[FFmpeg-devel,GSoC] audio filter-use cellular automata to generate tones

Use cellular automata to generate tones

On 3/18/20, Ashutosh Pradhan <ashutoshp012345@gmail.com> wrote:
> Use cellular automata to generate tones
>
> diff --git a/Changelog b/Changelog
> index d1572553a5..5ddd2484b0 100644
> --- a/Changelog
> +++ b/Changelog
> @@ -48,6 +48,7 @@ version <next>:
>  - AMQP 0-9-1 protocol (RabbitMQ)
>  - Vulkan support
>  - avgblur_vulkan, overlay_vulkan, scale_vulkan and chromaber_vulkan filters
> +- atone filter
>
>
>  version 4.2:
> diff --git a/configure b/configure
> index 6ceb0c7af4..2ec7e377f3 100755
> --- a/configure
> +++ b/configure
> @@ -233,6 +233,7 @@ External library support:
>                             and libraw1394 [no]
>    --enable-libfdk-aac      enable AAC de/encoding via libfdk-aac [no]
>    --enable-libflite        enable flite (voice synthesis) support via
> libflite [no]
> +  --enable-libfluidsynth   enable libfluidsynth support for fluidsynth [no]
>    --enable-libfontconfig   enable libfontconfig, useful for drawtext filter
> [no]
>    --enable-libfreetype     enable libfreetype, needed for drawtext filter
> [no]
>    --enable-libfribidi      enable libfribidi, improves drawtext filter [no]
> @@ -1770,6 +1771,7 @@ EXTERNAL_LIBRARY_LIST="
>      libdc1394
>      libdrm
>      libflite
> +    libfluidsynth
>      libfontconfig
>      libfreetype
>      libfribidi
> @@ -3465,6 +3467,7 @@ asr_filter_deps="pocketsphinx"
>  ass_filter_deps="libass"
>  atempo_filter_deps="avcodec"
>  atempo_filter_select="rdft"
> +atone_filter_deps="libfluidsynth"
>  avgblur_opencl_filter_deps="opencl"
>  avgblur_vulkan_filter_deps="vulkan libglslang"
>  azmq_filter_deps="libzmq"
> @@ -6270,6 +6273,7 @@ enabled libfdk_aac        && { check_pkg_config
> libfdk_aac fdk-aac "fdk-aac/aace
>                                   warn "using libfdk without pkg-config"; }
> }
>  flite_extralibs="-lflite_cmu_time_awb -lflite_cmu_us_awb -lflite_cmu_us_kal
> -lflite_cmu_us_kal16 -lflite_cmu_us_rms -lflite_cmu_us_slt -lflite_usenglish
> -lflite_cmulex -lflite"
>  enabled libflite          && require libflite "flite/flite.h" flite_init
> $flite_extralibs
> +enabled libfluidsynth     && require_pkg_config libfluidsynth fluidsynth
> "fluidsynth.h" fluid_log
>  enabled fontconfig        && enable libfontconfig
>  enabled libfontconfig     && require_pkg_config libfontconfig fontconfig
> "fontconfig/fontconfig.h" FcInit
>  enabled libfreetype       && require_pkg_config libfreetype freetype2
> "ft2build.h FT_FREETYPE_H" FT_Init_FreeType
> diff --git a/doc/filters.texi b/doc/filters.texi
> index 328e984e92..0eda8a4c6e 100644
> --- a/doc/filters.texi
> +++ b/doc/filters.texi
> @@ -6050,6 +6050,57 @@ anoisesrc=d=60:c=pink:r=44100:a=0.5
>  @end example
>  @end itemize
>
> +@section atone
> +
> +Generate legato tones using cellular automata.
> +To compile filter configure ffmpeg with @code{--enable-libfluidsynth}
> +
> +The filter accepts the following options:
> +
> +@table @option
> +@item sample_rate, r
> +Specify the sample rate. Default value is 44100 Hz.
> +
> +@item sfont
> +Specify the location of soundfont file. Default value is
> +"/usr/share/sounds/sf2/FluidR3_GM.sf2"(for linux).
> +
> +@item duration, d
> +Specify the duration of the generated audio stream. Not specifying this
> option
> +results in playing tones for infinite length.
> +
> +@item velocity, v
> +Specify the velocity of key press. Default value is 80.
> +
> +@item MIDI_channel
> +Specify the MIDI channel number between(0 to 16) to play tones. Default is
> 0.
> +
> +@item tone_change_interval, t
> +Specify the time interval between successive tones in seconds. Default is
> 0.2s.
> +
> +@item rule
> +Specify the rule between 0 to 255 to get the rule set. Default is 30.
> +
> +@item width
> +Specify the width of the cells array. Default is 64.
> +
> +@item samples_per_frame
> +Set the number of samples per each output frame. Default is 1024.
