[FFmpeg-cvslog] avfilter: Add FFT domain filter.
Arwa Arif
git at videolan.org
Sat Mar 7 13:16:07 CET 2015
ffmpeg | branch: master | Arwa Arif <arwaarif1994 at gmail.com> | Tue Feb 24 12:35:30 2015 +0530| [48b5be65aedbc4ae12b3a0c32ed5d86a12dd039a] | committer: Michael Niedermayer
avfilter: Add FFT domain filter.
Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=48b5be65aedbc4ae12b3a0c32ed5d86a12dd039a
---
Changelog | 1 +
doc/filters.texi | 38 +++++
libavfilter/Makefile | 1 +
libavfilter/allfilters.c | 1 +
libavfilter/version.h | 4 +-
libavfilter/vf_fftfilt.c | 343 ++++++++++++++++++++++++++++++++++++++++++++++
6 files changed, 386 insertions(+), 2 deletions(-)
diff --git a/Changelog b/Changelog
index 640b14d..7db8abd 100644
--- a/Changelog
+++ b/Changelog
@@ -2,6 +2,7 @@ Entries are sorted chronologically from oldest to youngest within each release,
releases are sorted from youngest to oldest.
version <next>:
+- FFT video filter
version 2.6:
diff --git a/doc/filters.texi b/doc/filters.texi
index 0c72145..b15087c 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -4591,6 +4591,44 @@ fade=t=in:st=5.5:d=0.5
@end itemize
+ at section fftfilt
+Apply arbitrary expressions to samples in frequency domain
+
+ at table @option
+ at item dc_Y
+Adjust the dc value (gain) of the luma plane of the image. The filter
+accepts an integer value in range @code{0} to @code{1000}. The default
+value is set to @code{0}.
+
+ at item dc_U
+Adjust the dc value (gain) of the 1st chroma plane of the image. The
+filter accepts an integer value in range @code{0} to @code{1000}. The
+default value is set to @code{0}.
+
+ at item dc_V
+Adjust the dc value (gain) of the 2nd chroma plane of the image. The
+filter accepts an integer value in range @code{0} to @code{1000}. The
+default value is set to @code{0}.
+
+ at item weight_Y
+Set the frequency domain weight expression for the luma plane.
+
+ at item weight_U
+Set the frequency domain weight expression for the 1st chroma plane.
+
+ at item weight_V
+Set the frequency domain weight expression for the 2nd chroma plane.
+
+The filter accepts the following variables:
+ at item X
+ at item Y
+The coordinates of the current sample.
+
+ at item W
+ at item H
+The width and height of the image.
+ at end table
+
@section field
Extract a single field from an interlaced image using stride
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index 289c63b..b184f07 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -120,6 +120,7 @@ OBJS-$(CONFIG_EDGEDETECT_FILTER) += vf_edgedetect.o
OBJS-$(CONFIG_EQ_FILTER) += vf_eq.o
OBJS-$(CONFIG_EXTRACTPLANES_FILTER) += vf_extractplanes.o
OBJS-$(CONFIG_FADE_FILTER) += vf_fade.o
+OBJS-$(CONFIG_FFTFILT_FILTER) += vf_fftfilt.o
OBJS-$(CONFIG_FIELD_FILTER) += vf_field.o
OBJS-$(CONFIG_FIELDMATCH_FILTER) += vf_fieldmatch.o
OBJS-$(CONFIG_FIELDORDER_FILTER) += vf_fieldorder.o
diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
index 55de154..043ac56 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -136,6 +136,7 @@ void avfilter_register_all(void)
REGISTER_FILTER(EQ, eq, vf);
REGISTER_FILTER(EXTRACTPLANES, extractplanes, vf);
REGISTER_FILTER(FADE, fade, vf);
+ REGISTER_FILTER(FFTFILT, fftfilt, vf);
REGISTER_FILTER(FIELD, field, vf);
REGISTER_FILTER(FIELDMATCH, fieldmatch, vf);
REGISTER_FILTER(FIELDORDER, fieldorder, vf);
diff --git a/libavfilter/version.h b/libavfilter/version.h
index 986d7a7..a964392 100644
--- a/libavfilter/version.h
+++ b/libavfilter/version.h
@@ -30,8 +30,8 @@
#include "libavutil/version.h"
#define LIBAVFILTER_VERSION_MAJOR 5
-#define LIBAVFILTER_VERSION_MINOR 11
-#define LIBAVFILTER_VERSION_MICRO 102
+#define LIBAVFILTER_VERSION_MINOR 12
+#define LIBAVFILTER_VERSION_MICRO 100
#define LIBAVFILTER_VERSION_INT AV_VERSION_INT(LIBAVFILTER_VERSION_MAJOR, \
LIBAVFILTER_VERSION_MINOR, \
diff --git a/libavfilter/vf_fftfilt.c b/libavfilter/vf_fftfilt.c
new file mode 100644
index 0000000..86acc62
--- /dev/null
+++ b/libavfilter/vf_fftfilt.c
@@ -0,0 +1,343 @@
+/*
+ * Copyright (c) 2015 Arwa Arif <arwaarif1994 at gmail.com>
+ *
+ * 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
+ * FFT domain filtering.
