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

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

552

553

554

555

556

557

558

559

560

561

562

563

564

565

566

567

568

569

570

571

572

573

574

575

576

577

578

579

580

581

582

583

584

585

586

587

588

589

590

591

592

593

594

595

596

597

598

599

600

601

602

603

604

605

606

607

608

609

610

611

612

613

614

615

616

617

618

619

620

621

622

623

624

625

626

627

628

629

630

631

632

633

634

635

636

637

638

639

640

641

642

643

644

645

646

647

648

649

650

651

652

653

654

655

656

657

658

659

660

661

662

663

664

665

666

667

668

669

670

671

672

673

674

675

676

677

678

679

680

# Author: Travis Oliphant 

# 2003 

# 

# Feb. 2010: Updated by Warren Weckesser: 

# Rewrote much of chirp() 

# Added sweep_poly() 

from __future__ import division, print_function, absolute_import 

 

import numpy as np 

from numpy import asarray, zeros, place, nan, mod, pi, extract, log, sqrt, \ 

exp, cos, sin, polyval, polyint 

 

from scipy._lib.six import string_types 

 

 

__all__ = ['sawtooth', 'square', 'gausspulse', 'chirp', 'sweep_poly', 

'unit_impulse'] 

 

 

def sawtooth(t, width=1): 

""" 

Return a periodic sawtooth or triangle waveform. 

 

The sawtooth waveform has a period ``2*pi``, rises from -1 to 1 on the 

interval 0 to ``width*2*pi``, then drops from 1 to -1 on the interval 

``width*2*pi`` to ``2*pi``. `width` must be in the interval [0, 1]. 

 

Note that this is not band-limited. It produces an infinite number 

of harmonics, which are aliased back and forth across the frequency 

spectrum. 

 

Parameters 

---------- 

t : array_like 

Time. 

width : array_like, optional 

Width of the rising ramp as a proportion of the total cycle. 

Default is 1, producing a rising ramp, while 0 produces a falling 

ramp. `width` = 0.5 produces a triangle wave. 

If an array, causes wave shape to change over time, and must be the 

same length as t. 

 

Returns 

------- 

y : ndarray 

Output array containing the sawtooth waveform. 

 

Examples 

-------- 

A 5 Hz waveform sampled at 500 Hz for 1 second: 

 

>>> from scipy import signal 

>>> import matplotlib.pyplot as plt 

>>> t = np.linspace(0, 1, 500) 

>>> plt.plot(t, signal.sawtooth(2 * np.pi * 5 * t)) 

 

""" 

t, w = asarray(t), asarray(width) 

w = asarray(w + (t - t)) 

t = asarray(t + (w - w)) 

if t.dtype.char in ['fFdD']: 

ytype = t.dtype.char 

else: 

ytype = 'd' 

y = zeros(t.shape, ytype) 

 

# width must be between 0 and 1 inclusive 

mask1 = (w > 1) | (w < 0) 

place(y, mask1, nan) 

 

# take t modulo 2*pi 

tmod = mod(t, 2 * pi) 

 

# on the interval 0 to width*2*pi function is 

# tmod / (pi*w) - 1 

mask2 = (1 - mask1) & (tmod < w * 2 * pi) 

tsub = extract(mask2, tmod) 

wsub = extract(mask2, w) 

place(y, mask2, tsub / (pi * wsub) - 1) 

 

# on the interval width*2*pi to 2*pi function is 

# (pi*(w+1)-tmod) / (pi*(1-w)) 

 

mask3 = (1 - mask1) & (1 - mask2) 

tsub = extract(mask3, tmod) 

wsub = extract(mask3, w) 

place(y, mask3, (pi * (wsub + 1) - tsub) / (pi * (1 - wsub))) 

return y 

 

 

def square(t, duty=0.5): 

""" 

Return a periodic square-wave waveform. 

 

The square wave has a period ``2*pi``, has value +1 from 0 to 

``2*pi*duty`` and -1 from ``2*pi*duty`` to ``2*pi``. `duty` must be in 

the interval [0,1]. 

 

Note that this is not band-limited. It produces an infinite number 

of harmonics, which are aliased back and forth across the frequency 

spectrum. 