> +@end table
> +
> +@subsection Examples
> +
> +@itemize
> +
> +@item
> +Generate 10 seconds of random tones, with a key velocity of 100, midi
> channel 1
> +and an tone change interval of 0.1s:
> +@example
> +atone=d=10:MIDI_channel=1:v=100:t=0.1:sfont="example.sf2":rule=193
> +atone=duration=10:MIDI_channel=1:velocity=100:tone_change_interval=0.1:sfont="example.sf2":rule=193
> +@end example
> +@end itemize
> +
>  @section hilbert
>
>  Generate odd-tap Hilbert transform FIR coefficients.
> diff --git a/libavfilter/Makefile b/libavfilter/Makefile
> index 750412da6b..b1f0c4be35 100644
> --- a/libavfilter/Makefile
> +++ b/libavfilter/Makefile
> @@ -151,6 +151,7 @@ OBJS-$(CONFIG_FLITE_FILTER)                  +=
> asrc_flite.o
>  OBJS-$(CONFIG_HILBERT_FILTER)                += asrc_hilbert.o
>  OBJS-$(CONFIG_SINC_FILTER)                   += asrc_sinc.o
>  OBJS-$(CONFIG_SINE_FILTER)                   += asrc_sine.o
> +OBJS-$(CONFIG_ATONE_FILTER)                  += asrc_atone.o
>
>  OBJS-$(CONFIG_ANULLSINK_FILTER)              += asink_anullsink.o
>
> diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
> index 501e5d041b..4d3efc7e15 100644
> --- a/libavfilter/allfilters.c
> +++ b/libavfilter/allfilters.c
> @@ -145,6 +145,7 @@ extern AVFilter ff_asrc_flite;
>  extern AVFilter ff_asrc_hilbert;
>  extern AVFilter ff_asrc_sinc;
>  extern AVFilter ff_asrc_sine;
> +extern AVFilter ff_asrc_atone;
>
>  extern AVFilter ff_asink_anullsink;
>
> diff --git a/libavfilter/asrc_atone.c b/libavfilter/asrc_atone.c
> new file mode 100644
> index 0000000000..04465ec4e0
> --- /dev/null
> +++ b/libavfilter/asrc_atone.c
> @@ -0,0 +1,275 @@
> +/*
> + * 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 <float.h>
> +#include <stdio.h>
> +#include <string.h>
> +#include <fluidsynth.h>
> +#include <stdlib.h>
> +#include <unistd.h>
> +
> +#include "libavutil/avassert.h"
> +#include "libavutil/channel_layout.h"
> +#include "libavutil/eval.h"
> +#include "libavutil/opt.h"
> +#include "libavutil/lfg.h"
> +#include "libavutil/random_seed.h"
> +#include "audio.h"
> +#include "avfilter.h"
> +#include "internal.h"
> +
> +typedef struct AtoneContext
> +{
> +    const AVClass* class;
> +    int64_t duration;
> +    int nb_samples;
> +    int sample_rate;
> +    int64_t pts;
> +    int infinite;
> +
> +    fluid_settings_t* settings;
> +    fluid_synth_t* synth;
> +    char* sfont;                      ///< soundfont file
> +    int sfont_id;
> +    int midi_chan;                   ///< midi channel number
> +    int velocity;                   ///< velocity of key
> +    double changerate;              ///< get the time interval of changing
> tones
> +    int key[7];                     ///< play tones within one octave
> +    int *cells;
> +    int *nextgen;
> +    int rule;                       ///< Rule for changing value in
> cell(taking 3 cells at time hence 2^8 rules possible)
> +    int width;                      ///< width of the cell array
> +    int ruleset[8];                 ///< taking 3 cells at a time
> +}AtoneContext;
> +
> +#define CONTEXT AtoneContext
> +#define FLAGS AV_OPT_FLAG_AUDIO_PARAM|AV_OPT_FLAG_FILTERING_PARAM
> +
> +#define OPT_GENERIC(name, field, def, min, max, descr, type, deffield, ...)
