Edinburgh Speech Tools  2.4-release
 All Classes Functions Variables Typedefs Enumerations Enumerator Friends Pages
EST_Wave.h
1  /*************************************************************************/
2  /* */
3  /* Centre for Speech Technology Research */
4  /* University of Edinburgh, UK */
5  /* Copyright (c) 1996 */
6  /* All Rights Reserved. */
7  /* */
8  /* Permission is hereby granted, free of charge, to use and distribute */
9  /* this software and its documentation without restriction, including */
10  /* without limitation the rights to use, copy, modify, merge, publish, */
11  /* distribute, sublicense, and/or sell copies of this work, and to */
12  /* permit persons to whom this work is furnished to do so, subject to */
13  /* the following conditions: */
14  /* 1. The code must retain the above copyright notice, this list of */
15  /* conditions and the following disclaimer. */
16  /* 2. Any modifications must be clearly marked as such. */
17  /* 3. Original authors' names are not deleted. */
18  /* 4. The authors' names are not used to endorse or promote products */
19  /* derived from this software without specific prior written */
20  /* permission. */
21  /* */
22  /* THE UNIVERSITY OF EDINBURGH AND THE CONTRIBUTORS TO THIS WORK */
23  /* DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING */
24  /* ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT */
25  /* SHALL THE UNIVERSITY OF EDINBURGH NOR THE CONTRIBUTORS BE LIABLE */
26  /* FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES */
27  /* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN */
28  /* AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, */
29  /* ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF */
30  /* THIS SOFTWARE. */
31  /* */
32  /*************************************************************************/
33  /* */
34  /* Author : Paul Taylor and Alan W Black */
35  /* Rewritten : Richard Caley */
36  /* ------------------------------------------------------------------- */
37  /* EST_Wave Class header file */
38  /* */
39  /*************************************************************************/
40 
41 #ifndef __Wave_H__
42 #define __Wave_H__
43 
44 #include <cstdio>
45 #include "EST_Featured.h"
46 #include "EST_rw_status.h"
47 #include "EST_types.h"
48 
49 class EST_Track;
50 class EST_String;
51 class EST_TokenStream;
52 
53 
54 /** A class for storing digital waveforms. The waveform is stored as
55 an array of 16 bit shorts. Multiple channels are supported, but if no
56 channel information is given the 0th channel is accessed.
57 <p>
58 
59 The waveforms can be of any sample rate, and can be changed to another
60 sampling rate using the <tt>resample</tt> function.
61 
62 */
63 
64 class EST_Wave : public EST_Featured
65 {
66 protected:
67  EST_SMatrix p_values;
68 
69  int p_sample_rate;
70 
71  void default_vals(int n=0, int c=1);
72  void free_wave();
73  void copy_data(const EST_Wave &w);
74  void copy_setup(const EST_Wave &w);
75 
76 public:
77 
78  static const int default_sample_rate;
79  static const int default_num_channels;
80 
81  /// default constructor
82  EST_Wave();
83  /// copy constructor
84  EST_Wave(const EST_Wave &a);
85 
86  EST_Wave(int n, int c, int sr);
87 
88  /// Construct from memory supplied by caller
89  EST_Wave(int samps, int chans,
90  short *memory, int offset=0, int sample_rate=default_sample_rate,
91  int free_when_destroyed=0);
92 
93  ~EST_Wave();
94 
95 
96  /**@name Access functions for finding amplitudes of samples */
97  //@{
98 
99  /** return amplitude of sample <tt>i</tt> from channel <tt>
100  channel</tt>. By default the 0th channel is selected. This
101  function can be used for assignment.
102  */
103  short &a(int i, int channel = 0);
104  short a(int i, int channel = 0) const;
105  INLINE short &a_no_check(int i, int channel = 0)
106  { return p_values.a_no_check(i,channel); }
107  INLINE short a_no_check(int i, int channel = 0) const
108  { return p_values.a_no_check(i,channel); }
109  INLINE short &a_no_check_1(int i, int channel = 0)
110  { return p_values.a_no_check_1(i,channel); }
111  INLINE short a_no_check_1(int i, int channel = 0) const
112  { return p_values.a_no_check_1(i,channel); }
113 
114 
115  /** explicit set_a, easier to wrap than assignment
116  */
117  INLINE short set_a(int i, int channel = 0, short val = 0)
118  { return a(i,channel) = val; }
119 
120  /** return amplitude of sample <tt>i</tt> from channel <tt>
121  channel</tt>. By default the 0th channel is selected.
