[FFmpeg-devel] [PATCH] lavfi: add field filter
Stefano Sabatini
stefasab at gmail.com
Tue Oct 30 22:36:21 CET 2012
On date Tuesday 2012-10-30 21:37:52 +0100, Clément Bœsch encoded:
> On Tue, Oct 30, 2012 at 01:31:14PM +0100, Stefano Sabatini wrote:
> > The filter is a port of libmpcodecs/vf_field.c, since there is no common
> > code I relicensed it as LGPL, while keeping the original author
> > copyright.
> >
> > TODO: bump lavfi minor, update Changelog
> > ---
> > doc/filters.texi | 25 ++++++++
> > libavfilter/Makefile | 1 +
> > libavfilter/allfilters.c | 1 +
> > libavfilter/vf_field.c | 145 +++++++++++++++++++++++++++++++++++++++++++++
> > tests/fate/avfilter.mak | 1 +
> > tests/lavfi-regression.sh | 1 +
> > tests/ref/lavfi/field | 83 ++++++++++++++++++++++++++
> > 7 files changed, 257 insertions(+), 0 deletions(-)
> > create mode 100644 libavfilter/vf_field.c
> > create mode 100644 tests/ref/lavfi/field
> >
> > diff --git a/doc/filters.texi b/doc/filters.texi
> > index 0e77914..215c12a 100644
> > --- a/doc/filters.texi
> > +++ b/doc/filters.texi
> > @@ -2115,6 +2115,31 @@ fade=in:5:20
> > fade=in:0:25:alpha=1
> > @end example
> >
> > + at section field
> > +
> > +Extract a single field from an interlaced image using stride
> > +arithmetic to avoid wasting CPU time. The output frames are marked as
> > +non-interlaced.
> > +
> > +This filter accepts the following named options:
> > + at table @option
> > + at item type
> > +Specify whether to extract the top (if the value is @code{0} or
> > + at code{top}) or the bottom field (if the value is @code{1} or
> > + at code{bottom}.
>
> missing closing ')'
Should be fixed in later version.
> > + at end table
> > +
> > +If the option key is not specified, the first value specifies the
> > +type. For example:
> > + at example
> > +field=type=bottom
> > + at end example
> > +
> > +is equivalent to:
> > + at example
> > +field=1
> > + at end example
> > +
>
> I'm still wondering about the kind of use-case for this filter though...
I don't know (never used it for practical purposes), but I suppose it
was implemented for a reason. Possible uses: testing (you want to test
the checksum of every single field), or you want to get a very
computationally cheap deinterlacer (yes it loses half of the
information, and squeeze the image).
>
> > @section fieldorder
> >
> > Transform the field order of the input video.
> > diff --git a/libavfilter/Makefile b/libavfilter/Makefile
> > index 3618f10..0da0645 100644
> > --- a/libavfilter/Makefile
> > +++ b/libavfilter/Makefile
> > @@ -98,6 +98,7 @@ OBJS-$(CONFIG_DRAWBOX_FILTER) += vf_drawbox.o
> > OBJS-$(CONFIG_DRAWTEXT_FILTER) += vf_drawtext.o
> > OBJS-$(CONFIG_EDGEDETECT_FILTER) += vf_edgedetect.o
> > OBJS-$(CONFIG_FADE_FILTER) += vf_fade.o
> > +OBJS-$(CONFIG_FIELD_FILTER) += vf_field.o
> > OBJS-$(CONFIG_FIELDORDER_FILTER) += vf_fieldorder.o
> > OBJS-$(CONFIG_FIFO_FILTER) += fifo.o
> > OBJS-$(CONFIG_FORMAT_FILTER) += vf_format.o
> > diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
> > index 348f369..fef40e9 100644
> > --- a/libavfilter/allfilters.c
> > +++ b/libavfilter/allfilters.c
> > @@ -90,6 +90,7 @@ void avfilter_register_all(void)
> > REGISTER_FILTER (DRAWTEXT, drawtext, vf);
> > REGISTER_FILTER (EDGEDETECT, edgedetect, vf);
> > REGISTER_FILTER (FADE, fade, vf);
> > + REGISTER_FILTER (FIELD, field, vf);
> > REGISTER_FILTER (FIELDORDER, fieldorder, vf);
> > REGISTER_FILTER (FIFO, fifo, vf);
> > REGISTER_FILTER (FORMAT, format, vf);
> > diff --git a/libavfilter/vf_field.c b/libavfilter/vf_field.c
> > new file mode 100644
> > index 0000000..db55b99
> > --- /dev/null
> > +++ b/libavfilter/vf_field.c
> > @@ -0,0 +1,145 @@
> > +/*
> > + * Copyright (c) 2003 Rich Felker
> > + * Copyright (c) 2012 Stefano Sabatini
> > + *
> > + * 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 field filter, based on libmpcodecs/vf_field.c by Rich
> > + * Felker
>
> missing \n (I think Paul pointed that out already)
Should be already fixed.