> \
> +    { name, descr, offsetof(CONTEXT, field), AV_OPT_TYPE_ ## type,
> \
> +      { .deffield = def }, min, max, FLAGS, __VA_ARGS__ }
> +
> +#define OPT_INT(name, field, def, min, max, descr, ...) \
> +    OPT_GENERIC(name, field, def, min, max, descr, INT, i64, __VA_ARGS__)
> +
> +#define OPT_DBL(name, field, def, min, max, descr, ...) \
> +    OPT_GENERIC(name, field, def, min, max, descr, DOUBLE, dbl,
> __VA_ARGS__)
> +
> +#define OPT_DUR(name, field, def, min, max, descr, ...) \
> +    OPT_GENERIC(name, field, def, min, max, descr, DURATION, str,
> __VA_ARGS__)
> +
> +#define OPT_STR(name, field, def, min, max, descr, ...) \
> +    OPT_GENERIC(name, field, def, min, max, descr, STRING, str,
> __VA_ARGS__)
> +
> +static const AVOption atone_options[] = {
> +    OPT_INT("velocity",          velocity,       80,
>                   0, INT_MAX,             "set the velocity of key press",),
> +    OPT_INT("v",                 velocity,       80,
>                   0, INT_MAX,             "set the velocity of key press",),
> +    OPT_INT("sample_rate",       sample_rate,    44100,
>                   1, INT_MAX,             "set the sample rate",),
> +    OPT_INT("r",                 sample_rate,    44100,
>                   1, INT_MAX,             "set the sample rate",),
> +    OPT_DUR("duration",          duration,       0,
>                   0, INT64_MAX,           "set the audio duration",),
> +    OPT_DUR("d",                 duration,       0,
>                   0, INT64_MAX,           "set the audio duration",),
> +    OPT_STR("sfont",             sfont,
> "/usr/share/sounds/sf2/FluidR3_GM.sf2",      0, 0,                   "set
> the soundfont file",),
> +    OPT_INT("samples_per_frame", nb_samples,     1024,
>                   0, INT_MAX,             "set the number of samples per
> frame",),
> +    OPT_INT("MIDI_channel",      midi_chan,      0,
>                   0, 16,                  "set the MIDI Channel",),
> +    OPT_DBL("tone_change_interval",changerate,   0.2,
>                   0, DBL_MAX,             "set the random tone change time
> in seconds",),
> +    OPT_DBL("t",                 changerate,     0.2,
>                   0, DBL_MAX,             "set the random tone change time
> in seconds",),
> +    OPT_INT("rule",               rule,          30,
>                   0, 256,                 "set the rule for changing value
> in cell",),
> +    OPT_INT("width",             width,          64,
>                   0, INT_MAX,             "set the width of the cell
> array",),
> +    {NULL}
> +};
> +
> +AVFILTER_DEFINE_CLASS(atone);
> +
> +static av_cold int init(AVFilterContext *ctx)
> +{
> +    AtoneContext *s = ctx->priv;
> +    int val;
> +    /*Initialise the fluidsynth settings object followed by synthesizer*/
> +    s->settings = new_fluid_settings();
> +    if (s->settings== NULL){
> +        av_log(s, AV_LOG_ERROR, "Failed to create the fluidsynth
> settings");
> +        return AVERROR_EXTERNAL;
> +    }
> +
> +    s->synth = new_fluid_synth(s->settings);
> +    if (s->synth== NULL){
> +        av_log(s, AV_LOG_ERROR, "Failed to create the fluidsynth synth");
> +        return AVERROR_EXTERNAL;
> +    }
> +
> +    s->sfont_id= fluid_synth_sfload(s->synth, s->sfont, 1);
> +    if(s->sfont_id== FLUID_FAILED){
> +        av_log(s, AV_LOG_ERROR, "Loading the Soundfont Failed");
> +        return AVERROR_EXTERNAL;
> +    }
> +    if (!(s->cells = av_malloc(sizeof(int)*(s->width))))
> +        return AVERROR(ENOMEM);
> +
> +    if (!(s->nextgen = av_malloc(sizeof(int)*(s->width))))
> +        return AVERROR(ENOMEM);
> +
> +    for (int i=0; i<7; i++)
> +        s->key[i]=0;
> +
> +    s->changerate= s->changerate*s->sample_rate/s->nb_samples;
> +    if (s->changerate<1.0)
> +        s->changerate= 1.0;
> +
> +    for (int i=0; i< s->width; i++)
> +        s->cells[i]=0;
> +    s->cells[s->width/2]=1;
> +
> +    val= s->rule;
> +    for (int i=0; i<8; i++){
> +        s->ruleset[i]= val%2;
> +        val= val>>1;
> +    }
> +    return 0;
> +}
> +
> +static av_cold void uninit(AVFilterContext *ctx)
> +{
> +    AtoneContext *s = ctx->priv;
> +
> +    delete_fluid_synth(s->synth);
> +    delete_fluid_settings(s->settings);
> +    av_freep(&s->cells);
> +    av_freep(&s->nextgen);
> +}
> +
> +static av_cold int config_props(AVFilterLink *outlink)
> +{
> +    AtoneContext *s = outlink->src->priv;
> +
> +    if (s->duration==0)
> +        s->infinite=1;
> +    s->duration = av_rescale(s->duration, s->sample_rate, AV_TIME_BASE);
> +    return 0;
> +}
> +/*
> +Generate keys using cellular automata. Reference taken from
> http://tones.wolfram.com/about/how-it-works
> +to generate the cells. Rule 30(default) gives random numbers. Here middle
> part(+-3) of the cell generation
> +is taken to get the keys. All the keys generated in the octave are played
> at the same time.
