• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 /*
2  * Copyright (c) 2013 Paul B Mahol
3  *
4  * This file is part of FFmpeg.
5  *
6  * FFmpeg is free software; you can redistribute it and/or
7  * modify it under the terms of the GNU Lesser General Public
8  * License as published by the Free Software Foundation; either
9  * version 2.1 of the License, or (at your option) any later version.
10  *
11  * FFmpeg is distributed in the hope that it will be useful,
12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
14  * Lesser General Public License for more details.
15  *
16  * You should have received a copy of the GNU Lesser General Public
17  * License along with FFmpeg; if not, write to the Free Software
18  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19  */
20 
21 /**
22  * @file
23  * audio to video multimedia vectorscope filter
24  */
25 
26 #include "libavutil/avassert.h"
27 #include "libavutil/channel_layout.h"
28 #include "libavutil/opt.h"
29 #include "libavutil/parseutils.h"
30 #include "avfilter.h"
31 #include "filters.h"
32 #include "formats.h"
33 #include "audio.h"
34 #include "video.h"
35 #include "internal.h"
36 
37 enum VectorScopeMode {
38     LISSAJOUS,
39     LISSAJOUS_XY,
40     POLAR,
41     MODE_NB,
42 };
43 
44 enum VectorScopeDraw {
45     DOT,
46     LINE,
47     DRAW_NB,
48 };
49 
50 enum VectorScopeScale {
51     LIN,
52     SQRT,
53     CBRT,
54     LOG,
55     SCALE_NB,
56 };
57 
58 typedef struct AudioVectorScopeContext {
59     const AVClass *class;
60     AVFrame *outpicref;
61     int w, h;
62     int hw, hh;
63     int mode;
64     int draw;
65     int scale;
66     int contrast[4];
67     int fade[4];
68     double zoom;
69     int swap;
70     int mirror;
71     unsigned prev_x, prev_y;
72     AVRational frame_rate;
73     int nb_samples;
74 } AudioVectorScopeContext;
75 
76 #define OFFSET(x) offsetof(AudioVectorScopeContext, x)
77 #define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM
78 #define TFLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_RUNTIME_PARAM
79 
80 static const AVOption avectorscope_options[] = {
81     { "mode", "set mode", OFFSET(mode), AV_OPT_TYPE_INT, {.i64=LISSAJOUS}, 0, MODE_NB-1, TFLAGS, "mode" },
82     { "m",    "set mode", OFFSET(mode), AV_OPT_TYPE_INT, {.i64=LISSAJOUS}, 0, MODE_NB-1, TFLAGS, "mode" },
83     { "lissajous",    "", 0, AV_OPT_TYPE_CONST, {.i64=LISSAJOUS},    0, 0, TFLAGS, "mode" },
84     { "lissajous_xy", "", 0, AV_OPT_TYPE_CONST, {.i64=LISSAJOUS_XY}, 0, 0, TFLAGS, "mode" },
85     { "polar",        "", 0, AV_OPT_TYPE_CONST, {.i64=POLAR},        0, 0, TFLAGS, "mode" },
86     { "rate", "set video rate", OFFSET(frame_rate), AV_OPT_TYPE_VIDEO_RATE, {.str="25"}, 0, INT_MAX, FLAGS },
87     { "r",    "set video rate", OFFSET(frame_rate), AV_OPT_TYPE_VIDEO_RATE, {.str="25"}, 0, INT_MAX, FLAGS },
88     { "size", "set video size", OFFSET(w), AV_OPT_TYPE_IMAGE_SIZE, {.str="400x400"}, 0, 0, FLAGS },
89     { "s",    "set video size", OFFSET(w), AV_OPT_TYPE_IMAGE_SIZE, {.str="400x400"}, 0, 0, FLAGS },
90     { "rc", "set red contrast",   OFFSET(contrast[0]), AV_OPT_TYPE_INT, {.i64=40},  0, 255, TFLAGS },
91     { "gc", "set green contrast", OFFSET(contrast[1]), AV_OPT_TYPE_INT, {.i64=160}, 0, 255, TFLAGS },
92     { "bc", "set blue contrast",  OFFSET(contrast[2]), AV_OPT_TYPE_INT, {.i64=80},  0, 255, TFLAGS },
93     { "ac", "set alpha contrast", OFFSET(contrast[3]), AV_OPT_TYPE_INT, {.i64=255}, 0, 255, TFLAGS },