 

Parameters 

---------- 

t : array_like 

The input time array. 

duty : array_like, optional 

Duty cycle. Default is 0.5 (50% duty cycle). 

If an array, causes wave shape to change over time, and must be the 

same length as t. 

 

Returns 

------- 

y : ndarray 

Output array containing the square waveform. 

 

Examples 

-------- 

A 5 Hz waveform sampled at 500 Hz for 1 second: 

 

>>> from scipy import signal 

>>> import matplotlib.pyplot as plt 

>>> t = np.linspace(0, 1, 500, endpoint=False) 

>>> plt.plot(t, signal.square(2 * np.pi * 5 * t)) 

>>> plt.ylim(-2, 2) 

 

A pulse-width modulated sine wave: 

 

>>> plt.figure() 

>>> sig = np.sin(2 * np.pi * t) 

>>> pwm = signal.square(2 * np.pi * 30 * t, duty=(sig + 1)/2) 

>>> plt.subplot(2, 1, 1) 

>>> plt.plot(t, sig) 

>>> plt.subplot(2, 1, 2) 

>>> plt.plot(t, pwm) 

>>> plt.ylim(-1.5, 1.5) 

 

""" 

t, w = asarray(t), asarray(duty) 

w = asarray(w + (t - t)) 

t = asarray(t + (w - w)) 

if t.dtype.char in ['fFdD']: 

ytype = t.dtype.char 

else: 

ytype = 'd' 

 

y = zeros(t.shape, ytype) 

 

# width must be between 0 and 1 inclusive 

mask1 = (w > 1) | (w < 0) 

place(y, mask1, nan) 

 

# on the interval 0 to duty*2*pi function is 1 

tmod = mod(t, 2 * pi) 

mask2 = (1 - mask1) & (tmod < w * 2 * pi) 

place(y, mask2, 1) 

 

# on the interval duty*2*pi to 2*pi function is 

# (pi*(w+1)-tmod) / (pi*(1-w)) 

mask3 = (1 - mask1) & (1 - mask2) 

place(y, mask3, -1) 

return y 

 

 

def gausspulse(t, fc=1000, bw=0.5, bwr=-6, tpr=-60, retquad=False, 

retenv=False): 

""" 

Return a Gaussian modulated sinusoid: 

 

``exp(-a t^2) exp(1j*2*pi*fc*t).`` 

 

If `retquad` is True, then return the real and imaginary parts 

(in-phase and quadrature). 

If `retenv` is True, then return the envelope (unmodulated signal). 

Otherwise, return the real part of the modulated sinusoid. 

 

Parameters 

---------- 

t : ndarray or the string 'cutoff' 

Input array. 

fc : int, optional 

Center frequency (e.g. Hz). Default is 1000. 

bw : float, optional 

Fractional bandwidth in frequency domain of pulse (e.g. Hz). 

Default is 0.5. 

bwr : float, optional 

Reference level at which fractional bandwidth is calculated (dB). 

Default is -6. 

tpr : float, optional 

If `t` is 'cutoff', then the function returns the cutoff 

time for when the pulse amplitude falls below `tpr` (in dB). 

Default is -60. 

retquad : bool, optional 

If True, return the quadrature (imaginary) as well as the real part 

of the signal. Default is False. 

retenv : bool, optional 

If True, return the envelope of the signal. Default is False. 

 

Returns 

------- 

yI : ndarray 

Real part of signal. Always returned. 

yQ : ndarray 

Imaginary part of signal. Only returned if `retquad` is True. 

yenv : ndarray 

Envelope of signal. Only returned if `retenv` is True. 

 

See Also 

-------- 

scipy.signal.morlet 

 

Examples 

-------- 

Plot real component, imaginary component, and envelope for a 5 Hz pulse, 

sampled at 100 Hz for 2 seconds: 

 

>>> from scipy import signal 

>>> import matplotlib.pyplot as plt 

>>> t = np.linspace(-1, 1, 2 * 100, endpoint=False) 

>>> i, q, e = signal.gausspulse(t, fc=5, retquad=True, retenv=True) 

>>> plt.plot(t, i, t, q, t, e, '--') 

 