122  */
123  short operator()(int i, int channel) const
124  { return a(i,channel); }
125 
126  /** return amplitude of sample <tt>i</tt> from channel 0.
127  */
128  short operator()(int i) const
129  { return a(i,0); }
130 
131  /** Version of a() that returns zero if index is out of array
132  bounds. This is particularly useful in signal processing when
133  you want to have windows going off the end of the waveform. */
134  short &a_safe(int i, int channel = 0);
135 
136  /// return the time position in seconds of the ith sample
137  float t(int i) const { return (float)i/(float)p_sample_rate; }
138  //@}
139 
140  /**@name Information functions */
141  //@{
142  /// return the number of samples in the waveform
143  int num_samples() const { return p_values.num_rows();}
144  /// return the number of channels in the waveform
145  int num_channels() const { return p_values.num_columns(); }
146  /// return the sampling rate (frequency)
147  int sample_rate() const { return p_sample_rate; }
148  /// Set sampling rate to <tt>n</tt>
149  void set_sample_rate(const int n){p_sample_rate = n;}
150  /// return the size of the waveform, i.e. the number of samples.
151  int length() const { return num_samples();}
152  /// return the time position of the last sample.
153  float end(){ return t(num_samples()-1); }
154 
155  /// Can we look N samples to the left?
156  bool have_left_context(unsigned int n) const
157  { return p_values.have_rows_before(n); }
158 
159  /** returns the file format of the file from which the waveform
160  was read. If the waveform has not been read from a file, this is set
161  to the default type */
162 
163  EST_String sample_type() const { return f_String("sample_type","short"); }
164  void set_sample_type(const EST_String t) { f_set("sample_type", t); }
165 
166  EST_String file_type() const { return f_String("file_type","riff"); }
167  void set_file_type(const EST_String t) { f_set("file_type", t); }
168 
169  /// A string identifying the waveform, commonly used to store the filename
170  EST_String name() const { return f_String("name"); }
171 
172  /// Sets name.
173  void set_name(const EST_String n){ f_set("name", n); }
174 
175  //@}
176 
177  const EST_SMatrix &values() const { return p_values; }
178  EST_SMatrix &values() { return p_values; }
179 
180  /**@name Waveform manipulation functions */
181  //@{
182 
183  /// resize the waveform
184  void resize(int num_samples, int num_channels = EST_ALL, int set=1)
185  { p_values.resize(num_samples, num_channels, set); }
186 
187  /// Resample waveform to <tt>rate</tt>
188  void resample(int rate);
189 
190  /** multiply all samples by a factor <tt>gain</tt>. This checks for
191  overflows and puts them to the maximum positive or negative value
192  as appropriate.
193  */
194  void rescale(float gain,int normalize=0);
195 
196  // multiply samples by a factor contour. The factor_contour track
197  // should contains factor targets at time points throughout the wave,
198  // between which linear interpolation is used to calculate the factor
199  // for each sample.
200  void rescale( const EST_Track &factor_contour );
201 
202  /// clear waveform and set size to 0.
