[FFmpeg-devel] [PATCH] Port mp=eq/eq2 to FFmpeg

arwa arif arwaarif1994 at gmail.com
Fri Jan 23 16:14:01 CET 2015


On Fri, Jan 23, 2015 at 8:05 PM, Stefano Sabatini <stefasab at gmail.com>
wrote:

> On date Thursday 2015-01-22 01:38:11 +0530, Arwa Arif encoded:
> > On Thu, Jan 22, 2015 at 12:09 AM, Paul B Mahol <onemda at gmail.com> wrote:
> [...]
> > From 703cc1887903c2868537e19e99b76927bec07884 Mon Sep 17 00:00:00 2001
> > From: Arwa Arif <arwaarif1994 at gmail.com>
> > Date: Mon, 19 Jan 2015 03:56:48 +0530
> > Subject: [PATCH] Port mp=eq/eq2 to FFmpeg
> >
>
> > Code adapted from James Darnley's previous commits
>
> Code adapted from James Darnley's port.
>
> There are no related commits in the FFmpeg master repo.
>
> > ---
> >  configure                |    1 +
> >  doc/filters.texi         |   44 ++++++++
> >  libavfilter/Makefile     |    1 +
> >  libavfilter/allfilters.c |    1 +
> >  libavfilter/vf_eq.c      |  282
> ++++++++++++++++++++++++++++++++++++++++++++++
> >  libavfilter/vf_eq.h      |   63 +++++++++++
> >  libavfilter/x86/Makefile |    1 +
> >  libavfilter/x86/vf_eq.c  |   94 ++++++++++++++++
> >  8 files changed, 487 insertions(+)
> >  create mode 100644 libavfilter/vf_eq.c
> >  create mode 100644 libavfilter/vf_eq.h
> >  create mode 100644 libavfilter/x86/vf_eq.c
> >
> > diff --git a/configure b/configure
> > index c73562b..138852e 100755
> > --- a/configure
> > +++ b/configure
> > @@ -2579,6 +2579,7 @@ delogo_filter_deps="gpl"
> >  deshake_filter_select="pixelutils"
> >  drawtext_filter_deps="libfreetype"
> >  ebur128_filter_deps="gpl"
> > +eq_filter_deps="gpl"
> >  flite_filter_deps="libflite"
> >  frei0r_filter_deps="frei0r dlopen"
> >  frei0r_src_filter_deps="frei0r dlopen"
> > diff --git a/doc/filters.texi b/doc/filters.texi
> > index d7b2273..70e0557 100644
> > --- a/doc/filters.texi
> > +++ b/doc/filters.texi
> > @@ -4320,6 +4320,50 @@ edgedetect=mode=colormix:high=0
> >  @end example
> >  @end itemize
> >
> > + at section eq
>
> > +Equalizer that uses lookup tables (very slow), allowing gamma correction
> > +in addition to simple brightness and contrast adjustment.
>
> This is not fitting (it's not always using LUTs). Also tells what the
> filter does, not what the filter is, as in the rest of the filters
> documentation. Something like this:
>
> Set brightness, contrast, saturation and gamma adjustment.
>
> > +
> > +The filter accepts the following options:
> > +
> > + at table @option
> > + at item brightness
> > +Set the brightness value. It accepts a float value in range @code{-1.0}
> to
> > + at code{1.0}. The default value is @code{0.0}.
> > +
> > + at item contrast
> > +Set the contrast value. It accepts a float value in range @code{-2.0} to
> > + at code{2.0}. The default value is @code{0.0}.
> > +
> > + at item gamma
> > +Set the gamma value. It accepts a float value in range @code{0.1} to
> @code{10.0}.
> > +The default value is @code{1.0}.
> > +
> > + at item gamma_y
> > +Set the gamma value for the luma plane. It accepts a float value in
> range
> > + at code{0.1} to @code{10.0}. The default value is @code{1.0}.
> > +
> > + at item gamma_u
> > +Set the gamma value for 1st chroma plane. It accepts a float value in
> range
> > + at code{0.1} to @code{10.0}. The default value is @code{1.0}.
> > +
> > + at item gamma_v
> > +Set the gamma value for 2nd chroma plane. It accepts a float value in
> range
> > + at code{0.1} to @code{10.0}. The default value is @code{1.0}.
