aphaser filter
authorPaul B Mahol <onemda@gmail.com>
Sat, 30 Mar 2013 04:03:39 +0000 (04:03 +0000)
committerPaul B Mahol <onemda@gmail.com>
Wed, 3 Apr 2013 08:33:59 +0000 (08:33 +0000)
Signed-off-by: Paul B Mahol <onemda@gmail.com>
Changelog
doc/filters.texi
libavfilter/Makefile
libavfilter/af_aphaser.c [new file with mode: 0644]
libavfilter/allfilters.c
libavfilter/version.h

index 9f57061..5faa414 100644 (file)
--- a/Changelog
+++ b/Changelog
@@ -15,6 +15,7 @@ version <next>:
 - new ffmpeg options -filter_script and -filter_complex_script, which allow a
   filtergraph description to be read from a file
 - OpenCL support
+- audio phaser filter
 
 
 version 1.2:
index 401125b..08c1945 100644 (file)
@@ -6266,6 +6266,43 @@ following one, the permission might not be received as expected in that
 following filter. Inserting a @ref{format} or @ref{aformat} filter before the
 perms/aperms filter can avoid this problem.
 
+@section aphaser
+Add a phasing effect to the input audio.
+
+A phaser filter creates series of peaks and troughs in the frequency spectrum.
+The position of the peaks and troughs are modulated so that they vary over time, creating a sweeping effect.
+
+The filter accepts parameters as a list of @var{key}=@var{value}
+pairs, separated by ":".
+
+A description of the accepted parameters follows.
+
+@table @option
+@item in_gain
+Set input gain. Default is 0.4.
+
+@item out_gain
+Set output gain. Default is 0.74
+
+@item delay
+Set delay in milliseconds. Default is 3.0.
+
+@item decay
+Set decay. Default is 0.4.
+
+@item speed
+Set modulation speed in Hz. Default is 0.5.
+
+@item type
+Set modulation type. Default is triangular.
+
+It accepts the following values:
+@table @samp
+@item triangular, t
+@item sinusoidal, s
+@end table
+@end table
+
 @section aselect, select
 Select frames to pass in output.
 
index e865aef..9a12273 100644 (file)
@@ -58,6 +58,7 @@ OBJS-$(CONFIG_AMIX_FILTER)                   += af_amix.o
 OBJS-$(CONFIG_ANULL_FILTER)                  += af_anull.o
 OBJS-$(CONFIG_APAD_FILTER)                   += af_apad.o
 OBJS-$(CONFIG_APERMS_FILTER)                 += f_perms.o
+OBJS-$(CONFIG_APHASER_FILTER)                += af_aphaser.o
 OBJS-$(CONFIG_ARESAMPLE_FILTER)              += af_aresample.o
 OBJS-$(CONFIG_ASELECT_FILTER)                += f_select.o
 OBJS-$(CONFIG_ASENDCMD_FILTER)               += f_sendcmd.o
diff --git a/libavfilter/af_aphaser.c b/libavfilter/af_aphaser.c
new file mode 100644 (file)
index 0000000..141278f
--- /dev/null
@@ -0,0 +1,360 @@
+/*
+ * Copyright (c) 2013 Paul B Mahol
+ *
+ * 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
+ * phaser audio filter
+ */
+
+#include "libavutil/avassert.h"
+#include "libavutil/opt.h"
+#include "audio.h"
+#include "avfilter.h"
+#include "internal.h"
+
+enum WaveType {
+    WAVE_SIN,
+    WAVE_TRI,
+    WAVE_NB,
+};
+
+typedef struct AudioPhaserContext {
+    const AVClass *class;
+    double in_gain, out_gain;
+    double delay;
+    double decay;
+    double speed;
+
+    enum WaveType type;
+
+    int delay_buffer_length;
+    double *delay_buffer;
+
+    int modulation_buffer_length;
+    int32_t *modulation_buffer;
+
+    int delay_pos, modulation_pos;
+
+    void (*phaser)(struct AudioPhaserContext *p,
+                   uint8_t * const *src, uint8_t **dst,
+                   int nb_samples, int channels);
+} AudioPhaserContext;
+
+#define OFFSET(x) offsetof(AudioPhaserContext, x)
+#define FLAGS AV_OPT_FLAG_AUDIO_PARAM|AV_OPT_FLAG_FILTERING_PARAM
+
+static const AVOption aphaser_options[] = {
+    { "in_gain",  "set input gain",            OFFSET(in_gain),  AV_OPT_TYPE_DOUBLE, {.dbl=.4},  0,  1,   FLAGS },
+    { "out_gain", "set output gain",           OFFSET(out_gain), AV_OPT_TYPE_DOUBLE, {.dbl=.74}, 0,  1e9, FLAGS },
+    { "delay",    "set delay in milliseconds", OFFSET(delay),    AV_OPT_TYPE_DOUBLE, {.dbl=3.},  0,  5,   FLAGS },
+    { "decay",    "set decay",                 OFFSET(decay),    AV_OPT_TYPE_DOUBLE, {.dbl=.4},  0, .99,  FLAGS },
+    { "speed",    "set modulation speed",      OFFSET(speed),    AV_OPT_TYPE_DOUBLE, {.dbl=.5}, .1,  2,   FLAGS },
+    { "type",     "set modulation type",       OFFSET(type),     AV_OPT_TYPE_INT,    {.i64=WAVE_TRI}, 0, WAVE_NB-1, FLAGS, "type" },