203  void clear() {resize(0,EST_ALL);}
204 
205  void copy(const EST_Wave &from);
206 
207  void fill(short v=0, int channel=EST_ALL);
208 
209  void empty(int channel=EST_ALL) { fill(0,channel); }
210 
211  void sample(EST_TVector<short> &sv, int n)
212  { p_values.row(sv, n); }
213  void channel(EST_TVector<short> &cv, int n)
214  { p_values.column(cv, n); }
215 
216  void copy_channel(int n, short *buf, int offset=0, int num=EST_ALL) const
217  { p_values.copy_column(n, buf, offset, num); }
218  void copy_sample(int n, short *buf, int offset=0, int num=EST_ALL) const
219  { p_values.copy_row(n, buf, offset, num); }
220 
221  void set_channel(int n, const short *buf, int offset=0, int num=EST_ALL)
222  { p_values.set_column(n, buf, offset, num); }
223  void set_sample(int n, const short *buf, int offset=0, int num=EST_ALL)
224  { p_values.set_row(n, buf, offset, num); }
225 
226 
227  void sub_wave(EST_Wave &sw,
228  int offset=0, int num=EST_ALL,
229  int start_c=0, int nchan=EST_ALL);
230 
231  void sub_wave(EST_Wave &sw,
232  int offset=0, int num=EST_ALL,
233  int start_c=0, int nchan=EST_ALL) const
234  { ((EST_Wave *)this)->sub_wave(sw, offset, num, start_c, nchan); }
235 
236  //@}
237 
238  /**@name File i/o functions */
239  //@{
240 
241  /** Load a file into the waveform. The load routine attempts to
242  automatically determine which file type is being loaded. A
243  portion of the waveform can be loaded by setting <tt>
244  offset</tt> to the sample position from the beginning and
245  <length> to the number of required samples after this. */
246 
247  EST_read_status load(const EST_String filename,
248  int offset=0,
249  int length = 0,
250  int rate = default_sample_rate);
251 
252  EST_read_status load(EST_TokenStream &ts,
253  int offset=0,
254  int length = 0,
255  int rate = default_sample_rate);
256 
257  EST_read_status load(const EST_String filename,
258  const EST_String filetype,
259  int offset=0,
260  int length = 0,
261  int rate = default_sample_rate);
262 
263  EST_read_status load(EST_TokenStream &ts,
264  const EST_String filetype,
265  int offset=0,
266  int length = 0,
267  int rate = default_sample_rate);
268 
269  /** Load a file of type <tt>filetype</tt> into the waveform. This
270  can be used to load unheadered files, in which case the fields
271  <tt>sample_rate, sample_type, bo</tt> and <tt>nc</tt> are used
272  to specify the sample rate, type, byte order and number of
273  channels. A portion of the waveform can be loaded by setting
274  <tt> offset</tt> to the sample position from the beginning and
275  <length> to the number of required samples after this.
276  */
277 
278  EST_read_status load_file(const EST_String filename,
279  const EST_String filetype, int sample_rate,
280  const EST_String sample_type, int bo, int nc,
281  int offset = 0, int length = 0);
282  EST_read_status load_file(EST_TokenStream &ts,
283  const EST_String filetype, int sample_rate,
284  const EST_String sample_type, int bo, int nc,
285  int offset = 0, int length = 0);
286 
287  /* Save waveform to a file called <tt>filename</tt> of file
288  format <tt>EST_filetype</tt>.
289  */
290  EST_write_status save(const EST_String filename,
291  const EST_String EST_filetype = "");
292 
293  EST_write_status save(FILE *fp,
294  const EST_String EST_filetype = "");
295 
296  EST_write_status save_file(const EST_String filename,
297  EST_String filetype,
298  EST_String sample_type, int bo);
299 
300  EST_write_status save_file(FILE *fp,
301  EST_String filetype,
302  EST_String sample_type, int bo);
303  //@}
304 
305  /// Assignment operator
306  EST_Wave& operator = (const EST_Wave& w);
307  /** Add to existing wave in serial. Waveforms must have the same
308  number of channels.
309  */
310  EST_Wave& operator +=(const EST_Wave &a);
311  /** Add wave in parallel, i.e. make wave <tt>a</tt> become new
312  channels in existing waveform.
313  */
314  EST_Wave& operator |=(const EST_Wave &a);
315 
316  /// print waveform
317  friend ostream& operator << (ostream& p_values, const EST_Wave &sig);
318 
319  // integrity check *** debug
320  void integrity() const { p_values.integrity() ; }
321 
322 };
323 
325 
326 int operator != (EST_Wave a, EST_Wave b);
327 int operator == (EST_Wave a, EST_Wave b);
328 
329 #endif /* __Wave_H__ */