[FFmpeg-devel] avfilter: add superequalizer filter

Submitted by Paul B Mahol on June 17, 2017, 4 p.m.

Details

Message ID 20170617160001.19604-1-onemda@gmail.com
State New
Headers show

Commit Message

Paul B Mahol June 17, 2017, 4 p.m.
Signed-off-by: Paul B Mahol <onemda@gmail.com>
---
 libavfilter/Makefile            |   1 +
 libavfilter/af_superequalizer.c | 381 ++++++++++++++++++++++++++++++++++++++++
 libavfilter/allfilters.c        |   1 +
 3 files changed, 383 insertions(+)
 create mode 100644 libavfilter/af_superequalizer.c

Comments

James Almer June 17, 2017, 5:19 p.m.
On 6/17/2017 1:00 PM, Paul B Mahol wrote:
> Signed-off-by: Paul B Mahol <onemda@gmail.com>
> ---
>  libavfilter/Makefile            |   1 +
>  libavfilter/af_superequalizer.c | 381 ++++++++++++++++++++++++++++++++++++++++
>  libavfilter/allfilters.c        |   1 +
>  3 files changed, 383 insertions(+)
>  create mode 100644 libavfilter/af_superequalizer.c
> 
> diff --git a/libavfilter/Makefile b/libavfilter/Makefile
> index 04ec9b8..52c44d2 100644
> --- a/libavfilter/Makefile
> +++ b/libavfilter/Makefile
> @@ -109,6 +109,7 @@ OBJS-$(CONFIG_SILENCEREMOVE_FILTER)          += af_silenceremove.o
>  OBJS-$(CONFIG_SOFALIZER_FILTER)              += af_sofalizer.o
>  OBJS-$(CONFIG_STEREOTOOLS_FILTER)            += af_stereotools.o
>  OBJS-$(CONFIG_STEREOWIDEN_FILTER)            += af_stereowiden.o
> +OBJS-$(CONFIG_SUPEREQUALIZER_FILTER)         += af_superequalizer.o
>  OBJS-$(CONFIG_SURROUND_FILTER)               += af_surround.o
>  OBJS-$(CONFIG_TREBLE_FILTER)                 += af_biquads.o
>  OBJS-$(CONFIG_TREMOLO_FILTER)                += af_tremolo.o
> diff --git a/libavfilter/af_superequalizer.c b/libavfilter/af_superequalizer.c
> new file mode 100644
> index 0000000..c569394
> --- /dev/null
> +++ b/libavfilter/af_superequalizer.c
> @@ -0,0 +1,381 @@
> +/*
> + * 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 "libavutil/opt.h"
> +
> +#include "libavcodec/avfft.h"
> +
> +#include "audio.h"
> +#include "avfilter.h"
> +#include "internal.h"
> +
> +#define NBANDS 17
> +#define M 15
> +
> +typedef struct EqParameter {
> +    float lower, upper, gain;
> +} EqParameter;
> +
> +typedef struct SuperEqualizerContext {
> +    const AVClass *class;
> +
> +    EqParameter params[NBANDS+1];
> +
> +    float gains[NBANDS+1];
> +
> +    float fact[M+1];
> +    float aa;
> +    float iza;
> +    float *lires, *rires, *irest;
> +    float *fsamples;
> +    int chg_ires, cur_ires;
> +    int winlen, winlenbit, tabsize;
> +
> +    AVFrame *in, *out;
> +    RDFTContext *rdft, *irdft;
> +} SuperEqualizerContext;
> +
> +static const float bands[] = {
> +    65.406392, 92.498606, 130.81278, 184.99721, 261.62557, 369.99442, 523.25113, 739.9884, 1046.5023,
> +    1479.9768, 2093.0045, 2959.9536, 4186.0091, 5919.9072, 8372.0181, 11839.814, 16744.036
> +};
> +
> +static float izero(SuperEqualizerContext *s, float x)
> +{
> +    float ret = 1;
> +    int m;
> +
> +    for (m = 1; m <= M; m++) {
> +        float t;
> +
> +        t = pow(x / 2, m) / s->fact[m];

powf

> +        ret += t*t;
> +    }
> +
> +    return ret;
> +}
> +
> +static float hn_lpf(int n, float f, float fs)
> +{
> +    float t = 1 / fs;
> +    float omega = 2 * M_PI * f;
> +
> +    if (n * omega * t == 0)
> +        return 2 * f * t;
> +    return 2 * f * t * sin(n * omega * t)/(n * omega * t);

sinf

> +}
> +
> +static float hn_imp(int n)
> +{
> +    return n == 0 ? 1.0 : 0.0;

Use float constants.