+ */
+
+#include "libavfilter/internal.h"
+#include "libavutil/common.h"
+#include "libavutil/imgutils.h"
+#include "libavutil/opt.h"
+#include "libavutil/pixdesc.h"
+#include "libavcodec/avfft.h"
+#include "libavutil/eval.h"
+
+typedef struct {
+ const AVClass *class;
+
+ RDFTContext *rdft;
+ int rdft_hbits[3];
+ int rdft_vbits[3];
+ size_t rdft_hlen[3];
+ size_t rdft_vlen[3];
+ FFTSample *rdft_hdata[3];
+ FFTSample *rdft_vdata[3];
+
+ int dc[3];
+ char *weight_str[3];
+ AVExpr *weight_expr[3];
+ double *weight[3];
+
+} FFTFILTContext;
+
+static const char *const var_names[] = { "X", "Y", "W", "H", NULL };
+enum { VAR_X, VAR_Y, VAR_W, VAR_H, VAR_VARS_NB };
+
+enum { Y = 0, U, V };
+
+#define OFFSET(x) offsetof(FFTFILTContext, x)
+#define FLAGS AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_FILTERING_PARAM
+
+static const AVOption fftfilt_options[] = {
+ { "dc_Y", "adjust gain in Y plane", OFFSET(dc[Y]), AV_OPT_TYPE_INT, {.i64 = 0}, 0, 1000, FLAGS },
+ { "dc_U", "adjust gain in U plane", OFFSET(dc[U]), AV_OPT_TYPE_INT, {.i64 = 0}, 0, 1000, FLAGS },
+ { "dc_V", "adjust gain in V plane", OFFSET(dc[V]), AV_OPT_TYPE_INT, {.i64 = 0}, 0, 1000, FLAGS },
+ { "weight_Y", "set luminance expression in Y plane", OFFSET(weight_str[Y]), AV_OPT_TYPE_STRING, {.str = "1"}, CHAR_MIN, CHAR_MAX, FLAGS },
+ { "weight_U", "set chrominance expression in U plane", OFFSET(weight_str[U]), AV_OPT_TYPE_STRING, {.str = NULL}, CHAR_MIN, CHAR_MAX, FLAGS },
+ { "weight_V", "set chrominance expression in V plane", OFFSET(weight_str[V]), AV_OPT_TYPE_STRING, {.str = NULL}, CHAR_MIN, CHAR_MAX, FLAGS },
+ {NULL},
+};
+
+AVFILTER_DEFINE_CLASS(fftfilt);
+
+static inline double lum(void *priv, double x, double y, int plane)
+{
+ FFTFILTContext *fftfilt = priv;
+ return fftfilt->rdft_vdata[plane][(int)x * fftfilt->rdft_vlen[plane] + (int)y];
+}
+
+static double weight_Y(void *priv, double x, double y) { return lum(priv, x, y, Y); }
+static double weight_U(void *priv, double x, double y) { return lum(priv, x, y, U); }
+static double weight_V(void *priv, double x, double y) { return lum(priv, x, y, V); }
+
+static void copy_rev (FFTSample *dest, int w, int w2)
+{
+ int i;
+
+ for (i = w; i < w + (w2-w)/2; i++)
+ dest[i] = dest[2*w - i - 1];
+
+ for (; i < w2; i++)
+ dest[i] = dest[w2 - i];
+}
+
+/*Horizontal pass - RDFT*/
+static void rdft_horizontal(FFTFILTContext *fftfilt, AVFrame *in, int w, int h, int plane)
+{
+ int i, j;
+ fftfilt->rdft = av_rdft_init(fftfilt->rdft_hbits[plane], DFT_R2C);
+
+ for (i = 0; i < h; i++) {
+ for (j = 0; j < w; j++)