+    { "triangular",  NULL, 0, AV_OPT_TYPE_CONST,  {.i64=WAVE_TRI}, 0, 0, FLAGS, "type" },
+    { "t",           NULL, 0, AV_OPT_TYPE_CONST,  {.i64=WAVE_TRI}, 0, 0, FLAGS, "type" },
+    { "sinusoidal",  NULL, 0, AV_OPT_TYPE_CONST,  {.i64=WAVE_SIN}, 0, 0, FLAGS, "type" },
+    { "s",           NULL, 0, AV_OPT_TYPE_CONST,  {.i64=WAVE_SIN}, 0, 0, FLAGS, "type" },
+    { NULL },
+};
+
+AVFILTER_DEFINE_CLASS(aphaser);
+
+static av_cold int init(AVFilterContext *ctx, const char *args)
+{
+    AudioPhaserContext *p = ctx->priv;
+
+    if (p->in_gain > (1 - p->decay * p->decay))
+        av_log(ctx, AV_LOG_WARNING, "in_gain may cause clipping\n");
+    if (p->in_gain / (1 - p->decay) > 1 / p->out_gain)
+        av_log(ctx, AV_LOG_WARNING, "out_gain may cause clipping\n");
+
+    return 0;
+}
+
+static int query_formats(AVFilterContext *ctx)
+{
+    AVFilterFormats *formats;
+    AVFilterChannelLayouts *layouts;
+    static const enum AVSampleFormat sample_fmts[] = {
+        AV_SAMPLE_FMT_DBL, AV_SAMPLE_FMT_DBLP,
+        AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_FLTP,
+        AV_SAMPLE_FMT_S32, AV_SAMPLE_FMT_S32P,
+        AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_S16P,
+        AV_SAMPLE_FMT_NONE
+    };
+
+    layouts = ff_all_channel_layouts();
+    if (!layouts)
+        return AVERROR(ENOMEM);
+    ff_set_common_channel_layouts(ctx, layouts);
+
+    formats = ff_make_format_list(sample_fmts);
+    if (!formats)
+        return AVERROR(ENOMEM);
+    ff_set_common_formats(ctx, formats);
+
+    formats = ff_all_samplerates();
+    if (!formats)
+        return AVERROR(ENOMEM);
+    ff_set_common_samplerates(ctx, formats);
+
+    return 0;
+}
+
+static void generate_wave_table(enum WaveType wave_type, enum AVSampleFormat sample_fmt,
+                                void *table, int table_size,
+                                double min, double max, double phase)
+{
+    uint32_t i, phase_offset = phase / M_PI / 2 * table_size + 0.5;
+
+    for (i = 0; i < table_size; i++) {
+        uint32_t point = (i + phase_offset) % table_size;
+        double d;
+
+        switch (wave_type) {
+        case WAVE_SIN:
+            d = (sin((double)point / table_size * 2 * M_PI) + 1) / 2;
+            break;
+        case WAVE_TRI:
+            d = (double)point * 2 / table_size;
+            switch (4 * point / table_size) {
+            case 0: d = d + 0.5; break;
+            case 1:
+            case 2: d = 1.5 - d; break;
+            case 3: d = d - 1.5; break;
+            }
+            break;
+        default:
+            av_assert0(0);
+        }
+
+        d  = d * (max - min) + min;
+        switch (sample_fmt) {
+        case AV_SAMPLE_FMT_FLT: {
+            float *fp = (float *)table;
+            *fp++ = (float)d;
+            table = fp;
+            continue; }
+        case AV_SAMPLE_FMT_DBL: {
+            double *dp = (double *)table;
+            *dp++ = d;
+            table = dp;
+            continue; }
+        }
+
+        d += d < 0 ? -0.5 : 0.5;
+        switch (sample_fmt) {
+        case AV_SAMPLE_FMT_S16: {
+            int16_t *sp = table;
+            *sp++ = (int16_t)d;
+            table = sp;
+            continue; }
+        case AV_SAMPLE_FMT_S32: {
+            int32_t *ip = table;
+            *ip++ = (int32_t)d;
+            table = ip;
+            continue; }
+        default:
+            av_assert0(0);
+        }
+    }
+}
+
+#define MOD(a, b) (((a) >= (b)) ? (a) - (b) : (a))
+
+#define PHASER_PLANAR(name, type)                                      \
+static void phaser_## name ##p(AudioPhaserContext *p,                  \
+                               uint8_t * const *src, uint8_t **dst,    \
+                               int nb_samples, int channels)           \
+{                                                                      \
+    int i, c, delay_pos, modulation_pos;                               \
+                                                                       \
+    for (c = 0; c < channels; c++) {                                   \
+        type *s = (type *)src[c];                                      \
+        type *d = (type *)dst[c];                                      \
+        double *buffer = p->delay_buffer +                             \