Same for every other constant below.

> +}
> +
> +static float hn(int n, EqParameter *param, float fs)
> +{
> +    float ret, lhn;
> +    int i;
> +
> +    lhn = hn_lpf(n, param[0].upper, fs);
> +    ret = param[0].gain*lhn;
> +
> +    for (i = 1; i < NBANDS+1 && param[i].upper < fs / 2; i++) {
> +        float lhn2 = hn_lpf(n, param[i].upper, fs);
> +        ret += param[i].gain * (lhn2-lhn);
> +        lhn = lhn2;
> +    }
> +
> +    ret += param[i].gain * (hn_imp(n) - lhn);
> +
> +    return ret;
> +}
> +
> +static float alpha(float a)
> +{
> +    if (a <= 21)
> +        return 0;
> +    if (a <= 50)
> +        return .5842 * pow(a - 21,0.4) + 0.07886 * (a - 21);
> +    return .1102 * (a - 8.7);
> +}
> +
> +static float win(SuperEqualizerContext *s, float n, int N)
> +{
> +    return izero(s, alpha(s->aa) * sqrt(1 - 4 * n * n / ((N - 1) * (N - 1)))) / s->iza;

sqrtf

All these are called inside loops so it's best to prevent the compiler
from pointlessly using doubles.

> +}
> +
> +static void
> +process_param(float *bc, EqParameter *param, float fs, int ch)
> +{
> +    int i;
> +
> +    for (i = 0; i <= NBANDS; i++) {
> +        param[i].lower = i == 0 ? 0 : bands[i - 1];
> +        param[i].upper = i == NBANDS - 1 ? fs : bands[i];
> +        param[i].gain  = bc[i];
> +    }
> +}
> +
> +static int equ_init(SuperEqualizerContext *s, int wb)
> +{
> +    int i,j;
> +
> +    s->rdft  = av_rdft_init(wb, DFT_R2C);
> +    s->irdft = av_rdft_init(wb, IDFT_C2R);
> +    if (!s->rdft || !s->irdft)
> +        return AVERROR(ENOMEM);
> +
> +    s->aa = 96;
> +    s->winlen = (1 << (wb-1))-1;
> +    s->winlenbit = wb;
> +    s->tabsize  = 1 << wb;
> +
> +    s->lires    = av_calloc(s->tabsize, sizeof(float));
> +    s->rires    = av_calloc(s->tabsize, sizeof(float));
> +    s->irest    = av_calloc(s->tabsize, sizeof(float));
> +    s->fsamples = av_calloc(s->tabsize, sizeof(float));

Missing check for allocation failure. Also, none of these seem to be
freed during uninit.

