lavfi: add showwavespic filter

This is a variant of showwaves. It is implemented as a different filter
so that the user is not allowed to use meaningless options which belong
to showwaves (such as rate).

Major edits done by Stefano Sabatini, from a patch by ubitux.

See thread:
From: Clément Bœsch <u@pkh.me>
To: ffmpeg-devel@ffmpeg.org
Date: Wed, 24 Dec 2014 15:03:26 +0100
Subject: [FFmpeg-devel] [PATCH] avfilter/showwaves: add single_pic option

6 files changed, 270 insertions, 36 deletions

diff --git a/Changelog b/Changelog
index f3b1266..3ab1d30 100644
--- a/Changelog
+++ b/Changelog
@@ -5,6 +5,7 @@ version <next>:
 - FFT video filter
 - TDSC decoder
 - DTS lossless extension (XLL) decoding (not lossless, disabled by default)
+- showwavespic filter
 
 
 version 2.6:
diff --git a/doc/filters.texi b/doc/filters.texi
index dbcd391..3acd3e8 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -11741,6 +11741,33 @@ aevalsrc=sin(1*2*PI*t)*sin(880*2*PI*t):cos(2*PI*200*t),asplit[out0],showwaves=r=
 @end example
 @end itemize
 
+@section showwavespic
+
+Convert input audio to a single video frame, representing the samples waves.
+
+The filter accepts the following options:
+
+@table @option
+@item size, s
+Specify the video size for the output. For the syntax of this option, check the
+@ref{video size syntax,,"Video size" section in the ffmpeg-utils manual,ffmpeg-utils}.
+Default value is @code{600x240}.
+
+@item split_channels
+Set if channels should be drawn separately or overlap. Default value is 0.
+@end table
+
+@subsection Examples
+
+@itemize
+@item
+Extract a channel split representation of the wave form of a whole audio track
+in a 1024x800 picture using @command{ffmpeg}:
+@example
+ffmpeg -i audio.flac -lavfi showwavespic=split_channels=1:s=1024x800 waveform.png
+@end example
+@end itemize
+
 @section split, asplit
 
 Split input into several identical outputs.
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index b184f07..2cde029 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -236,6 +236,7 @@ OBJS-$(CONFIG_CONCAT_FILTER)                 += avf_concat.o
 OBJS-$(CONFIG_SHOWCQT_FILTER)                += avf_showcqt.o
 OBJS-$(CONFIG_SHOWSPECTRUM_FILTER)           += avf_showspectrum.o
 OBJS-$(CONFIG_SHOWWAVES_FILTER)              += avf_showwaves.o
+OBJS-$(CONFIG_SHOWWAVESPIC_FILTER)           += avf_showwaves.o
 
 # multimedia sources
 OBJS-$(CONFIG_AMOVIE_FILTER)                 += src_movie.o
diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
index 043ac56..0288082 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -251,6 +251,7 @@ void avfilter_register_all(void)
     REGISTER_FILTER(SHOWCQT,        showcqt,        avf);
     REGISTER_FILTER(SHOWSPECTRUM,   showspectrum,   avf);
     REGISTER_FILTER(SHOWWAVES,      showwaves,      avf);
+    REGISTER_FILTER(SHOWWAVESPIC,   showwavespic,   avf);
 
     /* multimedia sources */
     REGISTER_FILTER(AMOVIE,         amovie,         avsrc);
diff --git a/libavfilter/avf_showwaves.c b/libavfilter/avf_showwaves.c
index 9cddc51..57a6b2e 100644
--- a/libavfilter/avf_showwaves.c
+++ b/libavfilter/avf_showwaves.c
@@ -23,6 +23,7 @@
  * audio to video multimedia filter
  */
 
+#include "libavutil/avassert.h"
 #include "libavutil/channel_layout.h"
 #include "libavutil/opt.h"
 #include "libavutil/parseutils.h"
@@ -40,6 +41,11 @@ enum ShowWavesMode {
     MODE_NB,
 };
 
+struct frame_node {
+    AVFrame *frame;
+    struct frame_node *next;
+};
+
 typedef struct {
     const AVClass *class;
     int w, h;
@@ -54,6 +60,13 @@ typedef struct {
     int split_channels;
     void (*draw_sample)(uint8_t *buf, int height, int linesize,
                         int16_t sample, int16_t *prev_y, int intensity);
+
+    /* single picture */
+    int single_pic;
+    struct frame_node *audio_frames;
+    struct frame_node *last_frame;
+    int64_t total_samples;
+    int64_t *sum; /* abs sum of the samples per channel */
 } ShowWavesContext;
 