+ fftfilt->rdft_hdata[plane][i * fftfilt->rdft_hlen[plane] + j] = *(in->data[plane] + in->linesize[plane] * i + j);
+
+ copy_rev(fftfilt->rdft_hdata[plane] + i * fftfilt->rdft_hlen[plane], w, fftfilt->rdft_hlen[plane]);
+ }
+
+ for (i = 0; i < h; i++)
+ av_rdft_calc(fftfilt->rdft, fftfilt->rdft_hdata[plane] + i * fftfilt->rdft_hlen[plane]);
+
+ av_rdft_end(fftfilt->rdft);
+}
+
+/*Vertical pass - RDFT*/
+static void rdft_vertical(FFTFILTContext *fftfilt, int h, int plane)
+{
+ int i, j;
+ fftfilt->rdft = av_rdft_init(fftfilt->rdft_vbits[plane], DFT_R2C);
+
+ for (i = 0; i < fftfilt->rdft_hlen[plane]; i++) {
+ for (j = 0; j < h; j++)
+ fftfilt->rdft_vdata[plane][i * fftfilt->rdft_vlen[plane] + j] =
+ fftfilt->rdft_hdata[plane][j * fftfilt->rdft_hlen[plane] + i];
+ copy_rev(fftfilt->rdft_vdata[plane] + i * fftfilt->rdft_vlen[plane], h, fftfilt->rdft_vlen[plane]);
+ }
+
+ for (i = 0; i < fftfilt->rdft_hlen[plane]; i++)
+ av_rdft_calc(fftfilt->rdft, fftfilt->rdft_vdata[plane] + i * fftfilt->rdft_vlen[plane]);
+
+ av_rdft_end(fftfilt->rdft);
+}
+/*Vertical pass - IRDFT*/
+static void irdft_vertical(FFTFILTContext *fftfilt, int h, int plane)
+{
+ int i, j;
+ fftfilt->rdft = av_rdft_init(fftfilt->rdft_vbits[plane], IDFT_C2R);
+ for (i = 0; i < fftfilt->rdft_hlen[plane]; i++)
+ av_rdft_calc(fftfilt->rdft, fftfilt->rdft_vdata[plane] + i * fftfilt->rdft_vlen[plane]);
+
+ for (i = 0; i < fftfilt->rdft_hlen[plane]; i++)
+ for (j = 0; j < h; j++)
+ fftfilt->rdft_hdata[plane][j * fftfilt->rdft_hlen[plane] + i] =
+ fftfilt->rdft_vdata[plane][i * fftfilt->rdft_vlen[plane] + j];
+
+ av_rdft_end(fftfilt->rdft);
+}
+
+/*Horizontal pass - IRDFT*/
+static void irdft_horizontal(FFTFILTContext *fftfilt, AVFrame *out, int w, int h, int plane)
+{
+ int i, j;
+ fftfilt->rdft = av_rdft_init(fftfilt->rdft_hbits[plane], IDFT_C2R);
+ for (i = 0; i < h; i++)
+ av_rdft_calc(fftfilt->rdft, fftfilt->rdft_hdata[plane] + i * fftfilt->rdft_hlen[plane]);
+
+ for (i = 0; i < h; i++)
+ for (j = 0; j < w; j++)
+ *(out->data[plane] + out->linesize[plane] * i + j) = av_clip(fftfilt->rdft_hdata[plane][i
+ *fftfilt->rdft_hlen[plane] + j] * 4 /
+ (fftfilt->rdft_hlen[plane] *
+ fftfilt->rdft_vlen[plane]), 0, 255);
+
+ av_rdft_end(fftfilt->rdft);
+}
+
+static av_cold int initialize(AVFilterContext *ctx)
+{
+ FFTFILTContext *fftfilt = ctx->priv;
+ int ret = 0, plane;
+
+ if (!fftfilt->dc[U] && !fftfilt->dc[V]) {
+ fftfilt->dc[U] = fftfilt->dc[Y];
+ fftfilt->dc[V] = fftfilt->dc[Y];
+ } else {
+ if (!fftfilt->dc[U]) fftfilt->dc[U] = fftfilt->dc[V];
