[FFmpeg-devel] [PATCH] lavfi: add kerndeint filter

Clément Bœsch ubitux at gmail.com
Fri Dec 28 18:18:27 CET 2012


On Fri, Dec 28, 2012 at 04:41:45PM +0100, Stefano Sabatini wrote:
> On date Thursday 2012-10-18 09:50:53 +0200, Jérémy Tran encoded:
> > This is a port of the vf_kerndeint filter (libmpcodecs/vf_kerndeint) by
> > Donal A. Graft (original avisynth plugin author).
> > 
> > The filter works fine using YUV colorspace.
> > When using RGB, it is not binary equal with mp=kerndeint (the output video
> > looks fine so this may be a little mistake I did somewhere).
> > 
> > I also removed many non-32bits colorspaces since the alogrithm does not
> > seem to be made to work on shorter pixels, is this okay ?
> > The only non-32-bits colorspace left is YUYV422 which is what mp=kerndeint
> > uses, but I don't seem to get the same results (my first thought was
> > because it does not have 32bpp).
> > 
> > The FATE test will come after this.
> > ---
> >  configure                  |   1 +
> >  doc/filters.texi           |  45 ++++++
> >  libavfilter/Makefile       |   1 +
> >  libavfilter/allfilters.c   |   1 +
> >  libavfilter/vf_kerndeint.c | 377 +++++++++++++++++++++++++++++++++++++++++++++
> >  5 files changed, 425 insertions(+)
> >  create mode 100644 libavfilter/vf_kerndeint.c
> 
> Updated work in progress.
> 
> TODO: add a FATE test, fix the yuyv422 which is broken (same in
> mp=kerndeint), add support to more formats.
> -- 
> FFmpeg = Faithful & Free Mega Philosofic Erroneous Gorilla

> From ec015bfbea71a0d0de7f367b8c56040731b484c3 Mon Sep 17 00:00:00 2001
> From: Stefano Sabatini <stefasab at gmail.com>
> Date: Thu, 18 Oct 2012 09:50:53 +0200
> Subject: [PATCH] lavfi: add kerndeint filter
> MIME-Version: 1.0
> Content-Type: text/plain; charset=UTF-8
> Content-Transfer-Encoding: 8bit
> 
> This is a port of the vf_kerndeint filter (libmpcodecs/vf_kerndeint) by
> Donal A. Graft (original avisynth plugin author), and is based on the
> work by Jérémy Tran <tran.jeremy.av at gmail.com> done for SOCIS 2012.
> ---
>  configure                  |    1 +
>  doc/filters.texi           |   62 ++++++++
>  libavfilter/Makefile       |    1 +
>  libavfilter/allfilters.c   |    1 +
>  libavfilter/vf_kerndeint.c |  375 ++++++++++++++++++++++++++++++++++++++++++++
>  5 files changed, 440 insertions(+)
>  create mode 100644 libavfilter/vf_kerndeint.c
> 
> diff --git a/configure b/configure
> index c8b8c2b..6c489e9 100755
> --- a/configure
> +++ b/configure
> @@ -1986,6 +1986,7 @@ frei0r_src_filter_extralibs='$ldl'
>  geq_filter_deps="gpl"
>  hqdn3d_filter_deps="gpl"
>  hue_filter_deps="gpl"
> +kerndeint_filter_deps="gpl"

Required an entry in LICENSE too.