> +
> +    s->cur_ires = 1;
> +    s->chg_ires = 1;
> +
> +    for (i = 0; i <= M; i++) {
> +        s->fact[i] = 1;
> +        for (j = 1; j <= i; j++)
> +            s->fact[i] *= j;
> +    }
> +
> +    s->iza = izero(s, alpha(s->aa));
> +
> +    return 0;
> +}
> +
> +static void make_fir(SuperEqualizerContext *s, float *lbc, float *rbc, EqParameter *param, float fs)
> +{
> +    const int winlen = s->winlen;
> +    const int tabsize = s->tabsize;
> +    int i, cires = s->cur_ires;
> +    float *nires;
> +
> +    if (fs <= 0)
> +        return;
> +
> +    process_param(lbc, param, fs, 0);
> +    for (i = 0; i < winlen; i++)
> +        s->irest[i] = hn(i - winlen / 2, param, fs) * win(s, i - winlen / 2, winlen);
> +    for (; i < tabsize; i++)
> +        s->irest[i] = 0;
> +
> +    av_rdft_calc(s->rdft, s->irest);
> +    nires = s->lires;
> +    for (i = 0; i < tabsize; i++)
> +        nires[i] = s->irest[i];
> +
> +    process_param(rbc, param, fs, 1);
> +    for (i = 0; i < winlen; i++)
> +        s->irest[i] = hn(i - winlen / 2, param, fs) * win(s, i - winlen / 2, winlen);
> +    for (; i < tabsize; i++)
> +        s->irest[i] = 0;
> +
> +    av_rdft_calc(s->rdft, s->irest);
> +    nires = s->rires;
> +    for (i = 0; i < tabsize; i++)
> +        nires[i] = s->irest[i];
> +    s->chg_ires = cires == 1 ? 2 : 1;
> +}
> +
> +static int filter_frame(AVFilterLink *inlink, AVFrame *in)
> +{
> +    AVFilterContext *ctx = inlink->dst;
> +    SuperEqualizerContext *s = ctx->priv;
> +    AVFilterLink *outlink = ctx->outputs[0];
> +    const float *ires = s->lires;
> +    float *fsamples = s->fsamples;
> +    int ch, i;
> +
> +    AVFrame *out = ff_get_audio_buffer(outlink, s->winlen);
> +    float *src, *dst, *ptr;
> +
> +    if (!out) {
> +        av_frame_free(&in);
> +        return AVERROR(ENOMEM);
> +    }
> +
> +    for (ch = 0; ch < in->channels; ch++) {
> +        ptr = (float *)out->extended_data[ch];
> +        dst = (float *)s->out->extended_data[ch];
> +        src = (float *)in->extended_data[ch];
> +
> +        for (i = 0; i < s->winlen; i++)
> +            fsamples[i] = src[i];
> +        for (; i < s->tabsize; i++)
> +            fsamples[i] = 0;
> +
> +        av_rdft_calc(s->rdft, fsamples);
> +
> +        fsamples[0] = ires[0] * fsamples[0];
> +        fsamples[1] = ires[1] * fsamples[1];
> +        for (i = 1; i < s->tabsize / 2; i++) {
> +            float re, im;
> +
> +            re = ires[i*2  ] * fsamples[i*2] - ires[i*2+1] * fsamples[i*2+1];
> +            im = ires[i*2+1] * fsamples[i*2] + ires[i*2  ] * fsamples[i*2+1];
> +
> +            fsamples[i*2  ] = re;
> +            fsamples[i*2+1] = im;
> +        }
> +
> +        av_rdft_calc(s->irdft, fsamples);
> +
> +        for (i = 0; i < s->winlen; i++)
> +            dst[i] += fsamples[i] / s->tabsize * 2;
> +        for (i = s->winlen; i < s->tabsize; i++)
> +            dst[i]  = fsamples[i] / s->tabsize * 2;
> +        for (i = 0; i < s->winlen; i++)
> +            ptr[i] = dst[i];
> +        for (i = 0; i < s->winlen; i++)
> +            dst[i] = dst[i+s->winlen];
> +    }
> +
> +    out->pts = in->pts;
> +    av_frame_free(&in);
> +
> +    return ff_filter_frame(outlink, out);
> +}
> +
> +static av_cold int init(AVFilterContext *ctx)
> +{
> +    SuperEqualizerContext *s = ctx->priv;
> +
> +    return equ_init(s, 14);
> +}
> +
> +static int query_formats(AVFilterContext *ctx)
> +{
> +    AVFilterFormats *formats;
> +    AVFilterChannelLayouts *layouts;
> +    static const enum AVSampleFormat sample_fmts[] = {
> +        AV_SAMPLE_FMT_FLTP,
> +        AV_SAMPLE_FMT_NONE
> +    };
> +    int 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_make_format_list(sample_fmts);
> +    if ((ret = ff_set_common_formats(ctx, formats)) < 0)
> +        return ret;
> +
> +    formats = ff_all_samplerates();
> +    return ff_set_common_samplerates(ctx, formats);
> +}
> +
> +static int config_input(AVFilterLink *inlink)
> +{
> +    AVFilterContext *ctx = inlink->dst;
> +    SuperEqualizerContext *s = ctx->priv;
> +
> +    inlink->partial_buf_size =
> +    inlink->min_samples =
> +    inlink->max_samples = s->winlen;
> +
> +    s->out = ff_get_audio_buffer(inlink, s->tabsize);
> +    if (!s->out)
> +        return AVERROR(ENOMEM);
> +
> +    return 0;
> +}
> +
> +static int config_output(AVFilterLink *outlink)
