1173,6 +1174,7 @@ CONFIG_LIST="
libpulse
librtmp
libschroedinger
+ libsoxr
libspeex
libstagefright_h264
libtheora
@@ -3839,6 +3841,7 @@ enabled libopus && require_pkg_config opus opus_multistream.h opus_multistrea
enabled libpulse && require_pkg_config libpulse-simple pulse/simple.h pa_simple_new
enabled librtmp && require_pkg_config librtmp librtmp/rtmp.h RTMP_Socket
enabled libschroedinger && require_pkg_config schroedinger-1.0 schroedinger/schro.h schro_init
+enabled libsoxr && require libsoxr soxr.h soxr_create -lsoxr
enabled libspeex && require libspeex speex/speex.h speex_decoder_init -lspeex
enabled libstagefright_h264 && require_cpp libstagefright_h264 "binder/ProcessState.h media/stagefright/MetaData.h
media/stagefright/MediaBufferGroup.h media/stagefright/MediaDebug.h media/stagefright/MediaDefs.h
@@ -4254,6 +4257,7 @@ echo "libopus enabled ${libopus-no}"
echo "libpulse enabled ${libpulse-no}"
echo "librtmp enabled ${librtmp-no}"
echo "libschroedinger enabled ${libschroedinger-no}"
+echo "libsoxr enabled ${libsoxr-no}"
echo "libspeex enabled ${libspeex-no}"
echo "libstagefright-h264 enabled ${libstagefright_h264-no}"
echo "libtheora enabled ${libtheora-no}"
diff --git a/libswresample/Makefile b/libswresample/Makefile
index 4c3ec1c..9d9f10c 100644
--- a/libswresample/Makefile
+++ b/libswresample/Makefile
@@ -13,4 +13,6 @@ OBJS = audioconvert.o \
resample.o \
swresample.o \
+OBJS-$(CONFIG_LIBSOXR) += soxr_resample.o
+
TESTPROGS = swresample
diff --git a/libswresample/resample.c b/libswresample/resample.c
index 7256fcf..60ba0e9 100644
--- a/libswresample/resample.c
+++ b/libswresample/resample.c
@@ -196,7 +196,8 @@ static int build_filter(ResampleContext *c, void *filter, double factor, int tap
}
static ResampleContext *resample_init(ResampleContext *c, int out_rate, int in_rate, int filter_size, int phase_shift, int linear,
- double cutoff0, enum AVSampleFormat format, enum SwrFilterType filter_type, int kaiser_beta){
+ double cutoff0, enum AVSampleFormat format, enum SwrFilterType filter_type, int kaiser_beta,
+ double precision, int cheby){
double cutoff = cutoff0? cutoff0 : 0.8;
double factor= FFMIN(out_rate * cutoff / in_rate, 1.0);
int phase_count= 1<<phase_shift;
diff --git a/libswresample/soxr_resample.c b/libswresample/soxr_resample.c
new file mode 100644
index 0000000..0e02fcf
--- /dev/null
+++ b/libswresample/soxr_resample.c
@@ -0,0 +1,89 @@
+/*
+ * audio resampling with soxr
+ * Copyright (c) 2012 Rob Sykes <aquegg at yahoo.co.uk>
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg 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
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file
+ * audio resampling with soxr
+ */
+
+#include "libavutil/log.h"
+#include "swresample_internal.h"
+
+#include <soxr.h>
+
+static struct ResampleContext *create(struct ResampleContext *c, int out_rate, int in_rate, int filter_size, int phase_shift, int linear,
+ double cutoff, enum AVSampleFormat format, enum SwrFilterType filter_type, int kaiser_beta, double precision, int cheby){
+ soxr_error_t error;
+
+ soxr_datatype_t type =