94     { "rf", "set red fade",       OFFSET(fade[0]), AV_OPT_TYPE_INT, {.i64=15}, 0, 255, TFLAGS },
95     { "gf", "set green fade",     OFFSET(fade[1]), AV_OPT_TYPE_INT, {.i64=10}, 0, 255, TFLAGS },
96     { "bf", "set blue fade",      OFFSET(fade[2]), AV_OPT_TYPE_INT, {.i64=5},  0, 255, TFLAGS },
97     { "af", "set alpha fade",     OFFSET(fade[3]), AV_OPT_TYPE_INT, {.i64=5},  0, 255, TFLAGS },
98     { "zoom", "set zoom factor",  OFFSET(zoom), AV_OPT_TYPE_DOUBLE, {.dbl=1},  0, 10, TFLAGS },
99     { "draw", "set draw mode", OFFSET(draw), AV_OPT_TYPE_INT, {.i64=DOT}, 0, DRAW_NB-1, TFLAGS, "draw" },
100     { "dot",   "", 0, AV_OPT_TYPE_CONST, {.i64=DOT} , 0, 0, TFLAGS, "draw" },
101     { "line",  "", 0, AV_OPT_TYPE_CONST, {.i64=LINE}, 0, 0, TFLAGS, "draw" },
102     { "scale", "set amplitude scale mode", OFFSET(scale), AV_OPT_TYPE_INT, {.i64=LIN}, 0, SCALE_NB-1, TFLAGS, "scale" },
103     { "lin",   "linear",      0, AV_OPT_TYPE_CONST, {.i64=LIN},  0, 0, TFLAGS, "scale" },
104     { "sqrt",  "square root", 0, AV_OPT_TYPE_CONST, {.i64=SQRT}, 0, 0, TFLAGS, "scale" },
105     { "cbrt",  "cube root",   0, AV_OPT_TYPE_CONST, {.i64=CBRT}, 0, 0, TFLAGS, "scale" },
106     { "log",   "logarithmic", 0, AV_OPT_TYPE_CONST, {.i64=LOG},  0, 0, TFLAGS, "scale" },
107     { "swap", "swap x axis with y axis", OFFSET(swap), AV_OPT_TYPE_BOOL, {.i64=1}, 0, 1, TFLAGS },
108     { "mirror", "mirror axis", OFFSET(mirror), AV_OPT_TYPE_INT, {.i64=0}, 0, 3, TFLAGS, "mirror" },
109     { "none",  "no mirror", 0, AV_OPT_TYPE_CONST, {.i64=0}, 0, 0, TFLAGS, "mirror" },
110     { "x",  "mirror x",     0, AV_OPT_TYPE_CONST, {.i64=1}, 0, 0, TFLAGS, "mirror" },
111     { "y",  "mirror y",     0, AV_OPT_TYPE_CONST, {.i64=2}, 0, 0, TFLAGS, "mirror" },
112     { "xy", "mirror both",  0, AV_OPT_TYPE_CONST, {.i64=3}, 0, 0, TFLAGS, "mirror" },
113     { NULL }
114 };
115 
116 AVFILTER_DEFINE_CLASS(avectorscope);
117 
draw_dot(AudioVectorScopeContext * s,unsigned x,unsigned y)118 static void draw_dot(AudioVectorScopeContext *s, unsigned x, unsigned y)
119 {
120     const int linesize = s->outpicref->linesize[0];
121     uint8_t *dst;
122 
123     if (s->zoom > 1) {
124         if (y >= s->h || x >= s->w)
125             return;
126     } else {
127         y = FFMIN(y, s->h - 1);
128         x = FFMIN(x, s->w - 1);
129     }
130 
131     dst = &s->outpicref->data[0][y * linesize + x * 4];
132     dst[0] = FFMIN(dst[0] + s->contrast[0], 255);
133     dst[1] = FFMIN(dst[1] + s->contrast[1], 255);
134     dst[2] = FFMIN(dst[2] + s->contrast[2], 255);
135     dst[3] = FFMIN(dst[3] + s->contrast[3], 255);
136 }
137 
draw_line(AudioVectorScopeContext * s,int x0,int y0,int x1,int y1)138 static void draw_line(AudioVectorScopeContext *s, int x0, int y0, int x1, int y1)
139 {
140     int dx = FFABS(x1-x0), sx = x0 < x1 ? 1 : -1;
141     int dy = FFABS(y1-y0), sy = y0 < y1 ? 1 : -1;
142     int err = (dx>dy ? dx : -dy) / 2, e2;
143 
144     for (;;) {
145         draw_dot(s, x0, y0);
146 
147         if (x0 == x1 && y0 == y1)
148             break;
149 
150         e2 = err;
151 
152         if (e2 >-dx) {
153             err -= dy;
154             x0 += sx;
155         }
156 
157         if (e2 < dy) {
158             err += dx;
159             y0 += sy;
160         }
161     }
162 }
163 
fade(AVFilterContext * ctx,void * arg,int jobnr,int nb_jobs)164 static int fade(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
165 {
166     AudioVectorScopeContext *s = ctx->priv;