 #define OFFSET(x) offsetof(ShowWavesContext, x)
@@ -82,6 +95,19 @@ static av_cold void uninit(AVFilterContext *ctx)
 
     av_frame_free(&showwaves->outpicref);
     av_freep(&showwaves->buf_idy);
+
+    if (showwaves->single_pic) {
+        struct frame_node *node = showwaves->audio_frames;
+        while (node) {
+            struct frame_node *tmp = node;
+
+            node = node->next;
+            av_frame_free(&tmp->frame);
+            av_freep(&tmp);
+        }
+        av_freep(&showwaves->sum);
+        showwaves->last_frame = NULL;
+    }
 }
 
 static int query_formats(AVFilterContext *ctx)
@@ -162,6 +188,55 @@ inline static int push_frame(AVFilterLink *outlink)
     return ret;
 }
 
+static int push_single_pic(AVFilterLink *outlink)
+{
+    AVFilterContext *ctx = outlink->src;
+    AVFilterLink *inlink = ctx->inputs[0];
+    ShowWavesContext *showwaves = ctx->priv;
+    int64_t n = 0, max_samples = showwaves->total_samples / outlink->w;
+    AVFrame *out = showwaves->outpicref;
+    struct frame_node *node;
+    const int nb_channels = inlink->channels;
+    const int x = 255 / (showwaves->split_channels ? 1 : nb_channels);
+    const int ch_height = showwaves->split_channels ? outlink->h / nb_channels : outlink->h;
+    const int linesize = out->linesize[0];
+    int col = 0;
+    int64_t *sum = showwaves->sum;
+
+    av_log(ctx, AV_LOG_DEBUG, "Create frame averaging %"PRId64" samples per column\n", max_samples);
+
+    memset(sum, 0, nb_channels);
+
+    for (node = showwaves->audio_frames; node; node = node->next) {
+        int i;
+        const AVFrame *frame = node->frame;
+        const int16_t *p = (const int16_t *)frame->data[0];
+
+        for (i = 0; i < frame->nb_samples; i++) {
+            int ch;
+
+            for (ch = 0; ch < nb_channels; ch++)
+                sum[ch] += abs(p[ch + i*nb_channels]) << 1;
+            if (n++ == max_samples) {
+                for (ch = 0; ch < nb_channels; ch++) {
+                    int16_t sample = sum[ch] / max_samples;
+                    uint8_t *buf = out->data[0] + col;
+                    if (showwaves->split_channels)
+                        buf += ch*ch_height*linesize;
+                    av_assert0(col < outlink->w);
+                    showwaves->draw_sample(buf, ch_height, linesize, sample, &showwaves->buf_idy[ch], x);
+                    sum[ch] = 0;
+                }
+                col++;
+                n = 0;
+            }
+        }
+    }
+
+    return push_frame(outlink);
+}
+
+
 static int request_frame(AVFilterLink *outlink)
 {
     ShowWavesContext *showwaves = outlink->src->priv;
@@ -173,8 +248,13 @@ static int request_frame(AVFilterLink *outlink)
         ret = ff_request_frame(inlink);
     } while (!showwaves->req_fullfilled && ret >= 0);
 
-    if (ret == AVERROR_EOF && showwaves->outpicref)
-        push_frame(outlink);
+    if (ret == AVERROR_EOF && showwaves->outpicref) {
+        if (showwaves->single_pic)
+            push_single_pic(outlink);
+        else
+            push_frame(outlink);
+    }
+
     return ret;
 }
 
@@ -229,14 +309,56 @@ static void draw_sample_cline(uint8_t *buf, int height, int linesize,
         buf[k * linesize] += intensity;
 }
 