> +{
> +    AVFilterContext *ctx = outlink->src;
> +    SuperEqualizerContext *s = ctx->priv;
> +
> +    make_fir(s, s->gains, s->gains, s->params, outlink->sample_rate);
> +
> +    return 0;
> +}
> +
> +static av_cold void uninit(AVFilterContext *ctx)
> +{
> +    SuperEqualizerContext *s = ctx->priv;
> +
> +    av_rdft_end(s->rdft);
> +    av_rdft_end(s->irdft);
> +}
> +
> +static const AVFilterPad superequalizer_inputs[] = {
> +    {
> +        .name         = "default",
> +        .type         = AVMEDIA_TYPE_AUDIO,
> +        .filter_frame = filter_frame,
> +        .config_props = config_input,
> +    },
> +    { NULL }
> +};
> +
> +static const AVFilterPad superequalizer_outputs[] = {
> +    {
> +        .name         = "default",
> +        .type         = AVMEDIA_TYPE_AUDIO,
> +        .config_props = config_output,
> +    },
> +    { NULL }
> +};
> +
> +#define AF AV_OPT_FLAG_AUDIO_PARAM|AV_OPT_FLAG_FILTERING_PARAM
> +#define OFFSET(x) offsetof(SuperEqualizerContext, x)
> +
> +static const AVOption superequalizer_options[] = {
> +    {  "1b", "set  1st band gain", OFFSET(gains [0]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
> +    {  "2b", "set  2nd band gain", OFFSET(gains [1]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
> +    {  "3b", "set  3rd band gain", OFFSET(gains [2]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
> +    {  "4b", "set  4th band gain", OFFSET(gains [3]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
> +    {  "5b", "set  5th band gain", OFFSET(gains [4]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
> +    {  "6b", "set  6th band gain", OFFSET(gains [5]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
> +    {  "7b", "set  7th band gain", OFFSET(gains [6]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
> +    {  "8b", "set  8th band gain", OFFSET(gains [7]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
> +    {  "9b", "set  9th band gain", OFFSET(gains [8]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
> +    { "10b", "set 10nd band gain", OFFSET(gains [9]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
> +    { "11b", "set 11nd band gain", OFFSET(gains[10]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
> +    { "12b", "set 12nd band gain", OFFSET(gains[11]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
> +    { "13b", "set 13nd band gain", OFFSET(gains[12]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
> +    { "14b", "set 14th band gain", OFFSET(gains[13]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
> +    { "15b", "set 15th band gain", OFFSET(gains[14]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
> +    { "16b", "set 16th band gain", OFFSET(gains[15]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
> +    { "17b", "set 17th band gain", OFFSET(gains[16]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
> +    { "18b", "set 18th band gain", OFFSET(gains[17]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
> +    { NULL }
> +};
> +
> +AVFILTER_DEFINE_CLASS(superequalizer);
> +
> +AVFilter ff_af_superequalizer = {
> +    .name          = "superequalizer",
> +    .description   = NULL_IF_CONFIG_SMALL("Apply 18-th band equalization filter."),
> +    .priv_size     = sizeof(SuperEqualizerContext),
> +    .priv_class    = &superequalizer_class,
> +    .init          = init,
> +    .query_formats = query_formats,
> +    .uninit        = uninit,
> +    .inputs        = superequalizer_inputs,
> +    .outputs       = superequalizer_outputs,
> +};
> diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
> index 94f7cf3..bd81091 100644
> --- a/libavfilter/allfilters.c
> +++ b/libavfilter/allfilters.c
> @@ -122,6 +122,7 @@ static void register_all(void)
>      REGISTER_FILTER(SOFALIZER,      sofalizer,      af);
>      REGISTER_FILTER(STEREOTOOLS,    stereotools,    af);
>      REGISTER_FILTER(STEREOWIDEN,    stereowiden,    af);
> +    REGISTER_FILTER(SUPEREQUALIZER, superequalizer, af);
>      REGISTER_FILTER(SURROUND,       surround,       af);
>      REGISTER_FILTER(TREBLE,         treble,         af);
>      REGISTER_FILTER(TREMOLO,        tremolo,        af);
>
Moritz Barsnick June 19, 2017, 12:38 p.m.
On Sat, Jun 17, 2017 at 18:00:01 +0200, Paul B Mahol wrote:
> + * This file is part of FFmpeg.
> + *