167     const int linesize = s->outpicref->linesize[0];
168     const int height = s->outpicref->height;
169     const int slice_start = (height *  jobnr   ) / nb_jobs;
170     const int slice_end   = (height * (jobnr+1)) / nb_jobs;
171 
172     if (s->fade[0] == 255 && s->fade[1] == 255 && s->fade[2] == 255) {
173         for (int i = slice_start; i < slice_end; i++)
174             memset(s->outpicref->data[0] + i * linesize, 0, s->outpicref->width * 4);
175         return 0;
176     }
177 
178     if (s->fade[0] || s->fade[1] || s->fade[2]) {
179         uint8_t *d = s->outpicref->data[0] + slice_start * linesize;
180         for (int i = slice_start; i < slice_end; i++) {
181             for (int j = 0; j < s->w*4; j+=4) {
182                 if (d[j+0])
183                     d[j+0] = FFMAX(d[j+0] - s->fade[0], 0);
184                 if (d[j+1])
185                     d[j+1] = FFMAX(d[j+1] - s->fade[1], 0);
186                 if (d[j+2])
187                     d[j+2] = FFMAX(d[j+2] - s->fade[2], 0);
188                 if (d[j+3])
189                     d[j+3] = FFMAX(d[j+3] - s->fade[3], 0);
190             }
191             d += linesize;
192         }
193     }
194 
195     return 0;
196 }
197 
query_formats(AVFilterContext * ctx)198 static int query_formats(AVFilterContext *ctx)
199 {
200     AVFilterFormats *formats = NULL;
201     AVFilterChannelLayouts *layout = NULL;
202     AVFilterLink *inlink = ctx->inputs[0];
203     AVFilterLink *outlink = ctx->outputs[0];
204     static const enum AVSampleFormat sample_fmts[] = { AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_FLT, AV_SAMPLE_FMT_NONE };
205     static const enum AVPixelFormat pix_fmts[] = { AV_PIX_FMT_RGBA, AV_PIX_FMT_NONE };
206     int ret;
207 
208     formats = ff_make_format_list(sample_fmts);
209     if ((ret = ff_formats_ref         (formats, &inlink->outcfg.formats        )) < 0 ||
210         (ret = ff_add_channel_layout  (&layout, &(AVChannelLayout)AV_CHANNEL_LAYOUT_STEREO)) < 0 ||
211         (ret = ff_channel_layouts_ref (layout , &inlink->outcfg.channel_layouts)) < 0)
212         return ret;
213 
214     formats = ff_all_samplerates();
215     if ((ret = ff_formats_ref(formats, &inlink->outcfg.samplerates)) < 0)
216         return ret;
217 
218     formats = ff_make_format_list(pix_fmts);
219     if ((ret = ff_formats_ref(formats, &outlink->incfg.formats)) < 0)
220         return ret;
221 
222     return 0;
223 }
224 
config_input(AVFilterLink * inlink)225 static int config_input(AVFilterLink *inlink)
226 {
227     AVFilterContext *ctx = inlink->dst;
228     AudioVectorScopeContext *s = ctx->priv;
229 
230     s->nb_samples = FFMAX(1, av_rescale(inlink->sample_rate, s->frame_rate.den, s->frame_rate.num));
231 
232     return 0;
233 }
234 
config_output(AVFilterLink * outlink)235 static int config_output(AVFilterLink *outlink)
236 {
237     AudioVectorScopeContext *s = outlink->src->priv;
238 
239     outlink->w = s->w;
240     outlink->h = s->h;
241     outlink->sample_aspect_ratio = (AVRational){1,1};
242     outlink->frame_rate = s->frame_rate;
243     outlink->time_base = av_inv_q(outlink->frame_rate);
244 
245     s->prev_x = s->hw = s->w / 2;
246     s->prev_y = s->hh = s->mode == POLAR ? s->h - 1 : s->h / 2;
247 
248     return 0;
249 }
250 
filter_frame(AVFilterLink * inlink,AVFrame * insamples)251 static int filter_frame(AVFilterLink *inlink, AVFrame *insamples)
252 {
253     AVFilterContext *ctx = inlink->dst;
254     AVFilterLink *outlink = ctx->outputs[0];