> > +
> > + at item saturation
> > +Set the saturation value. It accepts a float value in range @code{0.0}
> to
> > + at code{3.0}. The default value is @code{1.0}.
> > +
> > + at item weight
>
> probably gamma_weight it's better, to make clear it is related to
> gamma correction.
>
> > +Can be used to reduce the effect of a high gamma value on bright image
> areas,
> > +e.g. keep them from getting overamplified and just plain white. It
> accepts a
> > +float value in range @code{0.0} to @code{1.0}.A value of @code{0.0}
> turns the
> > +gamma correction all the way down while @code{1.0} leaves it at its
> full strength.
> > +Default is @code{1.0}.
> > +
> > + at end table
> > +
> >  @section extractplanes
> >
> >  Extract color channel components from input video stream into
> > diff --git a/libavfilter/Makefile b/libavfilter/Makefile
> > index e43d76d..8e94033 100644
> > --- a/libavfilter/Makefile
> > +++ b/libavfilter/Makefile
> > @@ -116,6 +116,7 @@ OBJS-$(CONFIG_DRAWGRID_FILTER)               +=
> vf_drawbox.o
> >  OBJS-$(CONFIG_DRAWTEXT_FILTER)               += vf_drawtext.o
> >  OBJS-$(CONFIG_ELBG_FILTER)                   += vf_elbg.o
> >  OBJS-$(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_FIELD_FILTER)                  += vf_field.o
> > diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
> > index 381da4f..db34cb9 100644
> > --- a/libavfilter/allfilters.c
> > +++ b/libavfilter/allfilters.c
> > @@ -132,6 +132,7 @@ void avfilter_register_all(void)
> >      REGISTER_FILTER(DRAWTEXT,       drawtext,       vf);
> >      REGISTER_FILTER(EDGEDETECT,     edgedetect,     vf);
> >      REGISTER_FILTER(ELBG,           elbg,           vf);
> > +    REGISTER_FILTER(EQ,             eq,             vf);
> >      REGISTER_FILTER(EXTRACTPLANES,  extractplanes,  vf);
> >      REGISTER_FILTER(FADE,           fade,           vf);
> >      REGISTER_FILTER(FIELD,          field,          vf);
> > diff --git a/libavfilter/vf_eq.c b/libavfilter/vf_eq.c
> > new file mode 100644
> > index 0000000..fa4a2ea
> > --- /dev/null
> > +++ b/libavfilter/vf_eq.c
> > @@ -0,0 +1,282 @@
> > +/*
> > + * Original MPlayer filters by Richard Felker, Hampa Hug, Daniel Moreno,
> > + * and Michael Niedermeyer.
> > + *
> > + * Copyright (c) 2014 James Darnley <james.darnley at gmail.com>
> > + * 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 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
> > + * very simple video equalizer
> > + */
> > +
> > +/**
> > + * TODO:
> > + * - Add support to process_command
> > + */
> > +
> > +#include "libavfilter/internal.h"
> > +#include "libavutil/common.h"
> > +#include "libavutil/imgutils.h"
> > +#include "libavutil/opt.h"
> > +#include "libavutil/pixdesc.h"
> > +#include "vf_eq.h"
> > +
> > +static void create_lut(EQParameters *param)
> > +{
> > +    int i;
> > +    double g = param->gamma;
> > +
> > +    g = 1.0 / g;
> > +
> > +    for (i = 0; i < 256; i++) {
> > +        double v = i / 255.0;
> > +        v = param->contrast * (v - 0.5) + 0.5 + param->brightness;
> > +
> > +        if (v <= 0.0)
> > +            param->lut[i] = 0;
> > +        else {
> > +            v = v * (1.0 - param->weight) + pow(v, g) * param->weight;
> > +
> > +            if (v >= 1.0)
> > +                param->lut[i] = 255;
> > +            else
> > +                param->lut[i] = 256.0 * v;
> > +        }
> > +    }
> > +
> > +    param->lut_clean = 1;
> > +}
> > +