+                         c * p->delay_buffer_length;                   \
+                                                                       \
+        delay_pos      = p->delay_pos;                                 \
+        modulation_pos = p->modulation_pos;                            \
+                                                                       \
+        for (i = 0; i < nb_samples; i++, s++, d++) {                   \
+            double v = *s * p->in_gain + buffer[                       \
+                       MOD(delay_pos + p->modulation_buffer[           \
+                       modulation_pos],                                \
+                       p->delay_buffer_length)] * p->decay;            \
+                                                                       \
+            modulation_pos = MOD(modulation_pos + 1,                   \
+                             p->modulation_buffer_length);             \
+            delay_pos = MOD(delay_pos + 1, p->delay_buffer_length);    \
+            buffer[delay_pos] = v;                                     \
+                                                                       \
+            *d = v * p->out_gain;                                      \
+        }                                                              \
+    }                                                                  \
+                                                                       \
+    p->delay_pos      = delay_pos;                                     \
+    p->modulation_pos = modulation_pos;                                \
+}
+
+#define PHASER(name, type)                                              \
+static void phaser_## name (AudioPhaserContext *p,                      \
+                            uint8_t * const *src, uint8_t **dst,        \
+                            int nb_samples, int channels)               \
+{                                                                       \
+    int i, c, delay_pos, modulation_pos;                                \
+    type *s = (type *)src[0];                                           \
+    type *d = (type *)dst[0];                                           \
+    double *buffer = p->delay_buffer;                                   \
+                                                                        \
+    delay_pos      = p->delay_pos;                                      \
+    modulation_pos = p->modulation_pos;                                 \
+                                                                        \
+    for (i = 0; i < nb_samples; i++) {                                  \
+        int pos = MOD(delay_pos + p->modulation_buffer[modulation_pos], \
+                   p->delay_buffer_length) * channels;                  \
+        int npos;                                                       \
+                                                                        \
+        delay_pos = MOD(delay_pos + 1, p->delay_buffer_length);         \
+        npos = delay_pos * channels;                                    \
+        for (c = 0; c < channels; c++, s++, d++) {                      \
+            double v = *s * p->in_gain + buffer[pos + c] * p->decay;    \
+                                                                        \
+            buffer[npos + c] = v;                                       \
+                                                                        \
+            *d = v * p->out_gain;                                       \
+        }                                                               \
+                                                                        \
+        modulation_pos = MOD(modulation_pos + 1,                        \
+                         p->modulation_buffer_length);                  \
+    }                                                                   \
+                                                                        \
+    p->delay_pos      = delay_pos;                                      \
+    p->modulation_pos = modulation_pos;                                 \
+}
+
+PHASER_PLANAR(dbl, double)
+PHASER_PLANAR(flt, float)
+PHASER_PLANAR(s16, int16_t)
