@@ -54,6 +54,7 @@ OBJS-$(CONFIG_AMERGE_FILTER) += af_amerge.o
OBJS-$(CONFIG_AMETADATA_FILTER) += f_metadata.o
OBJS-$(CONFIG_AMIX_FILTER) += af_amix.o
OBJS-$(CONFIG_ANEQUALIZER_FILTER) += af_anequalizer.o
+OBJS-$(CONFIG_ANLMEANS_FILTER) += af_anlmeans.o
OBJS-$(CONFIG_ANULL_FILTER) += af_anull.o
OBJS-$(CONFIG_APAD_FILTER) += af_apad.o
OBJS-$(CONFIG_APERMS_FILTER) += f_perms.o
new file mode 100644
@@ -0,0 +1,282 @@
+/*
+ * Copyright (c) 2018 Paul B Mahol
+ *
+ * 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
+ */
+
+#include <float.h>
+
+#include "libavutil/avassert.h"
+#include "libavutil/audio_fifo.h"
+#include "libavutil/opt.h"
+#include "avfilter.h"
+#include "audio.h"
+#include "formats.h"
+
+#define SQR(x) ((x) * (x))
+
+typedef struct AudioNLMeansContext {
+ const AVClass *class;
+
+ double g;
+ double B;
+ double h;
+ double a;
+ double m;
+ int n;
+ int K;
+ int S;
+
+ int N;
+ int hop_size;
+
+ AVFrame *in;
+
+ double *kernel;
+
+ int64_t pts;
+ uint64_t nb_samples;
+ int samples_left;
+
+ AVAudioFifo *fifo;
+} AudioNLMeansContext;
+
+#define OFFSET(x) offsetof(AudioNLMeansContext, x)
+#define AF AV_OPT_FLAG_AUDIO_PARAM|AV_OPT_FLAG_FILTERING_PARAM
+
+static const AVOption anlmeans_options[] = {
+ { "n", "set number of patches", OFFSET(n), AV_OPT_TYPE_INT, {.i64=1}, 1, 4, AF },
+ { "K", "set patch radius", OFFSET(K), AV_OPT_TYPE_INT, {.i64=32}, 0, 4096, AF },
+ { "S", "set research radius", OFFSET(S), AV_OPT_TYPE_INT, {.i64=64}, 1, 4096, AF },
+ { "h", "set strength", OFFSET(h), AV_OPT_TYPE_DOUBLE, {.dbl=1}, 1, FLT_MAX, AF },
+ { "m", "set max allowed diff", OFFSET(m), AV_OPT_TYPE_DOUBLE, {.dbl=0.1}, 0, 2, AF },
+ { "a", "set denoising amount", OFFSET(a), AV_OPT_TYPE_DOUBLE, {.dbl=1}, 1, FLT_MAX, AF },
+ { "B", "set smooth factor", OFFSET(B), AV_OPT_TYPE_DOUBLE, {.dbl=1}, 1, FLT_MAX, AF },
+ { "g", "set output gain", OFFSET(g), AV_OPT_TYPE_DOUBLE, {.dbl=1}, 0, 10, AF },
+ { NULL }
+};
+
+AVFILTER_DEFINE_CLASS(anlmeans);
+
+static int query_formats(AVFilterContext *ctx)
+{
+ AVFilterFormats *formats = NULL;
+ AVFilterChannelLayouts *layouts = NULL;
+ static const enum AVSampleFormat sample_fmts[] = {
+ AV_SAMPLE_FMT_FLTP,
+ AV_SAMPLE_FMT_NONE
+ };
+ int ret;
+
+ formats = ff_make_format_list(sample_fmts);
+ if (!formats)
+ return AVERROR(ENOMEM);
+ ret = ff_set_common_formats(ctx, formats);
+ if (ret < 0)
+ return ret;
+
+ layouts = ff_all_channel_counts();
+ if (!layouts)
+ return AVERROR(ENOMEM);
+
+ ret = ff_set_common_channel_layouts(ctx, layouts);
+ if (ret < 0)
+ return ret;
+
+ formats = ff_all_samplerates();
+ return ff_set_common_samplerates(ctx, formats);
+}
+
+static void compute_kernel(double *kernel, int K, double B2)
+{
+ int k;
+
+ for (k = 0; k <= 2 * K; k++) {
+ double kk = k - K;
+
+ kernel[k] = exp(-(kk * kk) / B2);
+ }
+}
+
+static int config_output(AVFilterLink *outlink)
+{
+ AVFilterContext *ctx = outlink->src;
+ AudioNLMeansContext *s = ctx->priv;
+
+ s->m = SQR(s->m * 65536.);
+ s->pts = AV_NOPTS_VALUE;
+ s->N = s->n * s->K * 2 + 1 + (s->K + s->S) * 2;
+ s->hop_size = s->n * s->K * 2 + 1;
+
+ av_frame_free(&s->in);
+ s->in = ff_get_audio_buffer(outlink, s->N);
+ if (!s->in)
+ return AVERROR(ENOMEM);
+
+ s->fifo = av_audio_fifo_alloc(outlink->format, outlink->channels, s->N);
+ if (!s->fifo)
+ return AVERROR(ENOMEM);
+
+ s->kernel = av_calloc(s->K * 2 + 1, sizeof(*s->kernel));
+ if (!s->kernel)
+ return AVERROR(ENOMEM);
+
+ compute_kernel(s->kernel, s->K, 1. / (s->B * s->B));
+
+ return 0;
+}
+