255     AudioVectorScopeContext *s = ctx->priv;
256     const int hw = s->hw;
257     const int hh = s->hh;
258     AVFrame *clone;
259     unsigned x, y;
260     unsigned prev_x = s->prev_x, prev_y = s->prev_y;
261     double zoom = s->zoom;
262     int i;
263 
264     if (!s->outpicref || s->outpicref->width  != outlink->w ||
265                          s->outpicref->height != outlink->h) {
266         av_frame_free(&s->outpicref);
267         s->outpicref = ff_get_video_buffer(outlink, outlink->w, outlink->h);
268         if (!s->outpicref) {
269             av_frame_free(&insamples);
270             return AVERROR(ENOMEM);
271         }
272 
273         s->outpicref->sample_aspect_ratio = (AVRational){1,1};
274         for (i = 0; i < outlink->h; i++)
275             memset(s->outpicref->data[0] + i * s->outpicref->linesize[0], 0, outlink->w * 4);
276     }
277     s->outpicref->pts = av_rescale_q(insamples->pts, inlink->time_base, outlink->time_base);
278 
279     av_frame_make_writable(s->outpicref);
280     ff_filter_execute(ctx, fade, NULL, NULL, FFMIN(outlink->h, ff_filter_get_nb_threads(ctx)));
281 
282     if (zoom < 1) {
283         float max = 0;
284 
285         switch (insamples->format) {
286         case AV_SAMPLE_FMT_S16: {
287             int16_t *samples = (int16_t *)insamples->data[0];
288 
289             for (i = 0; i < insamples->nb_samples * 2; i++) {
290                 float sample = samples[i] / (float)INT16_MAX;
291                 max = FFMAX(FFABS(sample), max);
292             }
293 
294             }
295             break;
296         case AV_SAMPLE_FMT_FLT: {
297             float *samples = (float *)insamples->data[0];
298 
299             for (i = 0; i < insamples->nb_samples * 2; i++) {
300                 max = FFMAX(FFABS(samples[i]), max);
301             }
302             }
303             break;
304         default:
305             av_assert2(0);
306         }
307 
308         switch (s->scale) {
309         case SQRT:
310             max = sqrtf(max);
311             break;
312         case CBRT:
313             max = cbrtf(max);
314             break;
315         case LOG:
316             max = logf(1 + max) / logf(2);
317             break;
318         }
319 
320         zoom = 1. / max;
321     }
322 
323     for (i = 0; i < insamples->nb_samples; i++) {
324         int16_t *samples = (int16_t *)insamples->data[0] + i * 2;
325         float *samplesf = (float *)insamples->data[0] + i * 2;
326         float src[2];
327 
328         switch (insamples->format) {
329         case AV_SAMPLE_FMT_S16:
330             src[0] = samples[0] / (float)INT16_MAX;
331             src[1] = samples[1] / (float)INT16_MAX;
332             break;
333         case AV_SAMPLE_FMT_FLT:
334             src[0] = samplesf[0];
335             src[1] = samplesf[1];
336             break;
337         default:
338             av_assert2(0);
339         }
340 
341         switch (s->scale) {
342         case SQRT:
343             src[0] = FFSIGN(src[0]) * sqrtf(FFABS(src[0]));
344             src[1] = FFSIGN(src[1]) * sqrtf(FFABS(src[1]));
345             break;
346         case CBRT:
347             src[0] = FFSIGN(src[0]) * cbrtf(FFABS(src[0]));
348             src[1] = FFSIGN(src[1]) * cbrtf(FFABS(src[1]));
349             break;
350         case LOG:
351             src[0] = FFSIGN(src[0]) * logf(1 + FFABS(src[0])) / logf(2);
352             src[1] = FFSIGN(src[1]) * logf(1 + FFABS(src[1])) / logf(2);
