1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
|
/* For information on usage and redistribution, and for a DISCLAIMER OF ALL
* WARRANTIES, see the file, "LICENSE.txt," in this distribution.
iemlib1 written by Thomas Musil, Copyright (c) IEM KUG Graz Austria 2000 - 2010 */
#include "m_pd.h"
#include "iemlib.h"
/* ---------- sparse_FIR~ - sparse_FIR-filter with coef-matrix-message ----------- */
/* the list or matrix message should have an index and its value, index , value, aso.... */
typedef struct _sparse_FIR_tilde
{
t_object x_obj;
t_sample *x_coef_beg;
int *x_index_beg;
int x_n_coef_resp_order;
int x_n_coef;
int x_n_coef_malloc;
t_sample *x_history_beg;
int x_n_order;
int x_n_order_malloc;
int x_rw_index;
t_float x_float_sig_in;
} t_sparse_FIR_tilde;
static t_class *sparse_FIR_tilde_class;
static t_int *sparse_FIR_tilde_perform(t_int *w)
{
t_sample *in = (t_sample *)(w[1]);
t_sample *out = (t_sample *)(w[2]);
t_sparse_FIR_tilde *x = (t_sparse_FIR_tilde *)(w[3]);
int n = (t_int)(w[4]);
int rw_index = x->x_rw_index;
int i, j, ix;
int order = x->x_n_order;
int n_coef = x->x_n_coef_resp_order;
int n_coef8;
t_sample sum=0.0;
t_sample *coef = x->x_coef_beg;
int *index = x->x_index_beg;
t_sample *write_hist1=x->x_history_beg;
t_sample *write_hist2;
t_sample *read_hist;
t_sample *coef_vec;
int *index_vec;
t_sample *hist_vec;
if((order < 1) || (n_coef < 1))
goto sparse_FIR_tilde_perf_zero;
n_coef8 = n_coef / 8;
write_hist1 = x->x_history_beg;
write_hist2 = write_hist1 + order;
read_hist = write_hist2;
for(i=0; i<n; i++)
{
write_hist1[rw_index] = in[i];
write_hist2[rw_index] = in[i];
sum = 0.0;
coef_vec = coef;
index_vec = index;
hist_vec = &read_hist[rw_index];
for(j=0; j<n_coef8; j++)
{
ix = index_vec[0];
sum += coef_vec[0] * hist_vec[ix];
ix = index_vec[1];
sum += coef_vec[1] * hist_vec[ix];
ix = index_vec[2];
sum += coef_vec[2] * hist_vec[ix];
ix = index_vec[3];
sum += coef_vec[3] * hist_vec[ix];
ix = index_vec[4];
sum += coef_vec[4] * hist_vec[ix];
ix = index_vec[5];
sum += coef_vec[5] * hist_vec[ix];
ix = index_vec[6];
sum += coef_vec[6] * hist_vec[ix];
ix = index_vec[7];
sum += coef_vec[7] * hist_vec[ix];
coef_vec += 8;
index_vec += 8;
}
for(j=n_coef8*8; j<n_coef; j++)
{
ix = index[j];
sum += coef[j] * read_hist[rw_index+ix];
}
out[i] = sum;
rw_index++;
if(rw_index >= order)
rw_index -= order;
}
x->x_rw_index = rw_index;
return(w+5);
sparse_FIR_tilde_perf_zero:
while(n--)
*out++ = 0.0;
return(w+5);
}
static void sparse_FIR_tilde_sort_within(t_sparse_FIR_tilde *x)
{
int cur_order = x->x_n_order;
int n_coef = x->x_n_coef;
int index, i;
int n_coef_resp_order = 0;
int *index_pointer_within = x->x_index_beg;
t_float *coef_pointer_within = x->x_coef_beg;
int *index_pointer = x->x_index_beg + x->x_n_coef_malloc;
t_float *coef_pointer = x->x_coef_beg + x->x_n_coef_malloc;
t_float coef;
for(i=0; i<n_coef; i++)
{
index = index_pointer[i];
coef = coef_pointer[i];