No copyright at all? (Not that I know better or care.)

> +static float alpha(float a)
> +{
> +    if (a <= 21)
> +        return 0;
> +    if (a <= 50)
> +        return .5842 * pow(a - 21,0.4) + 0.07886 * (a - 21);
> +    return .1102 * (a - 8.7);

I didn't try to compile this with the corresponding compiler warnings,
but I believe this is a typical float -> double promotion, as "8.7"
without an 'f' (8.7f) is a double. And unless you need the precision of
a double calculation, that's a big hit on softfloat hardware. (I'm not
sure it matters for ffmpeg, but ffmpeg *does* compile on armv5tejl for
example). So when mixing float variables with doubles, do add 'f'.
(Probably in other places as well.)

> +    .description   = NULL_IF_CONFIG_SMALL("Apply 18-th band equalization filter."),

18-th band? Not 18-band? It looks like it has 18 bands... (Or at least
"18th band", no dash.)

Cheers,
Moritz

Patch hide | download patch | download mbox

diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index 04ec9b8..52c44d2 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -109,6 +109,7 @@  OBJS-$(CONFIG_SILENCEREMOVE_FILTER)          += af_silenceremove.o
 OBJS-$(CONFIG_SOFALIZER_FILTER)              += af_sofalizer.o
 OBJS-$(CONFIG_STEREOTOOLS_FILTER)            += af_stereotools.o
 OBJS-$(CONFIG_STEREOWIDEN_FILTER)            += af_stereowiden.o
+OBJS-$(CONFIG_SUPEREQUALIZER_FILTER)         += af_superequalizer.o
 OBJS-$(CONFIG_SURROUND_FILTER)               += af_surround.o
 OBJS-$(CONFIG_TREBLE_FILTER)                 += af_biquads.o
 OBJS-$(CONFIG_TREMOLO_FILTER)                += af_tremolo.o
diff --git a/libavfilter/af_superequalizer.c b/libavfilter/af_superequalizer.c
new file mode 100644
index 0000000..c569394
--- /dev/null
+++ b/libavfilter/af_superequalizer.c
@@ -0,0 +1,381 @@ 
+/*
+ * 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 "libavutil/opt.h"
+
+#include "libavcodec/avfft.h"
+
+#include "audio.h"
+#include "avfilter.h"
+#include "internal.h"
+
+#define NBANDS 17
+#define M 15
+
+typedef struct EqParameter {
+    float lower, upper, gain;
+} EqParameter;
+
+typedef struct SuperEqualizerContext {
+    const AVClass *class;
+
+    EqParameter params[NBANDS+1];
+
+    float gains[NBANDS+1];
+
+    float fact[M+1];
+    float aa;
+    float iza;
+    float *lires, *rires, *irest;
+    float *fsamples;
+    int chg_ires, cur_ires;
+    int winlen, winlenbit, tabsize;
+
+    AVFrame *in, *out;
+    RDFTContext *rdft, *irdft;
+} SuperEqualizerContext;
+
+static const float bands[] = {
+    65.406392, 92.498606, 130.81278, 184.99721, 261.62557, 369.99442, 523.25113, 739.9884, 1046.5023,
+    1479.9768, 2093.0045, 2959.9536, 4186.0091, 5919.9072, 8372.0181, 11839.814, 16744.036
+};
+
+static float izero(SuperEqualizerContext *s, float x)
+{
+    float ret = 1;
+    int m;
+
+    for (m = 1; m <= M; m++) {
+        float t;
+
+        t = pow(x / 2, m) / s->fact[m];
+        ret += t*t;
+    }
+
+    return ret;
+}
+
+static float hn_lpf(int n, float f, float fs)
+{
+    float t = 1 / fs;
+    float omega = 2 * M_PI * f;
+
+    if (n * omega * t == 0)
+        return 2 * f * t;
+    return 2 * f * t * sin(n * omega * t)/(n * omega * t);
+}
+
+static float hn_imp(int n)
