FFmpeg coverage

Directory: ../../../ffmpeg/
File: src/libavfilter/avf_a3dscope.c
Date: 2024-11-20 23:03:26
Exec Total Coverage
Lines: 0 144 0.0%
Functions: 0 12 0.0%
Branches: 0 38 0.0%
Line Branch Exec Source
1 /*
2 * Copyright (c) 2022 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 #include "libavutil/avassert.h"
22 #include "libavutil/channel_layout.h"
23 #include "libavutil/opt.h"
24 #include "libavutil/parseutils.h"
25 #include "avfilter.h"
26 #include "filters.h"
27 #include "formats.h"
28 #include "audio.h"
29 #include "video.h"
30
31 typedef struct Audio3dScopeContext {
32 const AVClass *class;
33 int w, h;
34 int size;
35 float fov;
36 float roll;
37 float pitch;
38 float yaw;
39 float zoom[3];
40 float eye[3];
41
42 AVRational frame_rate;
43 int nb_samples;
44
45 float view_matrix[4][4];
46 float projection_matrix[4][4];
47
48 AVFrame *frames[60];
49 } Audio3dScopeContext;
50
51 #define OFFSET(x) offsetof(Audio3dScopeContext, x)
52 #define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM
53 #define TFLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_RUNTIME_PARAM
54
55 static const AVOption a3dscope_options[] = {
56 { "rate", "set video rate", OFFSET(frame_rate), AV_OPT_TYPE_VIDEO_RATE, {.str="25"}, 0, INT_MAX, FLAGS },
57 { "r", "set video rate", OFFSET(frame_rate), AV_OPT_TYPE_VIDEO_RATE, {.str="25"}, 0, INT_MAX, FLAGS },
58 { "size", "set video size", OFFSET(w), AV_OPT_TYPE_IMAGE_SIZE, {.str="hd720"}, 0, 0, FLAGS },
59 { "s", "set video size", OFFSET(w), AV_OPT_TYPE_IMAGE_SIZE, {.str="hd720"}, 0, 0, FLAGS },
60 { "fov", "set camera FoV", OFFSET(fov), AV_OPT_TYPE_FLOAT, {.dbl=90.f}, 40, 150, TFLAGS },
61 { "roll", "set camera roll",OFFSET(roll), AV_OPT_TYPE_FLOAT, {.dbl=0.f}, -180, 180, TFLAGS },
62 { "pitch","set camera pitch",OFFSET(pitch), AV_OPT_TYPE_FLOAT, {.dbl=0.f}, -180, 180, TFLAGS },
63 { "yaw", "set camera yaw", OFFSET(yaw), AV_OPT_TYPE_FLOAT, {.dbl=0.f}, -180, 180, TFLAGS },
64 { "xzoom","set camera zoom", OFFSET(zoom[0]),AV_OPT_TYPE_FLOAT, {.dbl=1.f}, 0.01, 10, TFLAGS },
65 { "yzoom","set camera zoom", OFFSET(zoom[1]),AV_OPT_TYPE_FLOAT, {.dbl=1.f}, 0.01, 10, TFLAGS },
66 { "zzoom","set camera zoom", OFFSET(zoom[2]),AV_OPT_TYPE_FLOAT, {.dbl=1.f}, 0.01, 10, TFLAGS },
67 { "xpos", "set camera position", OFFSET(eye[0]), AV_OPT_TYPE_FLOAT, {.dbl=0.f},-60.f, 60.f, TFLAGS },
68 { "ypos", "set camera position", OFFSET(eye[1]), AV_OPT_TYPE_FLOAT, {.dbl=0.f},-60.f, 60.f, TFLAGS },
69 { "zpos", "set camera position", OFFSET(eye[2]), AV_OPT_TYPE_FLOAT, {.dbl=0.f},-60.f, 60.f, TFLAGS },
70 { "length","set length", OFFSET(size), AV_OPT_TYPE_INT, {.i64=15}, 1, 60, FLAGS },
71 { NULL }
72 };
73
74 AVFILTER_DEFINE_CLASS(a3dscope);