353             break;
354         }
355 
356         if (s->mirror & 1)
357             src[0] = -src[0];
358 
359         if (s->mirror & 2)
360             src[1] = -src[1];
361 
362         if (s->swap)
363             FFSWAP(float, src[0], src[1]);
364 
365         if (s->mode == LISSAJOUS) {
366             x = ((src[1] - src[0]) * zoom / 2 + 1) * hw;
367             y = (1.0 - (src[0] + src[1]) * zoom / 2) * hh;
368         } else if (s->mode == LISSAJOUS_XY) {
369             x = (src[1] * zoom + 1) * hw;
370             y = (src[0] * zoom + 1) * hh;
371         } else {
372             float sx, sy, cx, cy;
373 
374             sx = src[1] * zoom;
375             sy = src[0] * zoom;
376             cx = sx * sqrtf(1 - 0.5 * sy * sy);
377             cy = sy * sqrtf(1 - 0.5 * sx * sx);
378             x = hw + hw * FFSIGN(cx + cy) * (cx - cy) * .7;
379             y = s->h - s->h * fabsf(cx + cy) * .7;
380         }
381 
382         if (s->draw == DOT) {
383             draw_dot(s, x, y);
384         } else {
385             draw_line(s, x, y, prev_x, prev_y);
386         }
387         prev_x = x;
388         prev_y = y;
389     }
390 
391     s->prev_x = x, s->prev_y = y;
392     av_frame_free(&insamples);
393 
394     clone = av_frame_clone(s->outpicref);
395     if (!clone)
396         return AVERROR(ENOMEM);
397 
398     return ff_filter_frame(outlink, clone);
399 }
400 
activate(AVFilterContext * ctx)401 static int activate(AVFilterContext *ctx)
402 {
403     AVFilterLink *inlink = ctx->inputs[0];
404     AVFilterLink *outlink = ctx->outputs[0];
405     AudioVectorScopeContext *s = ctx->priv;
406     AVFrame *in;
407     int ret;
408 
409     FF_FILTER_FORWARD_STATUS_BACK(outlink, inlink);
410 
411     ret = ff_inlink_consume_samples(inlink, s->nb_samples, s->nb_samples, &in);
412     if (ret < 0)
413         return ret;
414     if (ret > 0)
415         return filter_frame(inlink, in);
416 
417     if (ff_inlink_queued_samples(inlink) >= s->nb_samples) {
418         ff_filter_set_ready(ctx, 10);
419         return 0;
420     }
421 
422     FF_FILTER_FORWARD_STATUS(inlink, outlink);
423     FF_FILTER_FORWARD_WANTED(outlink, inlink);
424 
425     return FFERROR_NOT_READY;
426 }
427 
uninit(AVFilterContext * ctx)428 static av_cold void uninit(AVFilterContext *ctx)
429 {
430     AudioVectorScopeContext *s = ctx->priv;
431 
432     av_frame_free(&s->outpicref);
433 }
434 
435 static const AVFilterPad audiovectorscope_inputs[] = {
436     {
437         .name         = "default",
438         .type         = AVMEDIA_TYPE_AUDIO,
439         .config_props = config_input,
440     },
441 };
442 
443 static const AVFilterPad audiovectorscope_outputs[] = {
444     {
445         .name         = "default",
446         .type         = AVMEDIA_TYPE_VIDEO,
447         .config_props = config_output,
448     },
449 };
450 
451 const AVFilter ff_avf_avectorscope = {
452     .name          = "avectorscope",
453     .description   = NULL_IF_CONFIG_SMALL("Convert input audio to vectorscope video output."),
454     .uninit        = uninit,
455     .priv_size     = sizeof(AudioVectorScopeContext),
456     .activate      = activate,
457     FILTER_INPUTS(audiovectorscope_inputs),
458     FILTER_OUTPUTS(audiovectorscope_outputs),
459     FILTER_QUERY_FUNC(query_formats),
460     .priv_class    = &avectorscope_class,
461     .flags         = AVFILTER_FLAG_SLICE_THREADS,
462     .process_command = ff_filter_process_command,
463 };
464