+{
+    return n == 0 ? 1.0 : 0.0;
+}
+
+static float hn(int n, EqParameter *param, float fs)
+{
+    float ret, lhn;
+    int i;
+
+    lhn = hn_lpf(n, param[0].upper, fs);
+    ret = param[0].gain*lhn;
+
+    for (i = 1; i < NBANDS+1 && param[i].upper < fs / 2; i++) {
+        float lhn2 = hn_lpf(n, param[i].upper, fs);
+        ret += param[i].gain * (lhn2-lhn);
+        lhn = lhn2;
+    }
+
+    ret += param[i].gain * (hn_imp(n) - lhn);
+
+    return ret;
+}
+
+static float alpha(float a)
+{
+    if (a <= 21)
+        return 0;
+    if (a <= 50)
+        return .5842 * pow(a - 21,0.4) + 0.07886 * (a - 21);
+    return .1102 * (a - 8.7);
+}
+
+static float win(SuperEqualizerContext *s, float n, int N)
+{
+    return izero(s, alpha(s->aa) * sqrt(1 - 4 * n * n / ((N - 1) * (N - 1)))) / s->iza;
+}
+
+static void
+process_param(float *bc, EqParameter *param, float fs, int ch)
+{
+    int i;
+
+    for (i = 0; i <= NBANDS; i++) {
+        param[i].lower = i == 0 ? 0 : bands[i - 1];
+        param[i].upper = i == NBANDS - 1 ? fs : bands[i];
+        param[i].gain  = bc[i];
+    }
+}
+
+static int equ_init(SuperEqualizerContext *s, int wb)
+{
+    int i,j;
+
+    s->rdft  = av_rdft_init(wb, DFT_R2C);
+    s->irdft = av_rdft_init(wb, IDFT_C2R);
+    if (!s->rdft || !s->irdft)
+        return AVERROR(ENOMEM);
+
+    s->aa = 96;
+    s->winlen = (1 << (wb-1))-1;
+    s->winlenbit = wb;
+    s->tabsize  = 1 << wb;
+
+    s->lires    = av_calloc(s->tabsize, sizeof(float));
+    s->rires    = av_calloc(s->tabsize, sizeof(float));
+    s->irest    = av_calloc(s->tabsize, sizeof(float));
+    s->fsamples = av_calloc(s->tabsize, sizeof(float));
+
+    s->cur_ires = 1;
+    s->chg_ires = 1;
+
+    for (i = 0; i <= M; i++) {
+        s->fact[i] = 1;
+        for (j = 1; j <= i; j++)
+            s->fact[i] *= j;
+    }
+
+    s->iza = izero(s, alpha(s->aa));
+
+    return 0;
+}
+
+static void make_fir(SuperEqualizerContext *s, float *lbc, float *rbc, EqParameter *param, float fs)
+{
+    const int winlen = s->winlen;
+    const int tabsize = s->tabsize;
+    int i, cires = s->cur_ires;
+    float *nires;
+
+    if (fs <= 0)
+        return;
+
+    process_param(lbc, param, fs, 0);
+    for (i = 0; i < winlen; i++)
+        s->irest[i] = hn(i - winlen / 2, param, fs) * win(s, i - winlen / 2, winlen);
+    for (; i < tabsize; i++)
+        s->irest[i] = 0;
+
+    av_rdft_calc(s->rdft, s->irest);
+    nires = s->lires;
+    for (i = 0; i < tabsize; i++)
+        nires[i] = s->irest[i];
+
+    process_param(rbc, param, fs, 1);
+    for (i = 0; i < winlen; i++)
+        s->irest[i] = hn(i - winlen / 2, param, fs) * win(s, i - winlen / 2, winlen);
+    for (; i < tabsize; i++)
+        s->irest[i] = 0;
+
+    av_rdft_calc(s->rdft, s->irest);
+    nires = s->rires;
+    for (i = 0; i < tabsize; i++)
+        nires[i] = s->irest[i];
+    s->chg_ires = cires == 1 ? 2 : 1;
+}
+
+static int filter_frame(AVFilterLink *inlink, AVFrame *in)
+{
+    AVFilterContext *ctx = inlink->dst;
+    SuperEqualizerContext *s = ctx->priv;
+    AVFilterLink *outlink = ctx->outputs[0];
+    const float *ires = s->lires;
+    float *fsamples = s->fsamples;
+    int ch, i;
+
+    AVFrame *out = ff_get_audio_buffer(outlink, s->winlen);