75
76 ✗ static int query_formats(const AVFilterContext *ctx,
77 AVFilterFormatsConfig **cfg_in,
78 AVFilterFormatsConfig **cfg_out)
79 {
80 ✗ AVFilterFormats *formats = NULL;
81 static const enum AVSampleFormat sample_fmts[] = { AV_SAMPLE_FMT_FLTP, AV_SAMPLE_FMT_NONE };
82 static const enum AVPixelFormat pix_fmts[] = { AV_PIX_FMT_RGBA, AV_PIX_FMT_NONE };
83 int ret;
84
85 ✗ formats = ff_make_format_list(sample_fmts);
86 ✗ if ((ret = ff_formats_ref(formats, &cfg_in[0]->formats)) < 0)
87 ✗ return ret;
88
89 ✗ formats = ff_make_format_list(pix_fmts);
90 ✗ if ((ret = ff_formats_ref(formats, &cfg_out[0]->formats)) < 0)
91 ✗ return ret;
92
93 ✗ return 0;
94 }
95
96 ✗ static int config_input(AVFilterLink *inlink)
97 {
98 ✗ AVFilterContext *ctx = inlink->dst;
99 ✗ Audio3dScopeContext *s = ctx->priv;
100
101 ✗ s->nb_samples = FFMAX(1, av_rescale(inlink->sample_rate, s->frame_rate.den, s->frame_rate.num));
102
103 ✗ return 0;
104 }
105
106 ✗ static int config_output(AVFilterLink *outlink)
107 {
108 ✗ FilterLink *l = ff_filter_link(outlink);
109 ✗ Audio3dScopeContext *s = outlink->src->priv;
110
111 ✗ outlink->w = s->w;
112 ✗ outlink->h = s->h;
113 ✗ outlink->sample_aspect_ratio = (AVRational){1,1};
114 ✗ l->frame_rate = s->frame_rate;
115 ✗ outlink->time_base = av_inv_q(l->frame_rate);
116
117 ✗ return 0;
118 }
119
120 ✗ static void projection_matrix(float fov, float a, float near, float far,
121 float matrix[4][4])
122 {
123 float f;
124
125 ✗ memset(matrix, 0, sizeof(*matrix));
126
127 ✗ f = 1.0f / tanf(fov * 0.5f * M_PI / 180.f);
128 ✗ matrix[0][0] = f * a;
129 ✗ matrix[1][1] = f;
130 ✗ matrix[2][2] = -(far + near) / (far - near);
131 ✗ matrix[2][3] = -1.f;
132 ✗ matrix[3][2] = -(near * far) / (far - near);
133 ✗ }
134
135 ✗ static inline void vmultiply(const float v[4], const float m[4][4], float d[4])
136 {
137 ✗ d[0] = v[0] * m[0][0] + v[1] * m[1][0] + v[2] * m[2][0] + v[3] * m[3][0];
138 ✗ d[1] = v[0] * m[0][1] + v[1] * m[1][1] + v[2] * m[2][1] + v[3] * m[3][1];
139 ✗ d[2] = v[0] * m[0][2] + v[1] * m[1][2] + v[2] * m[2][2] + v[3] * m[3][2];
140 ✗ d[3] = v[0] * m[0][3] + v[1] * m[1][3] + v[2] * m[2][3] + v[3] * m[3][3];
141 ✗ }
142
143 ✗ static void mmultiply(const float m2[4][4], const float m1[4][4], float m[4][4])
144 {
145 ✗ vmultiply(m2[0], m1, m[0]);
146 ✗ vmultiply(m2[1], m1, m[1]);
147 ✗ vmultiply(m2[2], m1, m[2]);
148 ✗ vmultiply(m2[3], m1, m[3]);
149 ✗ }
150
151 ✗ static float vdot(const float x[3], const float y[3])
152 {
153 ✗ return x[0] * y[0] + x[1] * y[1] + x[2] * y[2];
154 }
155
156 ✗ static void view_matrix(const float eye[3],
157 const float z[3],
158 const float roll,
159 const float pitch,
160 const float yaw, float m[4][4])