-static int filter_frame(AVFilterLink *inlink, AVFrame *insamples)
+static int alloc_out_frame(ShowWavesContext *showwaves, const int16_t *p,
+                           const AVFilterLink *inlink, AVFilterLink *outlink,
+                           const AVFrame *in)
+{
+    if (!showwaves->outpicref) {
+        int j;
+        AVFrame *out = showwaves->outpicref =
+            ff_get_video_buffer(outlink, outlink->w, outlink->h);
+        if (!out)
+            return AVERROR(ENOMEM);
+        out->width  = outlink->w;
+        out->height = outlink->h;
+        out->pts = in->pts + av_rescale_q((p - (int16_t *)in->data[0]) / inlink->channels,
+                                          av_make_q(1, inlink->sample_rate),
+                                          outlink->time_base);
+        for (j = 0; j < outlink->h; j++)
+            memset(out->data[0] + j*out->linesize[0], 0, outlink->w);
+    }
+    return 0;
+}
+
+static av_cold int init(AVFilterContext *ctx)
+{
+    ShowWavesContext *showwaves = ctx->priv;
+
+    if (!strcmp(ctx->filter->name, "showwavespic")) {
+        showwaves->single_pic = 1;
+        showwaves->mode = MODE_CENTERED_LINE;
+    }
+
+    switch (showwaves->mode) {
+    case MODE_POINT:         showwaves->draw_sample = draw_sample_point; break;
+    case MODE_LINE:          showwaves->draw_sample = draw_sample_line;  break;
+    case MODE_P2P:           showwaves->draw_sample = draw_sample_p2p;   break;
+    case MODE_CENTERED_LINE: showwaves->draw_sample = draw_sample_cline; break;
+    default:
+        return AVERROR_BUG;
+    }
+    return 0;
+}
+
+#if CONFIG_SHOWWAVES_FILTER
+
+static int showwaves_filter_frame(AVFilterLink *inlink, AVFrame *insamples)
 {
     AVFilterContext *ctx = inlink->dst;
     AVFilterLink *outlink = ctx->outputs[0];
     ShowWavesContext *showwaves = ctx->priv;
     const int nb_samples = insamples->nb_samples;
     AVFrame *outpicref = showwaves->outpicref;
-    int linesize = outpicref ? outpicref->linesize[0] : 0;
     int16_t *p = (int16_t *)insamples->data[0];
     int nb_channels = inlink->channels;
     int i, j, ret = 0;
@@ -246,23 +368,15 @@ static int filter_frame(AVFilterLink *inlink, AVFrame *insamples)
 
     /* draw data in the buffer */
     for (i = 0; i < nb_samples; i++) {
-        if (!showwaves->outpicref) {
-            showwaves->outpicref = outpicref =
-                ff_get_video_buffer(outlink, outlink->w, outlink->h);
-            if (!outpicref)
-                return AVERROR(ENOMEM);
-            outpicref->width  = outlink->w;
-            outpicref->height = outlink->h;
-            outpicref->pts = insamples->pts +
-                             av_rescale_q((p - (int16_t *)insamples->data[0]) / nb_channels,
-                                          (AVRational){ 1, inlink->sample_rate },
-                                          outlink->time_base);
-            linesize = outpicref->linesize[0];
-            for (j = 0; j < outlink->h; j++)
-                memset(outpicref->data[0] + j * linesize, 0, outlink->w);
-        }
+
+        ret = alloc_out_frame(showwaves, p, inlink, outlink, insamples);
+        if (ret < 0)
+            goto end;
+        outpicref = showwaves->outpicref;
+
         for (j = 0; j < nb_channels; j++) {
             uint8_t *buf = outpicref->data[0] + showwaves->buf_idx;
+            const int linesize = outpicref->linesize[0];
             if (showwaves->split_channels)
                 buf += j*ch_height*linesize;
             showwaves->draw_sample(buf, ch_height, linesize, *p++,
@@ -280,30 +394,16 @@ static int filter_frame(AVFilterLink *inlink, AVFrame *insamples)
         outpicref = showwaves->outpicref;
     }
 