> > +static void apply_lut(EQParameters *param, uint8_t *dst, int dst_stride,
> > +                      uint8_t *src, int src_stride, int w, int h)
> > +{
> > +    int x, y;
> > +
> > +    if (!param->lut_clean)
> > +        create_lut(param);
> > +
> > +    for (y = 0; y < h; y++) {
> > +        for (x = 0; x < w; x++) {
> > +            dst[y*dst_stride+x] = param->lut[src[y*src_stride+x]];
> > +        }
> > +    }
> > +}
> > +
> > +static void process_c(EQParameters *param, uint8_t *dst, int dst_stride,
> > +                      uint8_t *src, int src_stride, int w, int h)
> > +{
> > +    int x, y, pel;
> > +
> > +    for (y = 0; y < h; y++) {
> > +        for (x = 0; x < w; x++) {
> > +            pel = ((src[y * src_stride + x] * param->c) >> 16) +
> param->b;
> > +
> > +            if (pel & 768)
> > +                pel = (-pel) >> 31;
> > +
> > +            dst[y * dst_stride + x] = pel;
> > +        }
> > +    }
> > +}
> > +
> > +static void check_values(EQParameters *param, EQContext *eq)
> > +{
> > +    if (param->contrast == 1.0 && param->brightness == 0.0 &&
> param->gamma == 1.0)
> > +        param->adjust = NULL;
> > +    else if (param->gamma == 1.0)
> > +        param->adjust = eq->process;
> > +    else
> > +        param->adjust = apply_lut;
> > +}
> > +
> > +static void set_contrast(EQContext *eq)
> > +{
> > +    eq->param[0].contrast = eq->contrast;
> > +    eq->param[0].lut_clean = 0;
> > +    check_values(&eq->param[0], eq);
> > +}
> > +
> > +static void set_brightness(EQContext *eq)
> > +{
> > +    eq->param[0].brightness = eq->brightness;
> > +    eq->param[0].lut_clean = 0;
> > +    check_values(&eq->param[0], eq);
> > +}
> > +
> > +static void set_gamma(EQContext *eq)
> > +{
> > +    int i;
> > +    eq->param[0].gamma = eq->gamma * eq->gamma_u;
> > +    eq->param[1].gamma = sqrt(eq->gamma_y / eq->gamma_u);
> > +    eq->param[2].gamma = sqrt(eq->gamma_v / eq->gamma_u);
> > +
> > +    for (i = 0; i < 3; i++) {
> > +        eq->param[i].weight = eq->weight;
> > +        eq->param[i].lut_clean = 0;
> > +        check_values(&eq->param[i], eq);
> > +    }
> > +}
> > +
> > +static void set_saturation(EQContext *eq)
> > +{
> > +    int i;
> > +    for (i = 1; i < 3; i++) {
> > +        eq->param[i].contrast = eq->saturation;
> > +        eq->param[i].lut_clean = 0;
> > +        check_values(&eq->param[i], eq);
> > +    }
> > +}
> > +
> > +static int initialize(AVFilterContext *ctx)
> > +{
> > +    EQContext *eq = ctx->priv;
> > +    int i;
> > +
>
> > +    set_gamma(eq);
> > +    set_contrast(eq);
> > +    set_brightness(eq);
> > +    set_saturation(eq);
> > +
> > +    for (i = 0; i < 3; i++) {
> > +        eq->param[i].c = (eq->param[i].contrast) * 65536.0;
> > +        eq->param[i].b = (eq->param[i].brightness + 1.0) * 255.5 -
> 128.0 - (eq->param[i].contrast) * 128.0;
> > +    }
> > +
> > +    eq->process = process_c;
>
> As already stated in another review, set_* are using eq->process, but
> you're only defining it here.
>
> > +
> > +    if (ARCH_X86)
> > +        ff_eq_init_x86(eq);
> > +
> > +    return 0;
> > +}
> > +
> > +static int query_formats(AVFilterContext *ctx)
> > +{
> > +    static const enum AVPixelFormat pixel_fmts_eq[] = {
> > +        AV_PIX_FMT_GRAY8,
> > +        AV_PIX_FMT_YUV410P,
> > +        AV_PIX_FMT_YUV411P,
> > +        AV_PIX_FMT_YUV420P,
> > +        AV_PIX_FMT_YUV422P,
> > +        AV_PIX_FMT_YUV444P,
> > +        AV_PIX_FMT_NONE
> > +    };