+ format == AV_SAMPLE_FMT_S16P? SOXR_INT16_S :
+ format == AV_SAMPLE_FMT_S16 ? SOXR_INT16_I :
+ format == AV_SAMPLE_FMT_S32P? SOXR_INT32_S :
+ format == AV_SAMPLE_FMT_S32 ? SOXR_INT32_I :
+ format == AV_SAMPLE_FMT_FLTP? SOXR_FLOAT32_S :
+ format == AV_SAMPLE_FMT_FLT ? SOXR_FLOAT32_I :
+ format == AV_SAMPLE_FMT_DBLP? SOXR_FLOAT64_S :
+ format == AV_SAMPLE_FMT_DBL ? SOXR_FLOAT64_I : (soxr_datatype_t)-1;
+
+ soxr_io_spec_t io_spec = soxr_io_spec(type, type);
+
+ soxr_quality_spec_t q_spec = soxr_quality_spec((int)((precision-2)/4), (SOXR_HI_PREC_CLOCK|SOXR_ROLLOFF_NONE)*!!cheby);
+ q_spec.bits = linear? 0 : precision;
+ q_spec.bw_pc = cutoff? FFMAX(FFMIN(cutoff,.995),.8)*100 : q_spec.bw_pc;
+
+ soxr_delete((soxr_t)c);
+ c = (struct ResampleContext *)
+ soxr_create(in_rate, out_rate, 0, &error, &io_spec, &q_spec, 0);
+ if (!c)
+ av_log(NULL, AV_LOG_ERROR, "soxr_create: %s\n", error);
+ return c;
+}
+
+static void destroy(struct ResampleContext * *c){
+ soxr_delete((soxr_t)*c);
+ *c = NULL;
+}
+
+static int flush(struct SwrContext *s){
+ soxr_process((soxr_t)s->resample, NULL, 0, NULL, NULL, 0, NULL);
+ return 0;
+}
+
+static int process(
+ struct ResampleContext * c, AudioData *dst, int dst_size,
+ AudioData *src, int src_size, int *consumed){
+ size_t idone, odone;
+ soxr_error_t error = soxr_set_error((soxr_t)c, soxr_set_num_channels((soxr_t)c, src->ch_count));
+ error = soxr_process((soxr_t)c, src->ch, (size_t)src_size,
+ &idone, dst->ch, (size_t)dst_size, &odone);
+ *consumed = (int)idone;
+ return error? -1 : odone;
+}
+
+static int64_t get_delay(struct SwrContext *s, int64_t base){
+ double delay_s = soxr_delay((soxr_t)s->resample) / s->out_sample_rate;
+ return (int64_t)(delay_s * base + .5);
+}
+
+struct Resampler const soxr_resampler={
+ create, destroy, process, flush, NULL /* set_compensation */, get_delay,
+};
+
diff --git a/libswresample/swresample.c b/libswresample/swresample.c
index 207abb4..af983a1 100644
--- a/libswresample/swresample.c
+++ b/libswresample/swresample.c
@@ -86,6 +86,9 @@ static const AVOption options[]={
{"cutoff" , "set cutoff frequency ratio" , OFFSET(cutoff) , AV_OPT_TYPE_DOUBLE,{.dbl=0. }, 0 , 1 , PARAM },
{"resampler" , "set resampling Engine" , OFFSET(engine) , AV_OPT_TYPE_INT , {.i64=0 }, 0 , SWR_ENGINE_NB-1, PARAM, "resampler"},
{"swr" , "select SW Resampler" , 0 , AV_OPT_TYPE_CONST, {.i64=SWR_ENGINE_SWR }, INT_MIN, INT_MAX , PARAM, "resampler"},
+{"soxr" , "select SoX Resampler" , 0 , AV_OPT_TYPE_CONST, {.i64=SWR_ENGINE_SOXR }, INT_MIN, INT_MAX , PARAM, "resampler"},
+{"precision" , "set resampling precision" , OFFSET(precision) , AV_OPT_TYPE_DOUBLE,{.dbl=20.0 }, 15.0 , 33.0 , PARAM },
+{"cheby" , "enable Chebyshev passband" , OFFSET(cheby) , AV_OPT_TYPE_INT , {.i64=0 }, 0 , 1 , PARAM },
{"min_comp" , "set minimum difference between timestamps and audio data (in seconds) below which no timestamp compensation of either kind is applied"
, OFFSET(min_compensation),AV_OPT_TYPE_FLOAT ,{.dbl=FLT_MAX }, 0 , FLT_MAX , PARAM },
{"min_hard_comp" , "set minimum difference between timestamps and audio data (in seconds) to trigger padding/trimming the data."