+end:
     av_frame_free(&insamples);
     return ret;
 }
 
-static av_cold int init(AVFilterContext *ctx)
-{
-    ShowWavesContext *showwaves = ctx->priv;
-
-    switch (showwaves->mode) {
-    case MODE_POINT:         showwaves->draw_sample = draw_sample_point; break;
-    case MODE_LINE:          showwaves->draw_sample = draw_sample_line;  break;
-    case MODE_P2P:           showwaves->draw_sample = draw_sample_p2p;   break;
-    case MODE_CENTERED_LINE: showwaves->draw_sample = draw_sample_cline; break;
-    default:
-        return AVERROR_BUG;
-    }
-    return 0;
-}
-
 static const AVFilterPad showwaves_inputs[] = {
     {
         .name         = "default",
         .type         = AVMEDIA_TYPE_AUDIO,
-        .filter_frame = filter_frame,
+        .filter_frame = showwaves_filter_frame,
     },
     { NULL }
 };
@@ -329,3 +429,107 @@ AVFilter ff_avf_showwaves = {
     .outputs       = showwaves_outputs,
     .priv_class    = &showwaves_class,
 };
+
+#endif // CONFIG_SHOWWAVES_FILTER
+
+#if CONFIG_SHOWWAVESPIC_FILTER
+
+#define OFFSET(x) offsetof(ShowWavesContext, x)
+#define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM
+
+static const AVOption showwavespic_options[] = {
+    { "size", "set video size", OFFSET(w), AV_OPT_TYPE_IMAGE_SIZE, {.str = "600x240"}, 0, 0, FLAGS },
+    { "s",    "set video size", OFFSET(w), AV_OPT_TYPE_IMAGE_SIZE, {.str = "600x240"}, 0, 0, FLAGS },
+    { "split_channels", "draw channels separately", OFFSET(split_channels), AV_OPT_TYPE_INT, {.i64 = 0}, 0, 1, FLAGS },
+    { NULL }
+};
+
+AVFILTER_DEFINE_CLASS(showwavespic);
+
+static int showwavespic_config_input(AVFilterLink *inlink)
+{
+    AVFilterContext *ctx = inlink->dst;
+    ShowWavesContext *showwaves = ctx->priv;
+
+    if (showwaves->single_pic) {
+        showwaves->sum = av_mallocz_array(inlink->channels, sizeof(*showwaves->sum));
+        if (!showwaves->sum)
+            return AVERROR(ENOMEM);
+    }
+
+    return 0;
+}
+
+static int showwavespic_filter_frame(AVFilterLink *inlink, AVFrame *insamples)
+{
+    AVFilterContext *ctx = inlink->dst;
+    AVFilterLink *outlink = ctx->outputs[0];
+    ShowWavesContext *showwaves = ctx->priv;
+    int16_t *p = (int16_t *)insamples->data[0];
+    int ret = 0;
+
+    if (showwaves->single_pic) {
+        struct frame_node *f;
+
+        ret = alloc_out_frame(showwaves, p, inlink, outlink, insamples);
+        if (ret < 0)
+            goto end;
+
+        /* queue the audio frame */
+        f = av_malloc(sizeof(*f));
+        if (!f) {
+            ret = AVERROR(ENOMEM);
+            goto end;
+        }
+        f->frame = insamples;
+        f->next = NULL;
+        if (!showwaves->last_frame) {
+            showwaves->audio_frames =
+            showwaves->last_frame   = f;
+        } else {
+            showwaves->last_frame->next = f;
+            showwaves->last_frame = f;
+        }
+        showwaves->total_samples += insamples->nb_samples;
+
+        return 0;
+    }
+
+end:
+    av_frame_free(&insamples);
+    return ret;
+}
+
+static const AVFilterPad showwavespic_inputs[] = {
+    {
+        .name         = "default",
+        .type         = AVMEDIA_TYPE_AUDIO,
+        .config_props = showwavespic_config_input,
+        .filter_frame = showwavespic_filter_frame,
+    },
+    { NULL }
+};
+
+static const AVFilterPad showwavespic_outputs[] = {
+    {
+        .name          = "default",
+        .type          = AVMEDIA_TYPE_VIDEO,
+        .config_props  = config_output,
+        .request_frame = request_frame,
+    },
+    { NULL }
+};
+
+AVFilter ff_avf_showwavespic = {
+    .name          = "showwavespic",
+    .description   = NULL_IF_CONFIG_SMALL("Convert input audio to a video output single picture."),
+    .init          = init,
+    .uninit        = uninit,
+    .query_formats = query_formats,
+    .priv_size     = sizeof(ShowWavesContext),
+    .inputs        = showwavespic_inputs,
+    .outputs       = showwavespic_outputs,
+    .priv_class    = &showwavespic_class,
+};
+
+#endif // CONFIG_SHOWWAVESPIC_FILTER
diff --git a/libavfilter/version.h b/libavfilter/version.h
index a964392..b349aaf 100644
--- a/libavfilter/version.h
+++ b/libavfilter/version.h
@@ -30,7 +30,7 @@
 #include "libavutil/version.h"
 
 #define LIBAVFILTER_VERSION_MAJOR  5
-#define LIBAVFILTER_VERSION_MINOR  12
+#define LIBAVFILTER_VERSION_MINOR  13
 #define LIBAVFILTER_VERSION_MICRO 100
 
 #define LIBAVFILTER_VERSION_INT AV_VERSION_INT(LIBAVFILTER_VERSION_MAJOR, \