> > +
> > +    ff_set_common_formats(ctx, ff_make_format_list(pixel_fmts_eq));
> > +
> > +    return 0;
> > +}
> > +
> > +static int filter_frame(AVFilterLink *inlink, AVFrame *in)
> > +{
> > +    AVFilterContext *ctx = inlink->dst;
> > +    AVFilterLink *outlink = inlink->dst->outputs[0];
> > +    EQContext *eq = ctx->priv;
> > +    AVFrame *out;
> > +    const AVPixFmtDescriptor *desc;
> > +    int i;
> > +
> > +    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 (i = 0; i < desc->nb_components; i++) {
> > +        int w = inlink->w;
> > +        int h = inlink->h;
> > +
> > +        if (i == 1 || i == 2) {
> > +            w = FF_CEIL_RSHIFT(w, desc->log2_chroma_w);
> > +            h = FF_CEIL_RSHIFT(h, desc->log2_chroma_h);
> > +        }
> > +
> > +        if (eq->param[i].adjust)
> > +            eq->param[i].adjust(&eq->param[i], out->data[i],
> out->linesize[i],
> > +                                 in->data[i], in->linesize[i], w, h);
> > +        else
> > +            av_image_copy_plane(out->data[i], out->linesize[i],
> > +                                in->data[i], in->linesize[i], w, h);
> > +    }
> > +
>
> > +    if (in != out) {
> > +        if (in->data[3])
> > +            av_image_copy_plane(out->data[3], out->linesize[3],
> > +                                in ->data[3], in ->linesize[3],
> > +                                inlink->w, inlink->h);
> > +        av_frame_free(&in);
> > +    }
>
> provided that alpha is not supported, this should be probably unneded
> (but keep it if you think it's a good idea)
>
> > +
> > +    return ff_filter_frame(outlink, out);
> > +}
> > +static const AVFilterPad eq_inputs[] = {
> > +    {
> > +        .name = "default",
> > +        .type = AVMEDIA_TYPE_VIDEO,
> > +        .filter_frame = filter_frame,
> > +    },
> > +    { NULL }
> > +};
> > +
> > +static const AVFilterPad eq_outputs[] = {
> > +    {
> > +        .name = "default",
> > +        .type = AVMEDIA_TYPE_VIDEO,
> > +    },
> > +    { NULL }
> > +};
> > +
> > +#define OFFSET(x) offsetof(EQContext, x)
> > +#define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM
> > +
> > +static const AVOption eq_options[] = {
> > +    { "brightness", "set the brightness adjustment",
> > +        OFFSET(brightness), AV_OPT_TYPE_DOUBLE, {.dbl = 0.0}, -1.0,
> 1.0, FLAGS },
> > +    { "contrast",   "set the contrast adjustment, negative values give
> a negative image",
> > +        OFFSET(contrast),   AV_OPT_TYPE_DOUBLE, {.dbl = 1.0}, -2.0,
> 2.0, FLAGS },
> > +    { "gamma",      "set the initial gamma value",
> > +        OFFSET(gamma),      AV_OPT_TYPE_DOUBLE, {.dbl = 1.0},  0.1,
> 10.0, FLAGS },
> > +    { "gamma_y",    "gamma value for the luma plane",
> > +        OFFSET(gamma_y),    AV_OPT_TYPE_DOUBLE, {.dbl = 1.0},  0.1,
> 10.0, FLAGS },
> > +    { "gamma_u",    "gamma value for the 1st chroma plane",
> > +        OFFSET(gamma_u),    AV_OPT_TYPE_DOUBLE, {.dbl = 1.0},  0.1,
> 10.0, FLAGS },
> > +    { "gamma_v",    "gamma value for the 2st chroma plane",
> > +        OFFSET(gamma_v),    AV_OPT_TYPE_DOUBLE, {.dbl = 1.0},  0.1,
> 10.0, FLAGS },
> > +    { "saturation", "set the saturation adjustment",
> > +        OFFSET(saturation), AV_OPT_TYPE_DOUBLE, {.dbl = 1.0},  0.0,
> 3.0, FLAGS },
> > +    { "weight",     "set the gamma weight which reduces the effect of
> gamma on bright areas",
> > +        OFFSET(weight),     AV_OPT_TYPE_DOUBLE, {.dbl = 1.0},  0.0,
> 1.0, FLAGS },