+PHASER_PLANAR(s32, int32_t)
+
+PHASER(dbl, double)
+PHASER(flt, float)
+PHASER(s16, int16_t)
+PHASER(s32, int32_t)
+
+static int config_output(AVFilterLink *outlink)
+{
+    AudioPhaserContext *p = outlink->src->priv;
+    AVFilterLink *inlink = outlink->src->inputs[0];
+
+    p->delay_buffer_length = p->delay * 0.001 * inlink->sample_rate + 0.5;
+    p->delay_buffer = av_calloc(p->delay_buffer_length, sizeof(*p->delay_buffer) * inlink->channels);
+    p->modulation_buffer_length = inlink->sample_rate / p->speed + 0.5;
+    p->modulation_buffer = av_malloc(p->modulation_buffer_length * sizeof(*p->modulation_buffer));
+
+    if (!p->modulation_buffer || !p->delay_buffer)
+        return AVERROR(ENOMEM);
+
+    generate_wave_table(p->type, AV_SAMPLE_FMT_S32,
+                        p->modulation_buffer, p->modulation_buffer_length,
+                        1., p->delay_buffer_length, M_PI / 2.0);
+
+    p->delay_pos = p->modulation_pos = 0;
+
+    switch (inlink->format) {
+    case AV_SAMPLE_FMT_DBL:  p->phaser = phaser_dbl;  break;
+    case AV_SAMPLE_FMT_DBLP: p->phaser = phaser_dblp; break;
+    case AV_SAMPLE_FMT_FLT:  p->phaser = phaser_flt;  break;
+    case AV_SAMPLE_FMT_FLTP: p->phaser = phaser_fltp; break;
+    case AV_SAMPLE_FMT_S16:  p->phaser = phaser_s16;  break;
+    case AV_SAMPLE_FMT_S16P: p->phaser = phaser_s16p; break;
+    case AV_SAMPLE_FMT_S32:  p->phaser = phaser_s32;  break;
+    case AV_SAMPLE_FMT_S32P: p->phaser = phaser_s32p; break;
+    default: av_assert0(0);
+    }
+
+    return 0;
+}
+
+static int filter_frame(AVFilterLink *inlink, AVFrame *inbuf)
+{
+    AudioPhaserContext *p = inlink->dst->priv;
+    AVFilterLink *outlink = inlink->dst->outputs[0];
+    AVFrame *outbuf;
+
+    if (av_frame_is_writable(inbuf)) {
+        outbuf = inbuf;
+    } else {
+        outbuf = ff_get_audio_buffer(inlink, inbuf->nb_samples);
+        if (!outbuf)
+            return AVERROR(ENOMEM);
+        av_frame_copy_props(outbuf, inbuf);
+    }
+
+    p->phaser(p, inbuf->extended_data, outbuf->extended_data,
+              outbuf->nb_samples, av_frame_get_channels(outbuf));
+
+    if (inbuf != outbuf)
+        av_frame_free(&inbuf);
+
+    return ff_filter_frame(outlink, outbuf);
+}
+
+static av_cold void uninit(AVFilterContext *ctx)
+{
+    AudioPhaserContext *p = ctx->priv;
+
+    av_freep(&p->delay_buffer);
+    av_freep(&p->modulation_buffer);
+}
+
+static const AVFilterPad aphaser_inputs[] = {
+    {
+        .name         = "default",
+        .type         = AVMEDIA_TYPE_AUDIO,
+        .filter_frame = filter_frame,
+    },
+    { NULL }
+};
+
+static const AVFilterPad aphaser_outputs[] = {
+    {
+        .name         = "default",
+        .type         = AVMEDIA_TYPE_AUDIO,
+        .config_props = config_output,
+    },
+    { NULL }
+};
+
+static const char *const shorthand[] = { "in_gain", "out_gain", "delay", "decay", "speed", "type", NULL };
+
+AVFilter avfilter_af_aphaser = {
+    .name          = "aphaser",
+    .description   = NULL_IF_CONFIG_SMALL("Add a phasing effect to the audio."),
+    .query_formats = query_formats,
+    .priv_size     = sizeof(AudioPhaserContext),
+    .init          = init,
+    .uninit        = uninit,
+    .inputs        = aphaser_inputs,
+    .outputs       = aphaser_outputs,
+    .priv_class    = &aphaser_class,
+    .shorthand     = shorthand,
+};
index 4ca180a..4972322 100644 (file)
@@ -54,6 +54,7 @@ void avfilter_register_all(void)
     REGISTER_FILTER(ANULL,          anull,          af);
     REGISTER_FILTER(APAD,           apad,           af);
     REGISTER_FILTER(APERMS,         aperms,         af);
+    REGISTER_FILTER(APHASER,        aphaser,        af);
     REGISTER_FILTER(ARESAMPLE,      aresample,      af);
     REGISTER_FILTER(ASELECT,        aselect,        af);
     REGISTER_FILTER(ASENDCMD,       asendcmd,       af);
index e623896..379b9ab 100644 (file)
@@ -29,8 +29,8 @@
 #include "libavutil/avutil.h"
 
 #define LIBAVFILTER_VERSION_MAJOR  3
-#define LIBAVFILTER_VERSION_MINOR  48
-#define LIBAVFILTER_VERSION_MICRO 105
+#define LIBAVFILTER_VERSION_MINOR  49
+#define LIBAVFILTER_VERSION_MICRO 100
 
 #define LIBAVFILTER_VERSION_INT AV_VERSION_INT(LIBAVFILTER_VERSION_MAJOR, \
                                                LIBAVFILTER_VERSION_MINOR, \