> +*/
> +static void generate(AtoneContext* s){
> +    for (int i=0; i< s->width; i++){
> +        int c=0;
> +        for (int j=0; j<3; j++){
> +            c+= s->cells[(i+j-1)%s->width]<<(j);
> +        }
> +        s->nextgen[i]= s->ruleset[c];
> +    }
> +    memcpy(s->cells, s->nextgen, s->width*sizeof(int));
> +    memcpy(s->key, &s->cells[s->width/2-3], 7*sizeof(int));
> +}
> +
> +static void gen_random_tone_samples(AtoneContext* s, int buf_size, float*
> data){
> +
> +    for(int i=0; i<7; i++)
> +        if (s->key[i]==1)
> +            fluid_synth_noteon(s->synth, s->midi_chan, (64+i),
> s->velocity);//pp range octave
> +
> +    fluid_synth_write_float(s->synth, buf_size, data, 0, 2, data, 1, 2);
> +    fluid_synth_all_notes_off(s->synth, s->midi_chan);
> +}
> +
> +static int request_frame(AVFilterLink *outlink)
> +{
> +    AVFilterContext *ctx = outlink->src;
> +    AtoneContext *s = ctx->priv;
> +    AVFrame *frame;
> +    int  nb_samples;
> +    if (!s->infinite && s->duration <= 0) {
> +        return AVERROR_EOF;
> +    } else if (!s->infinite && s->duration < s->nb_samples) {
> +        nb_samples = s->duration;
> +    } else {
> +        nb_samples = s->nb_samples;
> +    }
> +
> +    if (!(frame = ff_get_audio_buffer(outlink, nb_samples)))
> +        return AVERROR(ENOMEM);
> +
> +    if (s->pts % ((int)(s->changerate)*nb_samples) == 0)
> +       generate(s);

Why this approach?

> +
> +    gen_random_tone_samples(s, nb_samples, (float*)frame->data[0]);
> +
> +    if (!s->infinite)
> +        s->duration -= nb_samples;
> +
> +    frame->pts = s->pts;
> +    s->pts    += nb_samples;
> +    return ff_filter_frame(outlink, frame);
> +}
> +
> +static av_cold int query_formats(AVFilterContext *ctx)
> +{
> +    AtoneContext *s = ctx->priv;
> +    static const int64_t chlayouts[] = { AV_CH_LAYOUT_STEREO, -1 };
> +    int sample_rates[] = { s->sample_rate, -1 };
> +    static const enum AVSampleFormat sample_fmts[] = { AV_SAMPLE_FMT_FLT,
> AV_SAMPLE_FMT_NONE};
> +    AVFilterFormats *formats;
> +    AVFilterChannelLayouts *layouts;
> +    int ret;
> +
> +    formats = ff_make_format_list(sample_fmts);
> +    if (!formats)
> +        return AVERROR(ENOMEM);
> +    ret = ff_set_common_formats (ctx, formats);
> +    if (ret < 0)
> +        return ret;
> +
> +    layouts = avfilter_make_format64_list(chlayouts);
> +    if (!layouts)
> +        return AVERROR(ENOMEM);
> +    ret = ff_set_common_channel_layouts(ctx, layouts);
> +    if (ret < 0)
> +        return ret;
> +
> +    formats = ff_make_format_list(sample_rates);
> +    if (!formats)
> +        return AVERROR(ENOMEM);
> +    return ff_set_common_samplerates(ctx, formats);
> +}
> +
> +
> +static const AVFilterPad atone_outputs[] = {
> +    {
> +        .name          = "default",
> +        .type          = AVMEDIA_TYPE_AUDIO,
> +        .request_frame = request_frame,
> +        .config_props  = config_props,
> +    },
> +    { NULL }
> +};
> +
> +AVFilter ff_asrc_atone = {
> +    .name          = "atone",
> +    .description   = NULL_IF_CONFIG_SMALL("Generate tones using cellular
> automata."),
> +    .query_formats = query_formats,
> +    .init          = init,
> +    .uninit        = uninit,
> +    .priv_size     = sizeof(AtoneContext),
> +    .inputs        = NULL,
> +    .outputs       = atone_outputs,
> +    .priv_class    = &atone_class,
> +};
> +
> +
> +
> +
> diff --git a/libavfilter/version.h b/libavfilter/version.h
> index 7b41018be7..4c4e8afe2d 100644
> --- a/libavfilter/version.h
> +++ b/libavfilter/version.h
> @@ -30,7 +30,7 @@
>  #include "libavutil/version.h"
>
>  #define LIBAVFILTER_VERSION_MAJOR   7
> -#define LIBAVFILTER_VERSION_MINOR  77
> +#define LIBAVFILTER_VERSION_MINOR  78
>  #define LIBAVFILTER_VERSION_MICRO 100
>
>
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".