[FFmpeg-cvslog] lavfi: add lavr-based audio resampling filter.

Anton Khirnov git at videolan.org
Sun May 13 00:19:12 CEST 2012


ffmpeg | branch: master | Anton Khirnov <anton at khirnov.net> | Fri May  4 15:09:17 2012 +0200| [d371e7b9885aabdb29c038fa723bc890276aa366] | committer: Anton Khirnov

lavfi: add lavr-based audio resampling filter.

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=d371e7b9885aabdb29c038fa723bc890276aa366
---

 configure                 |    1 +
 doc/filters.texi          |    6 +
 libavfilter/Makefile      |    2 +
 libavfilter/af_resample.c |  225 +++++++++++++++++++++++++++++++++++++++++++++
 libavfilter/allfilters.c  |    1 +
 5 files changed, 235 insertions(+), 0 deletions(-)

diff --git a/configure b/configure
index 79b948e..f6cb0e2 100755
--- a/configure
+++ b/configure
@@ -1531,6 +1531,7 @@ frei0r_filter_extralibs='$ldl'
 frei0r_src_filter_deps="frei0r dlopen strtok_r"
 frei0r_src_filter_extralibs='$ldl'
 hqdn3d_filter_deps="gpl"
+resample_filter_deps="avresample"
 ocv_filter_deps="libopencv"
 yadif_filter_deps="gpl"
 
diff --git a/doc/filters.texi b/doc/filters.texi
index dbcc86a..8eff84a 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -111,6 +111,12 @@ Below is a description of the currently available audio filters.
 
 Pass the audio source unchanged to the output.
 
+ at section resample
+Convert the audio sample format, sample rate and channel layout. This filter is
+not meant to be used directly, it is inserted automatically by libavfilter
+whenever conversion is needed. Use the @var{aformat} filter to force a specific
+conversion.
+
 @c man end AUDIO FILTERS
 
 @chapter Audio Sources
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index 49a47d3..9cbb908 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -1,6 +1,7 @@
 NAME = avfilter
 FFLIBS = avutil swscale
 FFLIBS-$(CONFIG_MOVIE_FILTER) += avformat avcodec
+FFLIBS-$(CONFIG_RESAMPLE_FILTER) += avresample
 
 HEADERS = avfilter.h                                                    \
           avfiltergraph.h                                               \
@@ -22,6 +23,7 @@ OBJS = allfilters.o                                                     \
        vsrc_buffer.o                                                    \
 
 OBJS-$(CONFIG_ANULL_FILTER)                  += af_anull.o
+OBJS-$(CONFIG_RESAMPLE_FILTER)               += af_resample.o
 