@@ -262,6 +265,10 @@ av_cold int swr_init(struct SwrContext *s){
}
switch(s->engine){
+#if CONFIG_LIBSOXR
+ extern struct Resampler const soxr_resampler;
+ case SWR_ENGINE_SOXR: s->resampler = &soxr_resampler; break;
+#endif
case SWR_ENGINE_SWR : s->resampler = &swri_resampler; break;
default:
av_log(s, AV_LOG_ERROR, "Requested resampling engine is unavailable\n");
@@ -272,7 +279,7 @@ av_cold int swr_init(struct SwrContext *s){
set_audiodata_fmt(&s->out, s->out_sample_fmt);
if (s->out_sample_rate!=s->in_sample_rate || (s->flags & SWR_FLAG_RESAMPLE)){
- s->resample = s->resampler->init(s->resample, s->out_sample_rate, s->in_sample_rate, s->filter_size, s->phase_shift, s->linear_interp, s->cutoff, s->int_sample_fmt, s->filter_type, s->kaiser_beta);
+ s->resample = s->resampler->init(s->resample, s->out_sample_rate, s->in_sample_rate, s->filter_size, s->phase_shift, s->linear_interp, s->cutoff, s->int_sample_fmt, s->filter_type, s->kaiser_beta, s->precision, s->cheby);
}else
s->resampler->free(&s->resample);
if( s->int_sample_fmt != AV_SAMPLE_FMT_S16P
@@ -491,7 +498,7 @@ static int resample(SwrContext *s, AudioData *out_param, int out_count,
}
}
- if(in_count && !s->in_buffer_count){
+ if((s->flushed || in_count) && !s->in_buffer_count){
s->in_buffer_index=0;
ret= s->resampler->multiple_resample(s->resample, &out, out_count, &in, in_count, &consumed);
out_count -= ret;
diff --git a/libswresample/swresample.h b/libswresample/swresample.h
index 356fb61..a2b5cd0 100644
--- a/libswresample/swresample.h
+++ b/libswresample/swresample.h
@@ -117,6 +117,7 @@ enum SwrDitherType {
/** Resampling Engines */
enum SwrEngine {
SWR_ENGINE_SWR, /**< SW Resampler */
+ SWR_ENGINE_SOXR, /**< SoX Resampler */
SWR_ENGINE_NB, ///< not part of API/ABI
};
diff --git a/libswresample/swresample_internal.h b/libswresample/swresample_internal.h
index bc53e6a..19ca0d0 100644
--- a/libswresample/swresample_internal.h
+++ b/libswresample/swresample_internal.h
@@ -77,6 +77,8 @@ struct SwrContext {
double cutoff; /**< resampling cutoff frequency. 1.0 corresponds to half the output sample rate */
enum SwrFilterType filter_type; /**< resampling filter type */
int kaiser_beta; /**< beta value for Kaiser window (only applicable if filter_type == AV_FILTER_TYPE_KAISER) */
+ double precision; /**< resampling precision (in bits) */
+ int cheby; /**< if 1 then the resampling FIR filter will be configured for maximal passband flatness */
float min_compensation; ///< minimum below which no compensation will happen
float min_hard_compensation; ///< minimum below which no silence inject / sample drop will happen
@@ -125,7 +127,7 @@ struct SwrContext {
};
typedef struct ResampleContext * (* resample_init_func)(struct ResampleContext *c, int out_rate, int in_rate, int filter_size, int phase_shift, int linear,
- double cutoff, enum AVSampleFormat format, enum SwrFilterType filter_type, int kaiser_beta);
+ double cutoff, enum AVSampleFormat format, enum SwrFilterType filter_type, int kaiser_beta, double precision, int cheby);
typedef void (* resample_free_func)(struct ResampleContext **c);
typedef int (* multiple_resample_func)(struct ResampleContext *c, AudioData *dst, int dst_size, AudioData *src, int src_size, int *consumed);
typedef int (* resample_flush_func)(struct SwrContext *c);
More information about the ffmpeg-cvslog
mailing list