""" 

if fc < 0: 

raise ValueError("Center frequency (fc=%.2f) must be >=0." % fc) 

if bw <= 0: 

raise ValueError("Fractional bandwidth (bw=%.2f) must be > 0." % bw) 

if bwr >= 0: 

raise ValueError("Reference level for bandwidth (bwr=%.2f) must " 

"be < 0 dB" % bwr) 

 

# exp(-a t^2) <-> sqrt(pi/a) exp(-pi^2/a * f^2) = g(f) 

 

ref = pow(10.0, bwr / 20.0) 

# fdel = fc*bw/2: g(fdel) = ref --- solve this for a 

# 

# pi^2/a * fc^2 * bw^2 /4=-log(ref) 

a = -(pi * fc * bw) ** 2 / (4.0 * log(ref)) 

 

if isinstance(t, string_types): 

if t == 'cutoff': # compute cut_off point 

# Solve exp(-a tc**2) = tref for tc 

# tc = sqrt(-log(tref) / a) where tref = 10^(tpr/20) 

if tpr >= 0: 

raise ValueError("Reference level for time cutoff must be < 0 dB") 

tref = pow(10.0, tpr / 20.0) 

return sqrt(-log(tref) / a) 

else: 

raise ValueError("If `t` is a string, it must be 'cutoff'") 

 

yenv = exp(-a * t * t) 

yI = yenv * cos(2 * pi * fc * t) 

yQ = yenv * sin(2 * pi * fc * t) 

if not retquad and not retenv: 

return yI 

if not retquad and retenv: 

return yI, yenv 

if retquad and not retenv: 

return yI, yQ 

if retquad and retenv: 

return yI, yQ, yenv 

 

 

def chirp(t, f0, t1, f1, method='linear', phi=0, vertex_zero=True): 

"""Frequency-swept cosine generator. 

 

In the following, 'Hz' should be interpreted as 'cycles per unit'; 

there is no requirement here that the unit is one second. The 

important distinction is that the units of rotation are cycles, not 

radians. Likewise, `t` could be a measurement of space instead of time. 

 

Parameters 

---------- 

t : array_like 

Times at which to evaluate the waveform. 

f0 : float 

Frequency (e.g. Hz) at time t=0. 

t1 : float 

Time at which `f1` is specified. 

f1 : float 

Frequency (e.g. Hz) of the waveform at time `t1`. 

method : {'linear', 'quadratic', 'logarithmic', 'hyperbolic'}, optional 

Kind of frequency sweep. If not given, `linear` is assumed. See 

Notes below for more details. 

phi : float, optional 

Phase offset, in degrees. Default is 0. 

vertex_zero : bool, optional 

This parameter is only used when `method` is 'quadratic'. 

It determines whether the vertex of the parabola that is the graph 

of the frequency is at t=0 or t=t1. 

 

Returns 

------- 

y : ndarray 

A numpy array containing the signal evaluated at `t` with the 

requested time-varying frequency. More precisely, the function 

returns ``cos(phase + (pi/180)*phi)`` where `phase` is the integral 

(from 0 to `t`) of ``2*pi*f(t)``. ``f(t)`` is defined below. 

 

See Also 

-------- 

sweep_poly 

 

Notes 

----- 

There are four options for the `method`. The following formulas give 

the instantaneous frequency (in Hz) of the signal generated by 

`chirp()`. For convenience, the shorter names shown below may also be 

used. 

 

linear, lin, li: 

 

``f(t) = f0 + (f1 - f0) * t / t1`` 

 

quadratic, quad, q: 

 

The graph of the frequency f(t) is a parabola through (0, f0) and 

(t1, f1). By default, the vertex of the parabola is at (0, f0). 

If `vertex_zero` is False, then the vertex is at (t1, f1). The 

formula is: 

 

if vertex_zero is True: 

 

``f(t) = f0 + (f1 - f0) * t**2 / t1**2`` 

 

else: 

 

``f(t) = f1 - (f1 - f0) * (t1 - t)**2 / t1**2`` 

 

To use a more general quadratic function, or an arbitrary 

polynomial, use the function `scipy.signal.waveforms.sweep_poly`. 

 

logarithmic, log, lo: 

 

``f(t) = f0 * (f1/f0)**(t/t1)`` 

 

f0 and f1 must be nonzero and have the same sign. 

 

This signal is also known as a geometric or exponential chirp. 