161 {
162 ✗ float cr = cosf(roll * M_PI / 180.f);
163 ✗ float sr = sinf(roll * M_PI / 180.f);
164 ✗ float cp = cosf(pitch * M_PI / 180.f);
165 ✗ float sp = sinf(pitch * M_PI / 180.f);
166 ✗ float cy = cosf(yaw * M_PI / 180.f);
167 ✗ float sy = sinf(yaw * M_PI / 180.f);
168 float t[4][4];
169 ✗ float rx[4][4] = {
170 ✗ {z[0], 0.f, 0.f, 0.f },
171 ✗ { 0.f, cy, -sy, 0.f },
172 { 0.f, sy, cy, 0.f },
173 { 0.f, 0.f, 0.f, 1.f },
174 };
175
176 ✗ float ry[4][4] = {
177 { cp, 0.f, sp, 0.f },
178 ✗ { 0.f,z[1], 0.f, 0.f },
179 ✗ {-sp, 0.f, cp, 0.f },
180 { 0.f, 0.f, 0.f, 1.f },
181 };
182
183 ✗ float rz[4][4] = {
184 ✗ { cr, -sr, 0.f, 0.f },
185 { sr, cr, 0.f, 0.f },
186 ✗ { 0.f, 0.f,z[2], 0.f },
187 { 0.f, 0.f, 0.f, 1.f },
188 };
189
190 ✗ memset(m, 0, sizeof(*m));
191
192 ✗ mmultiply(rx, ry, t);
193 ✗ mmultiply(rz, t, m);
194
195 ✗ m[3][0] = -vdot(m[0], eye);
196 ✗ m[3][1] = -vdot(m[1], eye);
197 ✗ m[3][2] = -vdot(m[2], eye);
198 ✗ }
199
200 ✗ static void draw_dot(AVFrame *out, unsigned x, unsigned y, float z,
201 int r, int g, int b)
202 {
203 ✗ const ptrdiff_t linesize = out->linesize[0];
204 uint8_t *dst;
205
206 ✗ dst = out->data[0] + y * linesize + x * 4;
207 ✗ dst[0] = r * z;
208 ✗ dst[1] = g * z;
209 ✗ dst[2] = b * z;
210 ✗ dst[3] = 255 * z;
211 ✗ }
212
213 ✗ static int filter_frame(AVFilterLink *inlink, AVFrame *in)
214 {
215 ✗ AVFilterContext *ctx = inlink->dst;
216 ✗ AVFilterLink *outlink = ctx->outputs[0];
217 ✗ Audio3dScopeContext *s = ctx->priv;
218 ✗ const float half_height = (s->h - 1) * 0.5f;
219 ✗ const float half_width = (s->w - 1) * 0.5f;
220 float matrix[4][4];
221 ✗ const int w = s->w;
222 ✗ const int h = s->h;
223 AVFrame *out;
224
225 ✗ out = ff_get_video_buffer(outlink, outlink->w, outlink->h);
226 ✗ if (!out) {
227 ✗ av_frame_free(&in);
228 ✗ return AVERROR(ENOMEM);
229 }
230
231 ✗ s->frames[0] = in;
232
233 ✗ out->sample_aspect_ratio = (AVRational){1,1};
234 ✗ for (int y = 0; y < outlink->h; y++)
235 ✗ memset(out->data[0] + y * out->linesize[0], 0, outlink->w * 4);
236 ✗ out->pts = av_rescale_q(in->pts, inlink->time_base, outlink->time_base);
237 ✗ out->duration = 1;
238
239 ✗ projection_matrix(s->fov, half_width / half_height, 0.1f, 1000000.f, s->projection_matrix);
240 ✗ view_matrix(s->eye, s->zoom, s->roll, s->pitch, s->yaw, s->view_matrix);
241 ✗ mmultiply(s->projection_matrix, s->view_matrix, matrix);
242
243 ✗ for (int nb_frame = s->size - 1; nb_frame >= 0; nb_frame--) {
244 ✗ const float scale = 1.f / s->nb_samples;
245 ✗ AVFrame *frame = s->frames[nb_frame];
246 float channels;
247
248 ✗ if (!frame)
249 ✗ continue;
250
251 ✗ channels = frame->ch_layout.nb_channels;
252 ✗ for (int ch = 0; ch < channels; ch++) {