> > +    { NULL }
> > +};
> > +
> > +AVFILTER_DEFINE_CLASS(eq);
> > +
> > +AVFilter ff_vf_eq = {
> > +    .name          = "eq",
> > +    .description   = NULL_IF_CONFIG_SMALL("Adjust brightness, contrast,
> gamma, and saturation."),
> > +    .priv_size     = sizeof(EQContext),
> > +    .priv_class    = &eq_class,
> > +    .inputs        = eq_inputs,
> > +    .outputs       = eq_outputs,
> > +    .query_formats = query_formats,
> > +    .init          = initialize,
> > +};
> > diff --git a/libavfilter/vf_eq.h b/libavfilter/vf_eq.h
> > new file mode 100644
> > index 0000000..dcf8b5e
> > --- /dev/null
> > +++ b/libavfilter/vf_eq.h
> > @@ -0,0 +1,63 @@
> > +/*
> > + * Original MPlayer filters by Richard Felker, Hampa Hug, Daniel Moreno,
> > + * and Michael Niedermeyer.
> > + *
> > + * Copyright (c) 2014 James Darnley <james.darnley at gmail.com>
> > + * 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 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.
> > + */
> > +
> > +#ifndef AVFILTER_EQ_H
> > +#define AVFILTER_EQ_H
> > +
> > +#include "avfilter.h"
> > +
> > +typedef struct EQParameters {
> > +
> > +    void (*adjust)(struct EQParameters *eq, uint8_t *dst, int
> dst_stride,
> > +                   uint8_t *src, int src_stride, int w, int h);
> > +
> > +    uint8_t lut[256];
> > +    uint16_t lut16[256*256];
> > +
> > +    double brightness, contrast, gamma, weight;
> > +    int b, c, lut_clean;
> > +
> > +} EQParameters;
> > +
> > +typedef struct {
> > +    const AVClass *class;
> > +
> > +    EQParameters param[3];
> > +
> > +    double contrast;
> > +    double brightness;
> > +    double saturation;
> > +
> > +    double gamma;
> > +    double weight;
> > +    double gamma_y, gamma_u, gamma_v;
> > +
> > +    void (*process)(struct EQParameters *par, uint8_t *dst, int
> dst_stride,
> > +                    uint8_t *src, int src_stride, int w, int h);
> > +
> > +} EQContext;
> > +
> > +void ff_eq_init_x86(EQContext *eq);
> > +
> > +#endif /* AVFILTER_EQ_H */
> > diff --git a/libavfilter/x86/Makefile b/libavfilter/x86/Makefile
> > index b93154e..8222e3f 100644
> > --- a/libavfilter/x86/Makefile
> > +++ b/libavfilter/x86/Makefile
> > @@ -1,3 +1,4 @@
> > +OBJS-$(CONFIG_EQ_FILTER)                     += x86/vf_eq.o
> >  OBJS-$(CONFIG_FSPP_FILTER)                   += x86/vf_fspp.o
> >  OBJS-$(CONFIG_GRADFUN_FILTER)                += x86/vf_gradfun_init.o
> >  OBJS-$(CONFIG_HQDN3D_FILTER)                 += x86/vf_hqdn3d_init.o
>
> This is not updated against latest git. Please *always* send patches
> updated against latest master, so that we can easily apply the patch.
>
> [...]
>
> Looks good otherwise, assuming it is bitexact with the mp=eq2.
>

The default is bit-exact with mp=eq2, I can't check it with other values,
because the range of values in mp is different from the range of values in
this code.


>
> Also, what about adding expressions support (like done in hue)? (this
> will go of course to a separate patch).
>

What is meant by expressions support?


I have updated the patch(with timer.h)

> --
> FFmpeg = Fundamentalist and Fanciful Muttering Peaceless Eccentric
> Generator
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-Port-mp-eq-eq2-to-FFmpeg.patch
Type: text/x-patch
Size: 29297 bytes
Desc: not available
URL: <https://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20150123/72307687/attachment.bin>


More information about the ffmpeg-devel mailing list