+typedef struct ThreadData {
+ AVFrame *out;
+} ThreadData;
+
+static double compute_distance(const double *kernel,
+ const float *f1, const float *f2, int K)
+{
+ double sum = 0.;
+ int k;
+
+ for (k = -K; k <= K; k++) {
+ sum += kernel[k] * SQR(65536. * f1[k] - 65536. * f2[k]);
+ }
+
+ return sum;
+}
+
+static int filter_channel(AVFilterContext *ctx, void *arg, int ch, int nb_jobs)
+{
+ AudioNLMeansContext *s = ctx->priv;
+ ThreadData *td = arg;
+ AVFrame *out = td->out;
+ const int S = s->S;
+ const int K = s->K;
+ const double m = s->m;
+ const double g = s->g;
+ const double *kernel = s->kernel + K;
+ const float *f = (const float *)(s->in->extended_data[ch]) + K;
+ const double sw = s->a / (s->h * s->h);
+ float *dst = (float *)out->extended_data[ch];
+ int i, j;
+
+ for (i = S; i < s->hop_size + S; i++) {
+ double P = 0., Q = 0.;
+
+ for (j = i - S; j <= i + S; j++) {
+ double w, d2;
+
+ d2 = compute_distance(kernel, f + i, f + j, K);
+
+ if (d2 > m)
+ continue;
+
+ w = exp(-d2 * sw);
+ P += w * f[j];
+ Q += w;
+ }
+
+ dst[i - S] = g * P / Q;
+ }
+
+ return 0;
+}
+
+static int filter_frame(AVFilterLink *inlink, AVFrame *in)
+{
+ AVFilterContext *ctx = inlink->dst;
+ AVFilterLink *outlink = ctx->outputs[0];
+ AudioNLMeansContext *s = ctx->priv;
+ AVFrame *out = NULL;
+ int ret = 0;
+
+ if (s->pts == AV_NOPTS_VALUE)
+ s->pts = in->pts;
+
+ ret = av_audio_fifo_write(s->fifo, (void **)in->extended_data,
+ in->nb_samples);
+ av_frame_free(&in);
+
+ while (av_audio_fifo_size(s->fifo) >= s->N) {
+ ThreadData td;
+
+ out = ff_get_audio_buffer(outlink, s->hop_size);
+ if (!out)
+ return AVERROR(ENOMEM);
+
+ ret = av_audio_fifo_peek(s->fifo, (void **)s->in->extended_data,
+ s->N);
+ if (ret < 0)
+ break;
+
+ td.out = out;
+ ctx->internal->execute(ctx, filter_channel, &td, NULL, inlink->channels);
+
+ av_audio_fifo_drain(s->fifo, s->hop_size);
+
+ if (s->samples_left > 0)
+ out->nb_samples = FFMIN(s->hop_size, s->samples_left);
+
+ out->pts = s->pts;
+ s->pts += s->hop_size;
+
+ ret = ff_filter_frame(outlink, out);
+ if (ret < 0)
+ break;
+ }
+
+ if (ret < 0)
+ av_frame_free(&out);
+ return ret;
+}
+
+static av_cold void uninit(AVFilterContext *ctx)
+{
+ AudioNLMeansContext *s = ctx->priv;
+
+ av_audio_fifo_free(s->fifo);
+ av_frame_free(&s->in);
+ av_freep(&s->kernel);
+}
+
+static const AVFilterPad inputs[] = {
+ {
+ .name = "default",
+ .type = AVMEDIA_TYPE_AUDIO,
+ .filter_frame = filter_frame,
+ },
+ { NULL }
+};
+
+static const AVFilterPad outputs[] = {
+ {
+ .name = "default",
+ .type = AVMEDIA_TYPE_AUDIO,
+ .config_props = config_output,
+ },
+ { NULL }
+};
+
+AVFilter ff_af_anlmeans = {
+ .name = "anlmeans",
+ .description = NULL_IF_CONFIG_SMALL("Reduce broadband noise from input audio using Non-Local Means."),
+ .query_formats = query_formats,
+ .priv_size = sizeof(AudioNLMeansContext),
+ .priv_class = &anlmeans_class,
+ .uninit = uninit,
+ .inputs = inputs,
+ .outputs = outputs,
+ .flags = AVFILTER_FLAG_SLICE_THREADS,
+};
@@ -49,6 +49,7 @@ extern AVFilter ff_af_amerge;
extern AVFilter ff_af_ametadata;
extern AVFilter ff_af_amix;
extern AVFilter ff_af_anequalizer;
+extern AVFilter ff_af_anlmeans;
extern AVFilter ff_af_anull;
extern AVFilter ff_af_apad;
extern AVFilter ff_af_aperms;
Signed-off-by: Paul B Mahol <onemda@gmail.com> --- libavfilter/Makefile | 1 + libavfilter/af_anlmeans.c | 282 ++++++++++++++++++++++++++++++++++++++++++++++ libavfilter/allfilters.c | 1 + 3 files changed, 284 insertions(+) create mode 100644 libavfilter/af_anlmeans.c