 OBJS-$(CONFIG_ANULLSRC_FILTER)               += asrc_anullsrc.o
 
diff --git a/libavfilter/af_resample.c b/libavfilter/af_resample.c
new file mode 100644
index 0000000..f46e24b
--- /dev/null
+++ b/libavfilter/af_resample.c
@@ -0,0 +1,225 @@
+/*
+ *
+ * This file is part of Libav.
+ *
+ * Libav 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.
+ *
+ * Libav 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 Libav; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/**
+ * @file
+ * sample format and channel layout conversion audio filter
+ */
+
+#include "libavutil/avassert.h"
+#include "libavutil/avstring.h"
+#include "libavutil/mathematics.h"
+#include "libavutil/opt.h"
+
+#include "libavresample/avresample.h"
+
+#include "audio.h"
+#include "avfilter.h"
+#include "internal.h"
+
+typedef struct ResampleContext {
+    AVAudioResampleContext *avr;
+
+    int64_t next_pts;
+} ResampleContext;
+
+static av_cold void uninit(AVFilterContext *ctx)
+{
+    ResampleContext *s = ctx->priv;
+
+    if (s->avr) {
+        avresample_close(s->avr);
+        avresample_free(&s->avr);
+    }
+}
+
+static int query_formats(AVFilterContext *ctx)
+{
+    AVFilterLink *inlink  = ctx->inputs[0];
+    AVFilterLink *outlink = ctx->outputs[0];
+
+    AVFilterFormats        *in_formats      = avfilter_all_formats(AVMEDIA_TYPE_AUDIO);
+    AVFilterFormats        *out_formats     = avfilter_all_formats(AVMEDIA_TYPE_AUDIO);
+
+    avfilter_formats_ref(in_formats,  &inlink->out_formats);
+    avfilter_formats_ref(out_formats, &outlink->in_formats);
+
+    return 0;
+}
+
+static int config_output(AVFilterLink *outlink)
+{
+    AVFilterContext *ctx = outlink->src;
+    AVFilterLink *inlink = ctx->inputs[0];
+    ResampleContext   *s = ctx->priv;
+    char buf1[64], buf2[64];
+    int ret;
+
+    if (s->avr) {
+        avresample_close(s->avr);
+        avresample_free(&s->avr);
+    }
+
+    if (inlink->channel_layout == outlink->channel_layout &&
+        inlink->sample_rate    == outlink->sample_rate    &&
+        inlink->format         == outlink->format)
+        return 0;
+
+    if (!(s->avr = avresample_alloc_context()))
+        return AVERROR(ENOMEM);
+
+    av_opt_set_int(s->avr,  "in_channel_layout", inlink ->channel_layout, 0);
+    av_opt_set_int(s->avr, "out_channel_layout", outlink->channel_layout, 0);
+    av_opt_set_int(s->avr,  "in_sample_fmt",     inlink ->format,         0);
+    av_opt_set_int(s->avr, "out_sample_fmt",     outlink->format,         0);
+    av_opt_set_int(s->avr,  "in_sample_rate",    inlink ->sample_rate,    0);
+    av_opt_set_int(s->avr, "out_sample_rate",    outlink->sample_rate,    0);
+
+    /* if both the input and output formats are s16 or u8, use s16 as
+       the internal sample format */
+    if (av_get_bytes_per_sample(inlink->format)  <= 2 &&
+        av_get_bytes_per_sample(outlink->format) <= 2)
+        av_opt_set_int(s->avr, "internal_sample_fmt", AV_SAMPLE_FMT_S16P, 0);
+
+    if ((ret = avresample_open(s->avr)) < 0)
+        return ret;
+
+    outlink->time_base = (AVRational){ 1, outlink->sample_rate };
+    s->next_pts        = AV_NOPTS_VALUE;
+
+    av_get_channel_layout_string(buf1, sizeof(buf1),
+                                 -1, inlink ->channel_layout);
+    av_get_channel_layout_string(buf2, sizeof(buf2),
+                                 -1, outlink->channel_layout);
+    av_log(ctx, AV_LOG_VERBOSE,
+           "fmt:%s srate: %d cl:%s -> fmt:%s srate: %d cl:%s\n",
+           av_get_sample_fmt_name(inlink ->format), inlink ->sample_rate, buf1,
+           av_get_sample_fmt_name(outlink->format), outlink->sample_rate, buf2);
+
+    return 0;
+}
+
+static int request_frame(AVFilterLink *outlink)
+{
+    AVFilterContext *ctx = outlink->src;
+    ResampleContext   *s = ctx->priv;
+    int ret = avfilter_request_frame(ctx->inputs[0]);
+
+    /* flush the lavr delay buffer */
+    if (ret == AVERROR_EOF && s->avr) {
+        AVFilterBufferRef *buf;
+        int nb_samples = av_rescale_rnd(avresample_get_delay(s->avr),
+                                        outlink->sample_rate,
+                                        ctx->inputs[0]->sample_rate,
+                                        AV_ROUND_UP);