>  movie_filter_deps="avcodec avformat"
>  mp_filter_deps="gpl avcodec swscale inline_asm"
>  mptestsrc_filter_deps="gpl"
> diff --git a/doc/filters.texi b/doc/filters.texi
> index 765a9b4..9cd6ac1 100644
> --- a/doc/filters.texi
> +++ b/doc/filters.texi
> @@ -2862,6 +2862,68 @@ If a parameter is omitted, it is kept at its current value.
>  Interlaceing detect filter. This filter tries to detect if the input is
>  interlaced or progressive. Top or bottom field first.
>  
> + at section kerndeint
> +
> +Deinterlace input video by applying Donald Graft's adaptive kernel
> +deinterling. Work on interlaced parts of a video to produce
> +progressive frames.
> +
> +This filter accepts parameters as a list of @var{key}=@var{value}
> +pairs, separated by ":". The description of the accepted parameters
> +follows.
> +
> +If the key of the first options is omitted, the arguments are
> +interpreted according to the following syntax:
> + at var{thresh}:@var{map}:@var{order}:@var{sharp}:@var{twoway}
> +
> + at table @option
> + at item thresh
> +Set the threshold which affects the filter's tolerance when
> +determining if a pixel line must be processed. Must be an integer in
> +the range [0,255] and defaults to 10. A value of 0 will result in
> +applying the process on every pixels.
> +
> + at item map
> +Must be 0 (ignore pixels exceeding the threshold) or 1 (paint pixels
> +exceeding the threshold white) and defaults to 0.
> +
> + at item order
> +Must be 0 (leave fields alone) or 1 (swap fields) and defaults to 0.
> +
> + at item sharp
> +Must be 0 (disable additional sharpening) or 1 (enable additional
> +sharpening) and defaults to 0.
> +
> + at item twoway
> +Must be 0 (disable twoway sharpening) or 1 (enable twoway sharpening)
> +and defaults to 0.
> +
> + at item deinterlace
> +Disable deinterlacing if set to 0. Default is 0.

nit: I'd say "Enable deinterlacing if set to 1".

> + at end table
> +
> + at subsection Examples
> +
> + at itemize
> + at item
> +Apply default values:
> + at example
> +kerndeint=thresh=10:map=0:order=0:sharp=0:twoway=0
> + at end example
> +
> + at item
> +Enable additional sharpening:
> + at example
> +kerndeint=sharp=1
> + at end example
> +
> + at item
> +Paint processed pixels in white:
> + at example
> +kerndeint=map=1
> + at end example
> + at end itemize
> +
>  @section lut, lutrgb, lutyuv
>  
>  Compute a look-up table for binding each pixel component input value
> diff --git a/libavfilter/Makefile b/libavfilter/Makefile
> index 34baa6d..b09b59a 100644
> --- a/libavfilter/Makefile
> +++ b/libavfilter/Makefile
> @@ -112,6 +112,7 @@ OBJS-$(CONFIG_HFLIP_FILTER)                  += vf_hflip.o
>  OBJS-$(CONFIG_HQDN3D_FILTER)                 += vf_hqdn3d.o
>  OBJS-$(CONFIG_HUE_FILTER)                    += vf_hue.o
>  OBJS-$(CONFIG_IDET_FILTER)                   += vf_idet.o
> +OBJS-$(CONFIG_KERNDEINT_FILTER)              += vf_kerndeint.o
>  OBJS-$(CONFIG_LUT_FILTER)                    += vf_lut.o
>  OBJS-$(CONFIG_LUTRGB_FILTER)                 += vf_lut.o
>  OBJS-$(CONFIG_LUTYUV_FILTER)                 += vf_lut.o
> diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
> index 88b0979..fb192d6 100644
> --- a/libavfilter/allfilters.c
> +++ b/libavfilter/allfilters.c
> @@ -103,6 +103,7 @@ void avfilter_register_all(void)
>      REGISTER_FILTER (HQDN3D,      hqdn3d,      vf);
>      REGISTER_FILTER (HUE,         hue,         vf);
>      REGISTER_FILTER (IDET,        idet,        vf);
> +    REGISTER_FILTER (KERNDEINT,   kerndeint,   vf);
>      REGISTER_FILTER (LUT,         lut,         vf);
>      REGISTER_FILTER (LUTRGB,      lutrgb,      vf);
>      REGISTER_FILTER (LUTYUV,      lutyuv,      vf);
> diff --git a/libavfilter/vf_kerndeint.c b/libavfilter/vf_kerndeint.c
> new file mode 100644
> index 0000000..052d2a6
> --- /dev/null
> +++ b/libavfilter/vf_kerndeint.c
> @@ -0,0 +1,375 @@
> +/*
> + * Copyright (c) 2012 Jeremy Tran
> + * Copyright (c) 2004 Tobias Diedrich
> + * Copyright (c) 2003 Donald A. Graft
> + *
> + * This file is part of FFmpeg.
> + *
> + * FFmpeg is free software; you can redistribute it and/or modify
> + * it under the terms of the GNU General Public License as published by
> + * the Free Software Foundation; either version 2 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 General Public License for more details.
> + *
> + * You should have received a copy of the GNU 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
> + * Kernel Deinterlacer
> + * Ported from MPlayer libmpcodecs/vf_kerndeint.c.
> + */
> +
> +#include "libavutil/imgutils.h"
> +#include "libavutil/opt.h"
> +#include "libavutil/pixdesc.h"
> +
> +#include "avfilter.h"
> +#include "formats.h"
> +#include "internal.h"
> +
> +#define PLANAR_Y 0
> +#define PLANAR_U 1
> +#define PLANAR_V 2
> +
> +#define RGB  0
> +#define YUYV 1
> +#define YUV  2
> +
> +typedef struct {
> +    const AVClass *class;
> +    int           frame;
> +    int           map;
> +    int           order;
> +    int           thresh;
> +    int           sharp;
> +    int           twoway;
> +    int           do_deinterlace;
> +    int           hsub;
> +    int           vsub;
> +    int           format;        ///< Pixel format: RGB, YUYV or YUV
> +    uint8_t       *temp_data[4]; ///< Temporary video plane data buffer
> +    int           bpp;           ///< bytes per pixel
> +} KerndeintContext;
> +
> +#define OFFSET(x) offsetof(KerndeintContext, x)
> +#define FLAGS AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_FILTERING_PARAM
> +static const AVOption kerndeint_options[] = {
> +    { "thresh", "set the threshold", OFFSET(thresh), AV_OPT_TYPE_INT, { 10 }, 0, 255, FLAGS },
> +    { "map",    "set the map", OFFSET(map), AV_OPT_TYPE_INT, { 0 }, 0, 1, FLAGS },
> +    { "order",  "set the order", OFFSET(order), AV_OPT_TYPE_INT, { 0 }, 0, 1, FLAGS },
> +    { "sharp",  "set the sharpening mode", OFFSET(sharp), AV_OPT_TYPE_INT, { 0 }, 0, 1, FLAGS },
> +    { "twoway", "activate twoway", OFFSET(twoway), AV_OPT_TYPE_INT, { 0 }, 0, 1, FLAGS },
> +    { "deinterlace", "disable interlacing", OFFSET(do_deinterlace), AV_OPT_TYPE_INT, { 1 }, 0, 1, FLAGS },
> +    { NULL }
> +};
> +
> +AVFILTER_DEFINE_CLASS(kerndeint);
> +
> +static av_cold int init(AVFilterContext *ctx, const char *args)
> +{
> +    KerndeintContext *kerndeint = ctx->priv;
> +    int i;
> +    const char const * shorthand[] = { "thresh", "map", "order", "sharp", "twoway", NULL };
> +
> +    kerndeint->class = &kerndeint_class;