+    float *src, *dst, *ptr;
+
+    if (!out) {
+        av_frame_free(&in);
+        return AVERROR(ENOMEM);
+    }
+
+    for (ch = 0; ch < in->channels; ch++) {
+        ptr = (float *)out->extended_data[ch];
+        dst = (float *)s->out->extended_data[ch];
+        src = (float *)in->extended_data[ch];
+
+        for (i = 0; i < s->winlen; i++)
+            fsamples[i] = src[i];
+        for (; i < s->tabsize; i++)
+            fsamples[i] = 0;
+
+        av_rdft_calc(s->rdft, fsamples);
+
+        fsamples[0] = ires[0] * fsamples[0];
+        fsamples[1] = ires[1] * fsamples[1];
+        for (i = 1; i < s->tabsize / 2; i++) {
+            float re, im;
+
+            re = ires[i*2  ] * fsamples[i*2] - ires[i*2+1] * fsamples[i*2+1];
+            im = ires[i*2+1] * fsamples[i*2] + ires[i*2  ] * fsamples[i*2+1];
+
+            fsamples[i*2  ] = re;
+            fsamples[i*2+1] = im;
+        }
+
+        av_rdft_calc(s->irdft, fsamples);
+
+        for (i = 0; i < s->winlen; i++)
+            dst[i] += fsamples[i] / s->tabsize * 2;
+        for (i = s->winlen; i < s->tabsize; i++)
+            dst[i]  = fsamples[i] / s->tabsize * 2;
+        for (i = 0; i < s->winlen; i++)
+            ptr[i] = dst[i];
+        for (i = 0; i < s->winlen; i++)
+            dst[i] = dst[i+s->winlen];
+    }
+
+    out->pts = in->pts;
+    av_frame_free(&in);
+
+    return ff_filter_frame(outlink, out);
+}
+
+static av_cold int init(AVFilterContext *ctx)
+{
+    SuperEqualizerContext *s = ctx->priv;
+
+    return equ_init(s, 14);
+}
+
+static int query_formats(AVFilterContext *ctx)
+{
+    AVFilterFormats *formats;
+    AVFilterChannelLayouts *layouts;
+    static const enum AVSampleFormat sample_fmts[] = {
+        AV_SAMPLE_FMT_FLTP,
+        AV_SAMPLE_FMT_NONE
+    };
+    int 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_make_format_list(sample_fmts);
+    if ((ret = ff_set_common_formats(ctx, formats)) < 0)
+        return ret;
+
+    formats = ff_all_samplerates();
+    return ff_set_common_samplerates(ctx, formats);
+}
+
+static int config_input(AVFilterLink *inlink)
+{
+    AVFilterContext *ctx = inlink->dst;
+    SuperEqualizerContext *s = ctx->priv;
+
+    inlink->partial_buf_size =
+    inlink->min_samples =
+    inlink->max_samples = s->winlen;
+
+    s->out = ff_get_audio_buffer(inlink, s->tabsize);
+    if (!s->out)
+        return AVERROR(ENOMEM);
+
+    return 0;
+}
+
+static int config_output(AVFilterLink *outlink)
+{
+    AVFilterContext *ctx = outlink->src;
+    SuperEqualizerContext *s = ctx->priv;
+
+    make_fir(s, s->gains, s->gains, s->params, outlink->sample_rate);
+
+    return 0;
+}
+
+static av_cold void uninit(AVFilterContext *ctx)
+{
+    SuperEqualizerContext *s = ctx->priv;
+
+    av_rdft_end(s->rdft);
+    av_rdft_end(s->irdft);
+}
+
+static const AVFilterPad superequalizer_inputs[] = {
+    {
+        .name         = "default",
+        .type         = AVMEDIA_TYPE_AUDIO,
+        .filter_frame = filter_frame,
+        .config_props = config_input,
+    },
+    { NULL }
+};
+
+static const AVFilterPad superequalizer_outputs[] = {
+    {
+        .name         = "default",
+        .type         = AVMEDIA_TYPE_AUDIO,