+
+        if (!nb_samples)
+            return ret;
+
+        buf = ff_get_audio_buffer(outlink, AV_PERM_WRITE, nb_samples);
+        if (!buf)
+            return AVERROR(ENOMEM);
+
+        ret = avresample_convert(s->avr, (void**)buf->extended_data,
+                                 buf->linesize[0], nb_samples,
+                                 NULL, 0, 0);
+        if (ret <= 0) {
+            avfilter_unref_buffer(buf);
+            return (ret == 0) ? AVERROR_EOF : ret;
+        }
+
+        buf->pts = s->next_pts;
+        ff_filter_samples(outlink, buf);
+        return 0;
+    }
+    return ret;
+}
+
+static void filter_samples(AVFilterLink *inlink, AVFilterBufferRef *buf)
+{
+    AVFilterContext  *ctx = inlink->dst;
+    ResampleContext    *s = ctx->priv;
+    AVFilterLink *outlink = ctx->outputs[0];
+
+    if (s->avr) {
+        AVFilterBufferRef *buf_out;
+        int delay, nb_samples, ret;
+
+        /* maximum possible samples lavr can output */
+        delay      = avresample_get_delay(s->avr);
+        nb_samples = av_rescale_rnd(buf->audio->nb_samples + delay,
+                                    outlink->sample_rate, inlink->sample_rate,
+                                    AV_ROUND_UP);
+
+        buf_out = ff_get_audio_buffer(outlink, AV_PERM_WRITE, nb_samples);
+        ret     = avresample_convert(s->avr, (void**)buf_out->extended_data,
+                                     buf_out->linesize[0], nb_samples,
+                                     (void**)buf->extended_data, buf->linesize[0],
+                                     buf->audio->nb_samples);
+
+        av_assert0(!avresample_available(s->avr));
+
+        if (s->next_pts == AV_NOPTS_VALUE) {
+            if (buf->pts == AV_NOPTS_VALUE) {
+                av_log(ctx, AV_LOG_WARNING, "First timestamp is missing, "
+                       "assuming 0.\n");
+                s->next_pts = 0;
+            } else
+                s->next_pts = av_rescale_q(buf->pts, inlink->time_base,
+                                           outlink->time_base);
+        }
+
+        if (ret > 0) {
+            buf_out->audio->nb_samples = ret;
+            if (buf->pts != AV_NOPTS_VALUE) {
+                buf_out->pts = av_rescale_q(buf->pts, inlink->time_base,
+                                            outlink->time_base) -
+                               av_rescale(delay, outlink->sample_rate,
+                                          inlink->sample_rate);
+            } else
+                buf_out->pts = s->next_pts;
+
+            s->next_pts = buf_out->pts + buf_out->audio->nb_samples;
+
+            ff_filter_samples(outlink, buf_out);
+        }
+        avfilter_unref_buffer(buf);
+    } else
+        ff_filter_samples(outlink, buf);
+}
+
+AVFilter avfilter_af_resample = {
+    .name          = "resample",
+    .description   = NULL_IF_CONFIG_SMALL("Audio resampling and conversion."),
+    .priv_size     = sizeof(ResampleContext),
+
+    .uninit         = uninit,
+    .query_formats  = query_formats,
+
+    .inputs    = (const AVFilterPad[]) {{ .name            = "default",
+                                          .type            = AVMEDIA_TYPE_AUDIO,
+                                          .filter_samples  = filter_samples,
+                                          .min_perms       = AV_PERM_READ },
+                                        { .name = NULL}},
+    .outputs   = (const AVFilterPad[]) {{ .name          = "default",
+                                          .type          = AVMEDIA_TYPE_AUDIO,
+                                          .config_props  = config_output,
+                                          .request_frame = request_frame },
+                                        { .name = NULL}},
+};
diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
index f887002..66d890f 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -35,6 +35,7 @@ void avfilter_register_all(void)
     initialized = 1;
 
     REGISTER_FILTER (ANULL,       anull,       af);
+    REGISTER_FILTER (RESAMPLE,    resample,    af);
 
     REGISTER_FILTER (ANULLSRC,    anullsrc,    asrc);
 



More information about the ffmpeg-cvslog mailing list