1 |
|
|
/* |
2 |
|
|
* Copyright (c) 2011 Mina Nagy Zaki |
3 |
|
|
* Copyright (c) 2000 Edward Beingessner And Sundry Contributors. |
4 |
|
|
* This source code is freely redistributable and may be used for any purpose. |
5 |
|
|
* This copyright notice must be maintained. Edward Beingessner And Sundry |
6 |
|
|
* Contributors are not responsible for the consequences of using this |
7 |
|
|
* software. |
8 |
|
|
* |
9 |
|
|
* This file is part of FFmpeg. |
10 |
|
|
* |
11 |
|
|
* FFmpeg is free software; you can redistribute it and/or |
12 |
|
|
* modify it under the terms of the GNU Lesser General Public |
13 |
|
|
* License as published by the Free Software Foundation; either |
14 |
|
|
* version 2.1 of the License, or (at your option) any later version. |
15 |
|
|
* |
16 |
|
|
* FFmpeg is distributed in the hope that it will be useful, |
17 |
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of |
18 |
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
19 |
|
|
* Lesser General Public License for more details. |
20 |
|
|
* |
21 |
|
|
* You should have received a copy of the GNU Lesser General Public |
22 |
|
|
* License along with FFmpeg; if not, write to the Free Software |
23 |
|
|
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
24 |
|
|
*/ |
25 |
|
|
|
26 |
|
|
/** |
27 |
|
|
* @file |
28 |
|
|
* Stereo Widening Effect. Adds audio cues to move stereo image in |
29 |
|
|
* front of the listener. Adapted from the libsox earwax effect. |
30 |
|
|
*/ |
31 |
|
|
|
32 |
|
|
#include "libavutil/channel_layout.h" |
33 |
|
|
#include "avfilter.h" |
34 |
|
|
#include "audio.h" |
35 |
|
|
#include "formats.h" |
36 |
|
|
|
37 |
|
|
#define NUMTAPS 32 |
38 |
|
|
|
39 |
|
|
static const int8_t filt[NUMTAPS * 2] = { |
40 |
|
|
/* 30° 330° */ |
41 |
|
|
4, -6, /* 32 tap stereo FIR filter. */ |
42 |
|
|
4, -11, /* One side filters as if the */ |
43 |
|
|
-1, -5, /* signal was from 30 degrees */ |
44 |
|
|
3, 3, /* from the ear, the other as */ |
45 |
|
|
-2, 5, /* if 330 degrees. */ |
46 |
|
|
-5, 0, |
47 |
|
|
9, 1, |
48 |
|
|
6, 3, /* Input */ |
49 |
|
|
-4, -1, /* Left Right */ |
50 |
|
|
-5, -3, /* __________ __________ */ |
51 |
|
|
-2, -5, /* | | | | */ |
52 |
|
|
-7, 1, /* .---| Hh,0(f) | | Hh,0(f) |---. */ |
53 |
|
|
6, -7, /* / |__________| |__________| \ */ |
54 |
|
|
30, -29, /* / \ / \ */ |
55 |
|
|
12, -3, /* / X \ */ |
56 |
|
|
-11, 4, /* / / \ \ */ |
57 |
|
|
-3, 7, /* ____V_____ __________V V__________ _____V____ */ |
58 |
|
|
-20, 23, /* | | | | | | | | */ |
59 |
|
|
2, 0, /* | Hh,30(f) | | Hh,330(f)| | Hh,330(f)| | Hh,30(f) | */ |
60 |
|
|
1, -6, /* |__________| |__________| |__________| |__________| */ |
61 |
|
|
-14, -5, /* \ ___ / \ ___ / */ |
62 |
|
|
15, -18, /* \ / \ / _____ \ / \ / */ |
63 |
|
|
6, 7, /* `->| + |<--' / \ `-->| + |<-' */ |
64 |
|
|
15, -10, /* \___/ _/ \_ \___/ */ |
65 |
|
|
-14, 22, /* \ / \ / \ / */ |
66 |
|
|
-7, -2, /* `--->| | | |<---' */ |
67 |
|
|
-4, 9, /* \_/ \_/ */ |
68 |
|
|
6, -12, /* */ |
69 |
|
|
6, -6, /* Headphones */ |
70 |
|
|
0, -11, |
71 |
|
|
0, -5, |
72 |
|
|
4, 0}; |
73 |
|
|
|
74 |