> +    kerndeint->do_deinterlace = 1;

Looks unnecessary

> +    for (i = 0; i < 4; ++i)
> +        kerndeint->temp_data[i] = NULL;

ditto

> +    av_opt_set_defaults(kerndeint);
> +
> +    return av_opt_set_from_string(kerndeint, args, shorthand, "=", ":");
> +}
> +
> +static av_cold void uninit(AVFilterContext *ctx)
> +{
> +    KerndeintContext *kerndeint = ctx->priv;
> +    int i;
> +
> +    for (i = 0; i < 4; ++i)
> +        av_free(kerndeint->temp_data[i]);
> +    av_opt_free(kerndeint);
> +}
> +
> +static int query_formats(AVFilterContext *ctx)
> +{
> +    static const enum PixelFormat pix_fmts[] = {
> +        PIX_FMT_YUV420P,
> +        PIX_FMT_YUVJ420P,
> +        PIX_FMT_YUYV422,
> +        PIX_FMT_ARGB,
> +        PIX_FMT_NONE
> +    };
> +
> +    ff_set_common_formats(ctx, ff_make_format_list(pix_fmts));
> +
> +    return 0;
> +}
> +
> +static int config_props(AVFilterLink *inlink)
> +{
> +    KerndeintContext *kerndeint = inlink->dst->priv;
> +    const AVPixFmtDescriptor *desc = &av_pix_fmt_descriptors[inlink->format];
> +    int i;
> +
> +    kerndeint->hsub = desc->log2_chroma_w;
> +    kerndeint->vsub = desc->log2_chroma_h;
> +    kerndeint->bpp = av_get_bits_per_pixel(desc) / 8;
> +
> +    for (i = 0; i < 4; ++i)
> +        kerndeint->temp_data[i] = av_mallocz(inlink->h * inlink->w * kerndeint->bpp);
> +