 

hyperbolic, hyp: 

 

``f(t) = f0*f1*t1 / ((f0 - f1)*t + f1*t1)`` 

 

f0 and f1 must be nonzero. 

 

Examples 

-------- 

The following will be used in the examples: 

 

>>> from scipy.signal import chirp, spectrogram 

>>> import matplotlib.pyplot as plt 

 

For the first example, we'll plot the waveform for a linear chirp 

from 6 Hz to 1 Hz over 10 seconds: 

 

>>> t = np.linspace(0, 10, 5001) 

>>> w = chirp(t, f0=6, f1=1, t1=10, method='linear') 

>>> plt.plot(t, w) 

>>> plt.title("Linear Chirp, f(0)=6, f(10)=1") 

>>> plt.xlabel('t (sec)') 

>>> plt.show() 

 

For the remaining examples, we'll use higher frequency ranges, 

and demonstrate the result using `scipy.signal.spectrogram`. 

We'll use a 10 second interval sampled at 8000 Hz. 

 

>>> fs = 8000 

>>> T = 10 

>>> t = np.linspace(0, T, T*fs, endpoint=False) 

 

Quadratic chirp from 1500 Hz to 250 Hz over 10 seconds 

(vertex of the parabolic curve of the frequency is at t=0): 

 

>>> w = chirp(t, f0=1500, f1=250, t1=10, method='quadratic') 

>>> ff, tt, Sxx = spectrogram(w, fs=fs, noverlap=256, nperseg=512, 

... nfft=2048) 

>>> plt.pcolormesh(tt, ff[:513], Sxx[:513], cmap='gray_r') 

>>> plt.title('Quadratic Chirp, f(0)=1500, f(10)=250') 

>>> plt.xlabel('t (sec)') 

>>> plt.ylabel('Frequency (Hz)') 

>>> plt.grid() 

>>> plt.show() 

 

Quadratic chirp from 1500 Hz to 250 Hz over 10 seconds 

(vertex of the parabolic curve of the frequency is at t=10): 

 

>>> w = chirp(t, f0=1500, f1=250, t1=10, method='quadratic', 

... vertex_zero=False) 

>>> ff, tt, Sxx = spectrogram(w, fs=fs, noverlap=256, nperseg=512, 

... nfft=2048) 

>>> plt.pcolormesh(tt, ff[:513], Sxx[:513], cmap='gray_r') 

>>> plt.title('Quadratic Chirp, f(0)=2500, f(10)=250\\n' + 

... '(vertex_zero=False)') 

>>> plt.xlabel('t (sec)') 

>>> plt.ylabel('Frequency (Hz)') 

>>> plt.grid() 

>>> plt.show() 

 

Logarithmic chirp from 1500 Hz to 250 Hz over 10 seconds: 

 

>>> w = chirp(t, f0=1500, f1=250, t1=10, method='logarithmic') 

>>> ff, tt, Sxx = spectrogram(w, fs=fs, noverlap=256, nperseg=512, 

... nfft=2048) 

>>> plt.pcolormesh(tt, ff[:513], Sxx[:513], cmap='gray_r') 

>>> plt.title('Logarithmic Chirp, f(0)=1500, f(10)=250') 

>>> plt.xlabel('t (sec)') 

>>> plt.ylabel('Frequency (Hz)') 

>>> plt.grid() 

>>> plt.show() 

 

Hyperbolic chirp from 1500 Hz to 250 Hz over 10 seconds: 

 

>>> w = chirp(t, f0=1500, f1=250, t1=10, method='hyperbolic') 

>>> ff, tt, Sxx = spectrogram(w, fs=fs, noverlap=256, nperseg=512, 

... nfft=2048) 

>>> plt.pcolormesh(tt, ff[:513], Sxx[:513], cmap='gray_r') 

>>> plt.title('Hyperbolic Chirp, f(0)=1500, f(10)=250') 

>>> plt.xlabel('t (sec)') 

>>> plt.ylabel('Frequency (Hz)') 

>>> plt.grid() 

>>> plt.show() 

 

""" 

# 'phase' is computed in _chirp_phase, to make testing easier. 

phase = _chirp_phase(t, f0, t1, f1, method, vertex_zero) 

# Convert phi to radians. 

phi *= pi / 180 

return cos(phase + phi) 

 

 

def _chirp_phase(t, f0, t1, f1, method='linear', vertex_zero=True): 

""" 

Calculate the phase used by chirp_phase to generate its output. 