|
|
typedef struct EarwaxContext { |
75 |
|
|
int16_t filter[2][NUMTAPS]; |
76 |
|
|
int16_t taps[4][NUMTAPS * 2]; |
77 |
|
|
|
78 |
|
|
AVFrame *frame[2]; |
79 |
|
|
} EarwaxContext; |
80 |
|
|
|
81 |
|
1 |
static int query_formats(AVFilterContext *ctx) |
82 |
|
|
{ |
83 |
|
|
static const int sample_rates[] = { 44100, -1 }; |
84 |
|
|
int ret; |
85 |
|
|
|
86 |
|
1 |
AVFilterFormats *formats = NULL; |
87 |
|
1 |
AVFilterChannelLayouts *layout = NULL; |
88 |
|
|
|
89 |
✓✗✓✗
|
2 |
if ((ret = ff_add_format (&formats, AV_SAMPLE_FMT_S16P )) < 0 || |
90 |
✓✗ |
2 |
(ret = ff_set_common_formats (ctx , formats )) < 0 || |
91 |
✓✗ |
2 |
(ret = ff_add_channel_layout (&layout , AV_CH_LAYOUT_STEREO )) < 0 || |
92 |
✗✓ |
2 |
(ret = ff_set_common_channel_layouts (ctx , layout )) < 0 || |
93 |
|
1 |
(ret = ff_set_common_samplerates (ctx , ff_make_format_list(sample_rates) )) < 0) |
94 |
|
|
return ret; |
95 |
|
|
|
96 |
|
1 |
return 0; |
97 |
|
|
} |
98 |
|
|
|
99 |
|
|
//FIXME: replace with DSPContext.scalarproduct_int16 |
100 |
|
160 |
static inline int16_t *scalarproduct(const int16_t *in, const int16_t *endin, |
101 |
|
|
const int16_t *filt, int16_t *out) |
102 |
|
|
{ |
103 |
|
|
int32_t sample; |
104 |
|
|
int16_t j; |
105 |
|
|
|
106 |
✓✓ |
82080 |
while (in < endin) { |
107 |
|
81920 |
sample = 0; |
108 |
✓✓ |
2703360 |
for (j = 0; j < NUMTAPS; j++) |
109 |
|
2621440 |
sample += in[j] * filt[j]; |
110 |
|
81920 |
*out = av_clip_int16(sample >> 7); |
111 |
|
81920 |
out++; |
112 |
|
81920 |
in++; |
113 |
|
|
} |
114 |
|
|
|
115 |
|
160 |
return out; |
116 |
|
|
} |
117 |
|
|
|
118 |
|
1 |
static int config_input(AVFilterLink *inlink) |
119 |
|
|
{ |
120 |
|
1 |
EarwaxContext *s = inlink->dst->priv; |
121 |
|
|
|
122 |
✓✓ |
33 |
for (int i = 0; i < NUMTAPS; i++) { |
123 |
|
32 |
s->filter[0][i] = filt[i * 2]; |
124 |
|
32 |
s->filter[1][i] = filt[i * 2 + 1]; |
125 |
|
|
} |
126 |
|
|
|
127 |
|
1 |
return 0; |
128 |
|
|
} |
129 |
|
|
|
130 |
|
80 |
static void convolve(AVFilterContext *ctx, AVFrame *in, |
131 |
|
|
int input_ch, int output_ch, |
132 |
|
|
int filter_ch, int tap_ch) |
133 |
|
|
{ |
134 |
|
80 |
EarwaxContext *s = ctx->priv; |
135 |
|
|
int16_t *taps, *endin, *dst, *src; |
136 |
|
|
int len; |
137 |
|
|
|
138 |
|
80 |
taps = s->taps[tap_ch]; |
139 |
|
80 |
dst = (int16_t *)s->frame[input_ch]->data[output_ch]; |
140 |
|
80 |
src = (int16_t *)in->data[input_ch]; |
141 |
|
|
|
142 |
|
80 |
len = FFMIN(NUMTAPS, in->nb_samples); |
143 |
|
|
// copy part of new input and process with saved input |
144 |
|
80 |
memcpy(taps+NUMTAPS, src, len * sizeof(*taps)); |
145 |
|
80 |
dst = scalarproduct(taps, taps + len, s->filter[filter_ch], dst); |
146 |
|
|
|
147 |
|
|
// process current input |
148 |
✓✗ |
80 |
if (in->nb_samples >= NUMTAPS) { |
149 |
|
80 |
endin = src + in->nb_samples - NUMTAPS; |
150 |
|
80 |
scalarproduct(src, endin, s->filter[filter_ch], dst); |
151 |
|
|
|
152 |
|
|
// save part of input for next round |
153 |
|
80 |
memcpy(taps, endin, NUMTAPS * sizeof(*taps)); |
154 |
|
|
} else { |
155 |
|
|
memmove(taps, taps + in->nb_samples, NUMTAPS * sizeof(*taps)); |