253 ✗ const float *src = (float *)frame->extended_data[ch];
254 ✗ const int r = 128.f + 127.f * sinf(ch / (channels - 1) * M_PI);
255 ✗ const int g = 128.f + 127.f * ch / (channels - 1);
256 ✗ const int b = 128.f + 127.f * cosf(ch / (channels - 1) * M_PI);
257
258 ✗ for (int n = frame->nb_samples - 1, nn = s->nb_samples * nb_frame; n >= 0; n--, nn++) {
259 ✗ float v[4] = { src[n], ch - (channels - 1) * 0.5f, -0.1f + -nn * scale, 1.f };
260 float d[4];
261 int x, y;
262
263 ✗ vmultiply(v, matrix, d);
264
265 ✗ d[0] /= d[3];
266 ✗ d[1] /= d[3];
267
268 ✗ x = d[0] * half_width + half_width;
269 ✗ y = d[1] * half_height + half_height;
270
271 ✗ if (x >= w || y >= h || x < 0 || y < 0)
272 ✗ continue;
273
274 ✗ draw_dot(out, x, y, av_clipf(1.f / d[3], 0.f, 1.f),
275 r, g, b);
276 }
277 }
278 }
279
280 ✗ av_frame_free(&s->frames[59]);
281 ✗ memmove(&s->frames[1], &s->frames[0], 59 * sizeof(AVFrame *));
282 ✗ s->frames[0] = NULL;
283
284 ✗ return ff_filter_frame(outlink, out);
285 }
286
287 ✗ static int activate(AVFilterContext *ctx)
288 {
289 ✗ AVFilterLink *inlink = ctx->inputs[0];
290 ✗ AVFilterLink *outlink = ctx->outputs[0];
291 ✗ Audio3dScopeContext *s = ctx->priv;
292 AVFrame *in;
293 int ret;
294
295 ✗ FF_FILTER_FORWARD_STATUS_BACK(outlink, inlink);
296
297 ✗ ret = ff_inlink_consume_samples(inlink, s->nb_samples, s->nb_samples, &in);
298 ✗ if (ret < 0)
299 ✗ return ret;
300 ✗ if (ret > 0)
301 ✗ return filter_frame(inlink, in);
302
303 ✗ if (ff_inlink_queued_samples(inlink) >= s->nb_samples) {
304 ✗ ff_filter_set_ready(ctx, 10);
305 ✗ return 0;
306 }
307
308 ✗ FF_FILTER_FORWARD_STATUS(inlink, outlink);
309 ✗ FF_FILTER_FORWARD_WANTED(outlink, inlink);
310
311 ✗ return FFERROR_NOT_READY;
312 }
313
314 ✗ static av_cold void uninit(AVFilterContext *ctx)
315 {
316 ✗ Audio3dScopeContext *s = ctx->priv;
317
318 ✗ for (int n = 0; n < 60; n++)
319 ✗ av_frame_free(&s->frames[n]);
320 ✗ }
321
322 static const AVFilterPad audio3dscope_inputs[] = {
323 {
324 .name = "default",
325 .type = AVMEDIA_TYPE_AUDIO,
326 .config_props = config_input,
327 },
328 };
329
330 static const AVFilterPad audio3dscope_outputs[] = {
331 {
332 .name = "default",
333 .type = AVMEDIA_TYPE_VIDEO,
334 .config_props = config_output,
335 },
336 };
337
338 const AVFilter ff_avf_a3dscope = {
339 .name = "a3dscope",
340 .description = NULL_IF_CONFIG_SMALL("Convert input audio to 3d scope video output."),
341 .uninit = uninit,
342 .priv_size = sizeof(Audio3dScopeContext),
343 .activate = activate,
344 FILTER_INPUTS(audio3dscope_inputs),
345 FILTER_OUTPUTS(audio3dscope_outputs),
346 FILTER_QUERY_FUNC2(query_formats),
347 .priv_class = &a3dscope_class,
348 .process_command = ff_filter_process_command,
349 };
350