+ if (!fftfilt->dc[V]) fftfilt->dc[V] = fftfilt->dc[U];
+ }
+
+ if (!fftfilt->weight_str[U] && !fftfilt->weight_str[V]) {
+ fftfilt->weight_str[U] = av_strdup(fftfilt->weight_str[Y]);
+ fftfilt->weight_str[V] = av_strdup(fftfilt->weight_str[Y]);
+ } else {
+ if (!fftfilt->weight_str[U]) fftfilt->weight_str[U] = av_strdup(fftfilt->weight_str[V]);
+ if (!fftfilt->weight_str[V]) fftfilt->weight_str[V] = av_strdup(fftfilt->weight_str[U]);
+ }
+
+ for (plane = 0; plane < 3; plane++) {
+ static double (*p[])(void *, double, double) = { weight_Y, weight_U, weight_V };
+ const char *const func2_names[] = {"weight_Y", "weight_U", "weight_V", NULL };
+ double (*func2[])(void *, double, double) = { weight_Y, weight_U, weight_V, p[plane], NULL };
+
+ ret = av_expr_parse(&fftfilt->weight_expr[plane], fftfilt->weight_str[plane], var_names,
+ NULL, NULL, func2_names, func2, 0, ctx);
+ if (ret < 0)
+ break;
+ }
+ return ret;
+}
+
+static int config_props(AVFilterLink *inlink)
+{
+ FFTFILTContext *fftfilt = inlink->dst->priv;
+ const AVPixFmtDescriptor *desc;
+ int rdft_hbits, rdft_vbits, i, j, plane;
+ double values[VAR_VARS_NB];
+
+ desc = av_pix_fmt_desc_get(inlink->format);
+ for (i = 0; i < desc->nb_components; i++) {
+ int w = inlink->w;
+ int h = inlink->h;
+
+ /* RDFT - Array initialization for Horizontal pass*/
+ for (rdft_hbits = 1; 1 << rdft_hbits < w; rdft_hbits++);
+ fftfilt->rdft_hbits[i] = rdft_hbits;
+ fftfilt->rdft_hlen[i] = 1 << rdft_hbits;
+ if (!(fftfilt->rdft_hdata[i] = av_malloc_array(h, fftfilt->rdft_hlen[i] * sizeof(FFTSample))))
+ return AVERROR(ENOMEM);
+
+ /* RDFT - Array initialization for Vertical pass*/
+ for (rdft_vbits = 1; 1 << rdft_vbits < h; rdft_vbits++);
+ fftfilt->rdft_vbits[i] = rdft_vbits;
+ fftfilt->rdft_vlen[i] = 1 << rdft_vbits;
+ if (!(fftfilt->rdft_vdata[i] = av_malloc_array(fftfilt->rdft_hlen[i], fftfilt->rdft_vlen[i] * sizeof(FFTSample))))
+ return AVERROR(ENOMEM);
+ }
+
+ /*Luminance value - Array initialization*/
+ values[VAR_W] = inlink->w;
+ values[VAR_H] = inlink->h;
+ for (plane = 0; plane < 3; plane++)
+ {
+ if(!(fftfilt->weight[plane] = av_malloc_array(fftfilt->rdft_hlen[plane], fftfilt->rdft_vlen[plane] * sizeof(double))))
+ return AVERROR(ENOMEM);
+ for (i = 0; i < fftfilt->rdft_hlen[plane]; i++)
+ {
+ values[VAR_X] = i;
+ for (j = 0; j < fftfilt->rdft_vlen[plane]; j++)
+ {
+ values[VAR_Y] = j;
+ fftfilt->weight[plane][i * fftfilt->rdft_vlen[plane] + j] =
+ av_expr_eval(fftfilt->weight_expr[plane], values, fftfilt);
+ }
+ }
+ }
+ return 0;
+}
+