156 |
|
|
} |
157 |
|
80 |
} |
158 |
|
|
|
159 |
|
40 |
static void mix(AVFilterContext *ctx, AVFrame *out, |
160 |
|
|
int output_ch, int f0, int f1, int i0, int i1) |
161 |
|
|
{ |
162 |
|
40 |
EarwaxContext *s = ctx->priv; |
163 |
|
40 |
const int16_t *srcl = (const int16_t *)s->frame[f0]->data[i0]; |
164 |
|
40 |
const int16_t *srcr = (const int16_t *)s->frame[f1]->data[i1]; |
165 |
|
40 |
int16_t *dst = (int16_t *)out->data[output_ch]; |
166 |
|
|
|
167 |
✓✓ |
41000 |
for (int n = 0; n < out->nb_samples; n++) |
168 |
|
40960 |
dst[n] = av_clip_int16(srcl[n] + srcr[n]); |
169 |
|
40 |
} |
170 |
|
|
|
171 |
|
20 |
static int filter_frame(AVFilterLink *inlink, AVFrame *in) |
172 |
|
|
{ |
173 |
|
20 |
AVFilterContext *ctx = inlink->dst; |
174 |
|
20 |
EarwaxContext *s = ctx->priv; |
175 |
|
20 |
AVFilterLink *outlink = ctx->outputs[0]; |
176 |
|
20 |
AVFrame *out = ff_get_audio_buffer(outlink, in->nb_samples); |
177 |
|
|
|
178 |
✓✓ |
60 |
for (int ch = 0; ch < 2; ch++) { |
179 |
✓✓✗✓
|
40 |
if (!s->frame[ch] || s->frame[ch]->nb_samples < in->nb_samples) { |
180 |
|
2 |
av_frame_free(&s->frame[ch]); |
181 |
|
2 |
s->frame[ch] = ff_get_audio_buffer(outlink, in->nb_samples); |
182 |
✗✓ |
2 |
if (!s->frame[ch]) { |
183 |
|
|
av_frame_free(&in); |
184 |
|
|
av_frame_free(&out); |
185 |
|
|
return AVERROR(ENOMEM); |
186 |
|
|
} |
187 |
|
|
} |
188 |
|
|
} |
189 |
|
|
|
190 |
✗✓ |
20 |
if (!out) { |
191 |
|
|
av_frame_free(&in); |
192 |
|
|
return AVERROR(ENOMEM); |
193 |
|
|
} |
194 |
|
20 |
av_frame_copy_props(out, in); |
195 |
|
|
|
196 |
|
20 |
convolve(ctx, in, 0, 0, 0, 0); |
197 |
|
20 |
convolve(ctx, in, 0, 1, 1, 1); |
198 |
|
20 |
convolve(ctx, in, 1, 0, 0, 2); |
199 |
|
20 |
convolve(ctx, in, 1, 1, 1, 3); |
200 |
|
|
|
201 |
|
20 |
mix(ctx, out, 0, 0, 1, 1, 0); |
202 |
|
20 |
mix(ctx, out, 1, 0, 1, 0, 1); |
203 |
|
|
|
204 |
|
20 |
av_frame_free(&in); |
205 |
|
20 |
return ff_filter_frame(outlink, out); |
206 |
|
|
} |
207 |
|
|
|
208 |
|
1 |
static av_cold void uninit(AVFilterContext *ctx) |
209 |
|
|
{ |
210 |
|
1 |
EarwaxContext *s = ctx->priv; |
211 |
|
|
|
212 |
|
1 |
av_frame_free(&s->frame[0]); |
213 |
|
1 |
av_frame_free(&s->frame[1]); |
214 |
|
1 |
} |
215 |
|
|
|
216 |
|
|
static const AVFilterPad earwax_inputs[] = { |
217 |
|
|
{ |
218 |
|
|
.name = "default", |
219 |
|
|
.type = AVMEDIA_TYPE_AUDIO, |
220 |
|
|
.filter_frame = filter_frame, |
221 |
|
|
.config_props = config_input, |
222 |
|
|
}, |
223 |
|
|
{ NULL } |
224 |
|
|
}; |
225 |
|
|
|
226 |
|
|
static const AVFilterPad earwax_outputs[] = { |
227 |
|
|
{ |
228 |
|
|
.name = "default", |
229 |
|
|
.type = AVMEDIA_TYPE_AUDIO, |
230 |
|
|
}, |
231 |
|
|
{ NULL } |
232 |
|
|
}; |
233 |
|
|
|
234 |
|
|
AVFilter ff_af_earwax = { |
235 |
|
|
.name = "earwax", |
236 |
|
|
.description = NULL_IF_CONFIG_SMALL("Widen the stereo image."), |
237 |
|
|
.query_formats = query_formats, |
238 |
|
|
.priv_size = sizeof(EarwaxContext), |
239 |
|
|
.uninit = uninit, |
240 |
|
|
.inputs = earwax_inputs, |
241 |
|
|
.outputs = earwax_outputs, |
242 |
|
|
}; |