>
> > + */
> > +
> > +#include "libavutil/opt.h"
> > +#include "libavutil/pixdesc.h"
> > +#include "avfilter.h"
> > +#include "internal.h"
> > +
> > +enum FieldType { FIELD_TYPE_TOP = 0, FIELD_TYPE_BOTTOM };
> > +
> > +typedef struct {
> > + const AVClass *class;
> > + enum FieldType type;
> > + int nb_planes; ///< number of planes of the current format
> > +} FieldContext;
> > +
> > +#define OFFSET(x) offsetof(FieldContext, x)
> > +#define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM
> > +
> > +static const AVOption field_options[] = {
> > + {"type", "set field type (top or bottom)", OFFSET(type), AV_OPT_TYPE_INT, {.i64=FIELD_TYPE_TOP}, 0, 1, FLAGS, "field_type" },
> > + {"top", "select top field", 0, AV_OPT_TYPE_CONST, {.i64=FIELD_TYPE_TOP}, INT_MIN, INT_MAX, FLAGS, "field_type"},
> > + {"bottom", "select bottom field", 0, AV_OPT_TYPE_CONST, {.i64=FIELD_TYPE_BOTTOM}, INT_MIN, INT_MAX, FLAGS, "field_type"},
> > +
> > + {NULL}
> > +};
> > +
> > +AVFILTER_DEFINE_CLASS(field);
> > +
> > +static av_cold int init(AVFilterContext *ctx, const char *args)
> > +{
> > + FieldContext *field = ctx->priv;
> > + static const char *shorthand[] = { "type", NULL };
> > +
> > + field->class = &field_class;
> > + av_opt_set_defaults(field);
> > +
> > + return av_opt_set_from_string(field, args, shorthand, "=", ":");
> > +}
> > +
> > +static int config_props_output(AVFilterLink *outlink)
> > +{
> > + AVFilterContext *ctx = outlink->src;
> > + FieldContext *field = ctx->priv;
> > + AVFilterLink *inlink = ctx->inputs[0];
> > + const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(inlink->format);
> > + int i;
> > +
> > + for (i = 0; i < desc->nb_components; i++)
> > + field->nb_planes = FFMAX(field->nb_planes, desc->comp[i].plane);
> > + field->nb_planes++;
> > +
> > + outlink->w = inlink->w;
> > + outlink->h = inlink->h/2;
>
> If we have an odd height, won't this drop one of the line?
Fixed, although the case is not contemplated in mp=field.
[...]
> nit: weird align
Fixed.
> > +};
> > diff --git a/tests/fate/avfilter.mak b/tests/fate/avfilter.mak
> > index b931244..239fe64 100644
> > --- a/tests/fate/avfilter.mak
> > +++ b/tests/fate/avfilter.mak
> > @@ -11,6 +11,7 @@ FATE_LAVFI = fate-lavfi-alphaextract_rgb \
> > fate-lavfi-drawbox \
> > fate-lavfi-edgedetect \
> > fate-lavfi-fade \
> > + fate-lavfi-field \
> > fate-lavfi-idet \
> > fate-lavfi-life \
> > fate-lavfi-null \
> > diff --git a/tests/lavfi-regression.sh b/tests/lavfi-regression.sh
> > index 703d56c..76b2e5a 100755
> > --- a/tests/lavfi-regression.sh
> > +++ b/tests/lavfi-regression.sh
> > @@ -111,6 +111,7 @@ do_lavfi_pixfmts(){
> > do_lavfi_pixfmts "copy" ""
> > do_lavfi_pixfmts "crop" "100:100:100:100"
> > do_lavfi_pixfmts "hflip" ""
> > +do_lavfi_pixfmts "field" "field" "bottom"
>
> why the double "field"?
If there are three arguments, the first one is assumed to be the name
of the test (rather than "pixfmts_field"), seemed a good way to keep
the changes minimal.
>
> [...]
>
> Not much to say, that port looks pretty OK, feel free to dismiss any
> comment I have, and push with the mp drop when you think it's fine (unless
> someone doesn't agree).
Updated. I'll push tomorrow if I read no more comments.
--
FFmpeg = Fiendish Fierce Minimalistic Peaceful EniGma
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0005-lavfi-add-field-filter.patch
Type: text/x-diff
Size: 13840 bytes
Desc: not available
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20121030/effbf264/attachment.bin>
More information about the ffmpeg-devel
mailing list