Missing alloc check

> +    switch (inlink->format) {
> +        case PIX_FMT_YUV420P:
> +            kerndeint->format = YUV;
> +            break;
> +        case PIX_FMT_YUYV422:
> +            kerndeint->format = YUYV;
> +            break;
> +        case PIX_FMT_ARGB:
> +            kerndeint->format = RGB;
> +            break;
> +    }
> +
> +    return 0;
> +}
> +

> +#define AVWN32(buf, value) *(uint32_t *)(buf) = value

Don't we have safer macro for this?

> +
> +static int filter_frame(AVFilterLink *inlink, AVFilterBufferRef *inpic)
> +{
> +    KerndeintContext *kerndeint = inlink->dst->priv;
> +    AVFilterLink *outlink = inlink->dst->outputs[0];
> +    AVFilterBufferRef *outpic;
> +    int cw = inlink->w >> kerndeint->hsub;
> +    int ch = inlink->h >> kerndeint->vsub;

> +    const unsigned char *prvp;   ///< Previous field's pixel line number n
> +    const unsigned char *prvpp;  ///< Previous field's pixel line number (n - 1)
> +    const unsigned char *prvpn;  ///< Previous field's pixel line number (n + 1)
> +    const unsigned char *prvppp; ///< Previous field's pixel line number (n - 2)
> +    const unsigned char *prvpnn; ///< Previous field's pixel line number (n + 2)
> +    const unsigned char *prvp4p; ///< Previous field's pixel line number (n - 4)
> +    const unsigned char *prvp4n; ///< Previous field's pixel line number (n + 4)
> +
> +    const unsigned char *srcp;   ///< Current field's pixel line number n
> +    const unsigned char *srcpp;  ///< Current field's pixel line number (n - 1)
> +    const unsigned char *srcpn;  ///< Current field's pixel line number (n + 1)
> +    const unsigned char *srcppp; ///< Current field's pixel line number (n - 2)
> +    const unsigned char *srcpnn; ///< Current field's pixel line number (n + 2)
> +    const unsigned char *srcp3p; ///< Current field's pixel line number (n - 3)
> +    const unsigned char *srcp3n; ///< Current field's pixel line number (n + 3)
> +    const unsigned char *srcp4p; ///< Current field's pixel line number (n - 4)
> +    const unsigned char *srcp4n; ///< Current field's pixel line number (n + 4)
> +

nit: uint8_t

> +    unsigned char *dstp, *dstp_saved;
> +    const unsigned char *srcp_saved;
> +
> +    int src_linesize;
> +    int psrc_linesize;
> +    int dst_linesize;
> +    int x, y, plane, g;
> +    int n = kerndeint->frame++;
> +    int val, hi, lo, w, h;
> +    double valf;
> +    const int threshold = kerndeint->thresh;
> +    const int order     = kerndeint->order;
> +    const int map       = kerndeint->map;
> +    const int sharp     = kerndeint->sharp;
> +    const int twoway    = kerndeint->twoway;
> +
> +    if (!kerndeint->do_deinterlace)
> +        return ff_filter_frame(outlink, inpic);
> +
> +    outpic = ff_get_video_buffer(outlink, AV_PERM_WRITE|AV_PERM_ALIGN, outlink->w, outlink->h);
> +    if (!outpic) {
> +        avfilter_unref_bufferp(&inpic);
> +        return AVERROR(ENOMEM);
> +    }
> +    avfilter_copy_buffer_ref_props(outpic, inpic);
> +
> +    for (plane = 0; inpic->data[plane] && plane < 4; plane++) {

You should check for plane before dereferencing

> +        h = plane == PLANAR_Y ? inlink->h : ch;
> +        w = plane == PLANAR_Y ? inlink->w * kerndeint->bpp : cw;

not sure: no need to apply bpp to cw?