if((index >= 0) && (index < cur_order))
{
index_pointer_within[i] = -index; /* negate index for FIR direction */
coef_pointer_within[i] = coef;
n_coef_resp_order++;
}
}
x->x_n_coef_resp_order = n_coef_resp_order;
}
static void sparse_FIR_tilde_list(t_sparse_FIR_tilde *x, t_symbol *s, int argc, t_atom *argv)
{
int max_order = x->x_n_order_malloc;
int n_pair_arg = argc/2, index, i;
int n_coef = 0;
int *index_pointer;
t_float *coef_pointer;
t_float coef;
if(n_pair_arg > 0)
{
if(n_pair_arg > x->x_n_coef_malloc) /* resize */
{
x->x_index_beg = (int *)resizebytes(x->x_index_beg, 2*x->x_n_coef_malloc*sizeof(int), 2*n_pair_arg*sizeof(int));
x->x_coef_beg = (t_float *)resizebytes(x->x_coef_beg, 2*x->x_n_coef_malloc*sizeof(t_float), 2*n_pair_arg*sizeof(t_float));
x->x_n_coef_malloc = n_pair_arg;
}
index_pointer = x->x_index_beg + x->x_n_coef_malloc;
coef_pointer = x->x_coef_beg + x->x_n_coef_malloc;
for(i=0; i<n_pair_arg; i++)
{
index = (int)atom_getfloat(argv++);
coef = (t_float)atom_getfloat(argv++);
if((index >= 0) && (index < max_order))
{
*index_pointer++ = index;
*coef_pointer++ = coef;
n_coef++;
}
}
x->x_n_coef = n_coef;
sparse_FIR_tilde_sort_within(x);
}
}
static void sparse_FIR_tilde_matrix(t_sparse_FIR_tilde *x, t_symbol *s, int argc, t_atom *argv)
{
int row, col;
if(argc < 2)
{
post("sparse_FIR~ : corrupt matrix passed");
return;
}
row = (int)atom_getfloat(argv++);
col = (int)atom_getfloat(argv++);
if((row < 1)||(col < 1))
{
post("sparse_FIR~ : corrupt matrix passed");
return;
}
if((row*col) < (argc - 2))
{
post("sparse_FIR~ WARNING: row column product less than message content!");
sparse_FIR_tilde_list(x, &s_list, row*col, argv);
}
else if((row*col) > (argc-2))
{
post("sparse_FIR~ WARNING: row column product greater than message content!");
sparse_FIR_tilde_list(x, &s_list, argc-2, argv);
}
else
sparse_FIR_tilde_list(x, &s_list, argc-2, argv);
}
static void sparse_FIR_tilde_order(t_sparse_FIR_tilde *x, t_floatarg fn)
{
int n_order = (int)fn;
if(n_order > 0)
{
if(n_order > x->x_n_order_malloc) /* resize */
{
x->x_history_beg = (t_float *)resizebytes(x->x_history_beg, 2*x->x_n_order_malloc*sizeof(t_float), 2*n_order*sizeof(t_float));
x->x_n_order_malloc = n_order;
}
x->x_n_order = n_order;
x->x_rw_index = 0;
sparse_FIR_tilde_sort_within(x);
}
}
static void sparse_FIR_tilde_dsp(t_sparse_FIR_tilde *x, t_signal **sp)
{
dsp_add(sparse_FIR_tilde_perform, 4, sp[0]->s_vec, sp[1]->s_vec, x, sp[0]->s_n);
}
static void *sparse_FIR_tilde_new(t_floatarg fn)
{
t_sparse_FIR_tilde *x = (t_sparse_FIR_tilde *)pd_new(sparse_FIR_tilde_class);
int n_order=(int)fn;
int i;
outlet_new(&x->x_obj, &s_signal);
x->x_n_coef = 1;
x->x_n_coef_resp_order = 1;
x->x_n_coef_malloc = 1;
x->x_index_beg = (int *)getbytes(2*x->x_n_coef_malloc*sizeof(int));
x->x_coef_beg = (t_float *)getbytes(2*x->x_n_coef_malloc*sizeof(t_float));
x->x_index_beg[0] = 0;
x->x_index_beg[1] = 0;
x->x_coef_beg[0] = 0.0f;