+        .config_props = config_output,
+    },
+    { NULL }
+};
+
+#define AF AV_OPT_FLAG_AUDIO_PARAM|AV_OPT_FLAG_FILTERING_PARAM
+#define OFFSET(x) offsetof(SuperEqualizerContext, x)
+
+static const AVOption superequalizer_options[] = {
+    {  "1b", "set  1st band gain", OFFSET(gains [0]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
+    {  "2b", "set  2nd band gain", OFFSET(gains [1]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
+    {  "3b", "set  3rd band gain", OFFSET(gains [2]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
+    {  "4b", "set  4th band gain", OFFSET(gains [3]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
+    {  "5b", "set  5th band gain", OFFSET(gains [4]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
+    {  "6b", "set  6th band gain", OFFSET(gains [5]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
+    {  "7b", "set  7th band gain", OFFSET(gains [6]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
+    {  "8b", "set  8th band gain", OFFSET(gains [7]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
+    {  "9b", "set  9th band gain", OFFSET(gains [8]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
+    { "10b", "set 10nd band gain", OFFSET(gains [9]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
+    { "11b", "set 11nd band gain", OFFSET(gains[10]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
+    { "12b", "set 12nd band gain", OFFSET(gains[11]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
+    { "13b", "set 13nd band gain", OFFSET(gains[12]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
+    { "14b", "set 14th band gain", OFFSET(gains[13]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
+    { "15b", "set 15th band gain", OFFSET(gains[14]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
+    { "16b", "set 16th band gain", OFFSET(gains[15]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
+    { "17b", "set 17th band gain", OFFSET(gains[16]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
+    { "18b", "set 18th band gain", OFFSET(gains[17]), AV_OPT_TYPE_FLOAT, {.dbl=1}, 0, 20, AF },
+    { NULL }
+};
+
+AVFILTER_DEFINE_CLASS(superequalizer);
+
+AVFilter ff_af_superequalizer = {
+    .name          = "superequalizer",
+    .description   = NULL_IF_CONFIG_SMALL("Apply 18-th band equalization filter."),
+    .priv_size     = sizeof(SuperEqualizerContext),
+    .priv_class    = &superequalizer_class,
+    .init          = init,
+    .query_formats = query_formats,
+    .uninit        = uninit,
+    .inputs        = superequalizer_inputs,
+    .outputs       = superequalizer_outputs,
+};
diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
index 94f7cf3..bd81091 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -122,6 +122,7 @@  static void register_all(void)
     REGISTER_FILTER(SOFALIZER,      sofalizer,      af);
     REGISTER_FILTER(STEREOTOOLS,    stereotools,    af);
     REGISTER_FILTER(STEREOWIDEN,    stereowiden,    af);
+    REGISTER_FILTER(SUPEREQUALIZER, superequalizer, af);
     REGISTER_FILTER(SURROUND,       surround,       af);
     REGISTER_FILTER(TREBLE,         treble,         af);
     REGISTER_FILTER(TREMOLO,        tremolo,        af);