+static int filter_frame(AVFilterLink *inlink, AVFrame *in)
+{
+ AVFilterContext *ctx = inlink->dst;
+ AVFilterLink *outlink = inlink->dst->outputs[0];
+ const AVPixFmtDescriptor *desc;
+ FFTFILTContext *fftfilt = ctx->priv;
+ AVFrame *out;
+ int i, j, plane;
+
+ out = ff_get_video_buffer(outlink, inlink->w, inlink->h);
+ if (!out)
+ return AVERROR(ENOMEM);
+
+ av_frame_copy_props(out, in);
+
+ desc = av_pix_fmt_desc_get(inlink->format);
+ for (plane = 0; plane < desc->nb_components; plane++) {
+ int w = inlink->w;
+ int h = inlink->h;
+
+ if (plane == 1 || plane == 2) {
+ w = FF_CEIL_RSHIFT(w, desc->log2_chroma_w);
+ h = FF_CEIL_RSHIFT(h, desc->log2_chroma_h);
+ }
+
+ rdft_horizontal(fftfilt, in, w, h, plane);
+ rdft_vertical(fftfilt, h, plane);
+
+ /*Change user defined parameters*/
+ for (i = 0; i < fftfilt->rdft_hlen[plane]; i++)
+ for (j = 0; j < fftfilt->rdft_vlen[plane]; j++)
+ fftfilt->rdft_vdata[plane][i * fftfilt->rdft_vlen[plane] + j] *=
+ fftfilt->weight[plane][i * fftfilt->rdft_vlen[plane] + j];
+
+ fftfilt->rdft_vdata[plane][0] += fftfilt->rdft_hlen[plane] * fftfilt->rdft_vlen[plane] * fftfilt->dc[plane];
+
+ irdft_vertical(fftfilt, h, plane);
+ irdft_horizontal(fftfilt, out, w, h, plane);
+ }
+
+ av_frame_free(&in);
+ return ff_filter_frame(outlink, out);
+}
+
+static av_cold void uninit(AVFilterContext *ctx)
+{
+ FFTFILTContext *fftfilt = ctx->priv;
+ int i;
+ for (i = 0; i < 3; i++) {
+ av_free(fftfilt->rdft_hdata[i]);
+ av_free(fftfilt->rdft_vdata[i]);
+ av_expr_free(fftfilt->weight_expr[i]);
+ av_free(fftfilt->weight[i]);
+ }
+}
+
+static int query_formats(AVFilterContext *ctx)
+{
+ static const enum AVPixelFormat pixel_fmts_fftfilt[] = {
+ AV_PIX_FMT_GRAY8,
+ AV_PIX_FMT_YUV444P,
+ AV_PIX_FMT_NONE
+ };
+
+ ff_set_common_formats(ctx, ff_make_format_list(pixel_fmts_fftfilt));
+
+ return 0;
+}
+
+static const AVFilterPad fftfilt_inputs[] = {
+ {
+ .name = "default",
+ .type = AVMEDIA_TYPE_VIDEO,
+ .config_props = config_props,
+ .filter_frame = filter_frame,
+ },
+ { NULL }
+};
+
+static const AVFilterPad fftfilt_outputs[] = {
+ {
+ .name = "default",
+ .type = AVMEDIA_TYPE_VIDEO,
+ },
+ { NULL }
+};
+
+AVFilter ff_vf_fftfilt = {
+ .name = "fftfilt",
+ .description = NULL_IF_CONFIG_SMALL("Apply arbitrary expressions to samples in frequency domain"),
+ .priv_size = sizeof(FFTFILTContext),
+ .priv_class = &fftfilt_class,
+ .inputs = fftfilt_inputs,
+ .outputs = fftfilt_outputs,
+ .query_formats = query_formats,
+ .init = initialize,
+ .uninit = uninit,
+};
More information about the ffmpeg-cvslog
mailing list