> +
> +        srcp = srcp_saved = inpic->data[plane];
> +        src_linesize      = inpic->linesize[plane];
> +        psrc_linesize     = outpic->linesize[plane];
> +        dstp = dstp_saved = outpic->data[plane];
> +        dst_linesize      = outpic->linesize[plane];
> +        srcp              = srcp_saved + (1 - order) * src_linesize;
> +        dstp              = dstp_saved + (1 - order) * dst_linesize;
> +
> +        for (y = 0; y < h; y += 2) {
> +            memcpy(dstp, srcp, w);
> +            srcp += 2 * src_linesize;
> +            dstp += 2 * dst_linesize;
> +        }
> +
> +        // Copy through the lines that will be missed below.
> +        memcpy(dstp_saved + order           * dst_linesize, srcp_saved + (1 - order)     * src_linesize, w);
> +        memcpy(dstp_saved + (2 + order)     * dst_linesize, srcp_saved + (3 - order)     * src_linesize, w);
> +        memcpy(dstp_saved + (h - 2 + order) * dst_linesize, srcp_saved + (h - 1 - order) * src_linesize, w);
> +        memcpy(dstp_saved + (h - 4 + order) * dst_linesize, srcp_saved + (h - 3 - order) * src_linesize, w);
> +
> +        /* For the other field choose adaptively between using the previous field
> +           or the interpolant from the current field. */
> +        prvp   = kerndeint->temp_data[plane] + 5 * psrc_linesize - (1 - order) * psrc_linesize;
> +        prvpp  = prvp - psrc_linesize;
> +        prvppp = prvp - 2 * psrc_linesize;
> +        prvp4p = prvp - 4 * psrc_linesize;
> +        prvpn  = prvp + psrc_linesize;
> +        prvpnn = prvp + 2 * psrc_linesize;
> +        prvp4n = prvp + 4 * psrc_linesize;
> +
> +        srcp   = srcp_saved + 5 * src_linesize - (1 - order) * src_linesize;
> +        srcpp  = srcp - src_linesize;
> +        srcppp = srcp - 2 * src_linesize;
> +        srcp3p = srcp - 3 * src_linesize;
> +        srcp4p = srcp - 4 * src_linesize;
> +
> +        srcpn  = srcp + src_linesize;
> +        srcpnn = srcp + 2 * src_linesize;
> +        srcp3n = srcp + 3 * src_linesize;
> +        srcp4n = srcp + 4 * src_linesize;
> +
> +        dstp   = dstp_saved  + 5 * dst_linesize - (1 - order) * dst_linesize;
> +
> +        for (y = 5 - (1 - order); y <= h - 5 - (1 - order); y += 2) {
> +            for (x = 0; x < w; x++) {
> +                if ((threshold == 0) || (n == 0) ||
> +                    (abs((int)prvp[x]  - (int)srcp[x])  > threshold) ||
> +                    (abs((int)prvpp[x] - (int)srcpp[x]) > threshold) ||
> +                    (abs((int)prvpn[x] - (int)srcpn[x]) > threshold)) {

nit: looks like there is a bunch of pointless ()