x->x_coef_beg[1] = 0.0f;
if(n_order < 1)
n_order = 1;
x->x_n_order = n_order;
x->x_n_order_malloc = n_order;
x->x_history_beg = (t_float *)getbytes((2*x->x_n_order_malloc)*sizeof(t_float));
x->x_rw_index = 0;
n_order = 2*x->x_n_order_malloc;
for(i=0; i<n_order; i++)
x->x_history_beg[i] = 0.0f;
x->x_float_sig_in = 0.0f;
post("NEW: n_coef_resp_order = %d, n_coef = %d, n_coef_malloc = %d, n_order = %d, n_order_malloc = %d", x->x_n_coef_resp_order, x->x_n_coef, x->x_n_coef_malloc, x->x_n_order, x->x_n_order_malloc);
return(x);
}
static void sparse_FIR_tilde_free(t_sparse_FIR_tilde *x)
{
freebytes(x->x_history_beg, (2*x->x_n_order_malloc)*sizeof(t_float)); /* twice, because of my simple circle-buffer */
freebytes(x->x_index_beg, 2*x->x_n_coef_malloc*sizeof(int)); /* twice, because of buffering both, all coefficients and only the relevant for current order */
freebytes(x->x_coef_beg, 2*x->x_n_coef_malloc*sizeof(t_float)); /* twice, because of buffering both, all coefficients and only the relevant for current order */
}
/*static void sparse_FIR_tilde_dump(t_sparse_FIR_tilde *x)
{
t_float *hist=x->x_history_beg;
int *ix=x->x_index_beg;
int n=x->x_n_order;
post("n_coef_resp_order = %d, n_coef = %d, n_coef_malloc = %d, n_order = %d, n_order_malloc = %d", x->x_n_coef_resp_order, x->x_n_coef, x->x_n_coef_malloc, x->x_n_order, x->x_n_order_malloc);
post("HIST:");
while(n > 8)
{
post("hist = %g, %g, %g, %g, %g, %g, %g, %g", hist[n-1], hist[n-2], hist[n-3], hist[n-4], hist[n-5], hist[n-6], hist[n-7], hist[n-8]);
n -= 8;
hist -= 8;
}
while(n > 0)
{
post("hist = %g", hist[n-1]);
n--;
hist--;
}
post("COEF:");
hist = x->x_coef_beg;
n = x->x_n_coef_resp_order;
while(n > 8)
{
post("coef = %d@%g, %d@%g, %d@%g, %d@%g, %d@%g, %d@%g, %d@%g, %d@%g", ix[n-1],hist[n-1], ix[n-2],hist[n-2], ix[n-3],hist[n-3], ix[n-4],hist[n-4], ix[n-5],hist[n-5], ix[n-6],hist[n-6], ix[n-7],hist[n-7], ix[n-8],hist[n-8]);
n -= 8;
hist -= 8;
}
while(n > 0)
{
post("coef = %d@%g", ix[n-1],hist[n-1]);
n--;
hist--;
}
post("***********************");
}*/
void sparse_FIR_tilde_setup(void)
{
sparse_FIR_tilde_class = class_new(gensym("sparse_FIR~"), (t_newmethod)sparse_FIR_tilde_new,
(t_method)sparse_FIR_tilde_free, sizeof(t_sparse_FIR_tilde), 0, A_DEFFLOAT, 0);
CLASS_MAINSIGNALIN(sparse_FIR_tilde_class, t_sparse_FIR_tilde, x_float_sig_in);
class_addmethod(sparse_FIR_tilde_class, (t_method)sparse_FIR_tilde_dsp, gensym("dsp"), 0);
class_addlist(sparse_FIR_tilde_class, (t_method)sparse_FIR_tilde_list);
class_addmethod(sparse_FIR_tilde_class, (t_method)sparse_FIR_tilde_matrix, gensym("matrix"), A_GIMME, 0);
class_addmethod(sparse_FIR_tilde_class, (t_method)sparse_FIR_tilde_order, gensym("order"), A_FLOAT, 0);
class_addmethod(sparse_FIR_tilde_class, (t_method)sparse_FIR_tilde_order, gensym("size"), A_FLOAT, 0);
//class_addmethod(sparse_FIR_tilde_class, (t_method)sparse_FIR_tilde_dump, gensym("dump"), 0);
}
|