 

See `chirp` for a description of the arguments. 

 

""" 

t = asarray(t) 

f0 = float(f0) 

t1 = float(t1) 

f1 = float(f1) 

if method in ['linear', 'lin', 'li']: 

beta = (f1 - f0) / t1 

phase = 2 * pi * (f0 * t + 0.5 * beta * t * t) 

 

elif method in ['quadratic', 'quad', 'q']: 

beta = (f1 - f0) / (t1 ** 2) 

if vertex_zero: 

phase = 2 * pi * (f0 * t + beta * t ** 3 / 3) 

else: 

phase = 2 * pi * (f1 * t + beta * ((t1 - t) ** 3 - t1 ** 3) / 3) 

 

elif method in ['logarithmic', 'log', 'lo']: 

if f0 * f1 <= 0.0: 

raise ValueError("For a logarithmic chirp, f0 and f1 must be " 

"nonzero and have the same sign.") 

if f0 == f1: 

phase = 2 * pi * f0 * t 

else: 

beta = t1 / log(f1 / f0) 

phase = 2 * pi * beta * f0 * (pow(f1 / f0, t / t1) - 1.0) 

 

elif method in ['hyperbolic', 'hyp']: 

if f0 == 0 or f1 == 0: 

raise ValueError("For a hyperbolic chirp, f0 and f1 must be " 

"nonzero.") 

if f0 == f1: 

# Degenerate case: constant frequency. 

phase = 2 * pi * f0 * t 

else: 

# Singular point: the instantaneous frequency blows up 

# when t == sing. 

sing = -f1 * t1 / (f0 - f1) 

phase = 2 * pi * (-sing * f0) * log(np.abs(1 - t/sing)) 

 

else: 

raise ValueError("method must be 'linear', 'quadratic', 'logarithmic'," 

" or 'hyperbolic', but a value of %r was given." 

% method) 

 

return phase 

 

 

def sweep_poly(t, poly, phi=0): 

""" 

Frequency-swept cosine generator, with a time-dependent frequency. 

 

This function generates a sinusoidal function whose instantaneous 

frequency varies with time. The frequency at time `t` is given by 

the polynomial `poly`. 

 

Parameters 

---------- 

t : ndarray 

Times at which to evaluate the waveform. 

poly : 1-D array_like or instance of numpy.poly1d 

The desired frequency expressed as a polynomial. If `poly` is 

a list or ndarray of length n, then the elements of `poly` are 

the coefficients of the polynomial, and the instantaneous 

frequency is 

 

``f(t) = poly[0]*t**(n-1) + poly[1]*t**(n-2) + ... + poly[n-1]`` 

 

If `poly` is an instance of numpy.poly1d, then the 

instantaneous frequency is 

 

``f(t) = poly(t)`` 

 

phi : float, optional 

Phase offset, in degrees, Default: 0. 

 

Returns 

------- 

sweep_poly : ndarray 

A numpy array containing the signal evaluated at `t` with the 

requested time-varying frequency. More precisely, the function 

returns ``cos(phase + (pi/180)*phi)``, where `phase` is the integral 

(from 0 to t) of ``2 * pi * f(t)``; ``f(t)`` is defined above. 

 

See Also 

-------- 

chirp 

 

Notes 

----- 

.. versionadded:: 0.8.0 

 

If `poly` is a list or ndarray of length `n`, then the elements of 

`poly` are the coefficients of the polynomial, and the instantaneous 

frequency is: 

 

``f(t) = poly[0]*t**(n-1) + poly[1]*t**(n-2) + ... + poly[n-1]`` 

 

If `poly` is an instance of `numpy.poly1d`, then the instantaneous 

frequency is: 

 

``f(t) = poly(t)`` 

 

Finally, the output `s` is: 

 

``cos(phase + (pi/180)*phi)`` 

 

where `phase` is the integral from 0 to `t` of ``2 * pi * f(t)``, 

``f(t)`` as defined above. 