> +                    if (map == 1) {
> +                        g = x & ~3;
> +
> +                        if (kerndeint->format == RGB) {
> +                            AVWN32(dstp + g, 0xffffffff);
> +                            x = g + 3;
> +                        } else if (kerndeint->format == YUYV) {
> +                            AVWN32(dstp + g, 0xeb80eb80);
> +                            x = g + 3;
> +                        } else {
> +                            dstp[x] = plane == PLANAR_Y ? 235 : 128;
> +                        }
> +                    } else {
> +                        if (kerndeint->format == RGB) {
> +                            hi = 255;
> +                            lo = 0;
> +                        } else if (kerndeint->format == YUYV) {
> +                            hi = (x & 1) ? 240 : 235;
> +                            lo = 16;
> +                        } else {
> +                            hi = (plane == PLANAR_Y) ? 235 : 240;
> +                            lo = 16;
> +                        }
> +
> +                        if (sharp == 1) {
> +                            if (twoway == 1) {
> +                                valf =   0.526 * ((int)srcpp[x] + (int)srcpn[x])
> +                                    + 0.170 * ((int)srcp[x] + (int)prvp[x])
> +                                    - 0.116 * ((int)srcppp[x]
> +                                               + (int)srcpnn[x]
> +                                               + (int)prvppp[x]
> +                                               + (int)prvpnn[x])
> +                                    - 0.026 * ((int)srcp3p[x] + (int)srcp3n[x])
> +                                    + 0.031 * ((int)srcp4p[x]
> +                                               + (int)srcp4n[x]
> +                                               + (int)prvp4p[x]
> +                                               + (int)prvp4n[x]);
> +                            } else {
> +                                valf =   0.526 * ((int)srcpp[x] + (int)srcpn[x])
> +                                    + 0.170 * ((int)prvp[x])
> +                                    - 0.116 * ((int)prvppp[x] + (int)prvpnn[x])
> +                                    - 0.026 * ((int)srcp3p[x] + (int)srcp3n[x])
> +                                    + 0.031 * ((int)prvp4p[x] + (int)prvp4p[x]);
> +                            }
> +
> +                            valf = av_clip(valf, lo, hi);
> +                            dstp[x] = (int) valf;
> +                        } else {
> +                            if (twoway == 1) {
> +                                val = (8 * ((int)srcpp[x] + (int)srcpn[x])
> +                                       + 2 * ((int)srcp[x] + (int)prvp[x])
> +                                       - (int)(srcppp[x]) - (int)(srcpnn[x])
> +                                       - (int)(prvppp[x]) - (int)(prvpnn[x]))
> +                                    >> 4;
> +                            } else {
> +                                val = (8 * ((int)srcpp[x]
> +                                            + (int)srcpn[x])
> +                                       + 2 * ((int)prvp[x])
> +                                       - (int)(prvppp[x]) - (int)(prvpnn[x]))
> +                                    >> 4;
> +                            }
> +
> +                            val = av_clip(val, lo, hi);
> +                            dstp[x] = val;
> +                        }
> +                    }
> +                } else {
> +                    dstp[x] = srcp[x];
> +                }
> +            }
> +            prvp   += 2 * psrc_linesize;
> +            prvpp  += 2 * psrc_linesize;
> +            prvppp += 2 * psrc_linesize;
> +            prvpn  += 2 * psrc_linesize;
> +            prvpnn += 2 * psrc_linesize;
> +            prvp4p += 2 * psrc_linesize;
> +            prvp4n += 2 * psrc_linesize;
> +            srcp   += 2 * src_linesize;
> +            srcpp  += 2 * src_linesize;
> +            srcppp += 2 * src_linesize;
> +            srcp3p += 2 * src_linesize;
> +            srcp4p += 2 * src_linesize;
> +            srcpn  += 2 * src_linesize;
> +            srcpnn += 2 * src_linesize;
> +            srcp3n += 2 * src_linesize;
> +            srcp4n += 2 * src_linesize;
> +            dstp   += 2 * dst_linesize;
> +        }
> +
> +        srcp = inpic->data[plane];
> +        dstp = kerndeint->temp_data[plane];
> +        av_image_copy_plane(dstp, psrc_linesize, srcp, src_linesize, w, h);
> +    }
> +
> +    avfilter_unref_buffer(inpic);
> +    return ff_filter_frame(outlink, outpic);
> +}
> +
> +AVFilter avfilter_vf_kerndeint = {
> +    .name        = "kerndeint",
> +    .description = NULL_IF_CONFIG_SMALL("Kernel deinterlacer"),
> +
> +    .priv_size = sizeof(KerndeintContext),
> +
> +    .init          = init,
> +    .uninit        = uninit,
> +    .query_formats = query_formats,
> +
> +    .inputs = (const AVFilterPad[]) {
> +        {
> +            .name         = "default",
> +            .type         = AVMEDIA_TYPE_VIDEO,
> +            .filter_frame = filter_frame,
> +            .config_props = config_props,
> +            .min_perms    = AV_PERM_READ,
> +        },
> +        { .name = NULL }
> +    },
> +    .outputs = (const AVFilterPad[]) {
> +        {
> +            .name         = "default",
> +            .type         = AVMEDIA_TYPE_VIDEO,
> +        },
> +        { .name = NULL }
> +    },

Please split this like the other filters

> +    .priv_class = &kerndeint_class,
> +};

-- 
Clément B.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 490 bytes
Desc: not available
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20121228/9cded55e/attachment.asc>


More information about the ffmpeg-devel mailing list