 

Examples 

-------- 

Compute the waveform with instantaneous frequency:: 

 

f(t) = 0.025*t**3 - 0.36*t**2 + 1.25*t + 2 

 

over the interval 0 <= t <= 10. 

 

>>> from scipy.signal import sweep_poly 

>>> p = np.poly1d([0.025, -0.36, 1.25, 2.0]) 

>>> t = np.linspace(0, 10, 5001) 

>>> w = sweep_poly(t, p) 

 

Plot it: 

 

>>> import matplotlib.pyplot as plt 

>>> plt.subplot(2, 1, 1) 

>>> plt.plot(t, w) 

>>> plt.title("Sweep Poly\\nwith frequency " + 

... "$f(t) = 0.025t^3 - 0.36t^2 + 1.25t + 2$") 

>>> plt.subplot(2, 1, 2) 

>>> plt.plot(t, p(t), 'r', label='f(t)') 

>>> plt.legend() 

>>> plt.xlabel('t') 

>>> plt.tight_layout() 

>>> plt.show() 

 

""" 

# 'phase' is computed in _sweep_poly_phase, to make testing easier. 

phase = _sweep_poly_phase(t, poly) 

# Convert to radians. 

phi *= pi / 180 

return cos(phase + phi) 

 

 

def _sweep_poly_phase(t, poly): 

""" 

Calculate the phase used by sweep_poly to generate its output. 

 

See `sweep_poly` for a description of the arguments. 

 

""" 

# polyint handles lists, ndarrays and instances of poly1d automatically. 

intpoly = polyint(poly) 

phase = 2 * pi * polyval(intpoly, t) 

return phase 

 

 

def unit_impulse(shape, idx=None, dtype=float): 

""" 

Unit impulse signal (discrete delta function) or unit basis vector. 

 

Parameters 

---------- 

shape : int or tuple of int 

Number of samples in the output (1-D), or a tuple that represents the 

shape of the output (N-D). 

idx : None or int or tuple of int or 'mid', optional 

Index at which the value is 1. If None, defaults to the 0th element. 

If ``idx='mid'``, the impulse will be centered at ``shape // 2`` in 

all dimensions. If an int, the impulse will be at `idx` in all 

dimensions. 

dtype : data-type, optional 

The desired data-type for the array, e.g., `numpy.int8`. Default is 

`numpy.float64`. 

 

Returns 

------- 

y : ndarray 

Output array containing an impulse signal. 

 

Notes 

----- 

The 1D case is also known as the Kronecker delta. 

 

.. versionadded:: 0.19.0 

 

Examples 

-------- 

An impulse at the 0th element (:math:`\\delta[n]`): 

 

>>> from scipy import signal 

>>> signal.unit_impulse(8) 

array([ 1., 0., 0., 0., 0., 0., 0., 0.]) 

 

Impulse offset by 2 samples (:math:`\\delta[n-2]`): 

 

>>> signal.unit_impulse(7, 2) 

array([ 0., 0., 1., 0., 0., 0., 0.]) 

 

2-dimensional impulse, centered: 

 

>>> signal.unit_impulse((3, 3), 'mid') 

array([[ 0., 0., 0.], 

[ 0., 1., 0.], 

[ 0., 0., 0.]]) 

 

Impulse at (2, 2), using broadcasting: 

 

>>> signal.unit_impulse((4, 4), 2) 

array([[ 0., 0., 0., 0.], 

[ 0., 0., 0., 0.], 

[ 0., 0., 1., 0.], 

[ 0., 0., 0., 0.]]) 

 

Plot the impulse response of a 4th-order Butterworth lowpass filter: 

 

>>> imp = signal.unit_impulse(100, 'mid') 

>>> b, a = signal.butter(4, 0.2) 

>>> response = signal.lfilter(b, a, imp) 

 

>>> import matplotlib.pyplot as plt 

>>> plt.plot(np.arange(-50, 50), imp) 

>>> plt.plot(np.arange(-50, 50), response) 

>>> plt.margins(0.1, 0.1) 

>>> plt.xlabel('Time [samples]') 

>>> plt.ylabel('Amplitude') 

>>> plt.grid(True) 

>>> plt.show() 

 

""" 

out = zeros(shape, dtype) 

 

shape = np.atleast_1d(shape) 

 

if idx is None: 

idx = (0,) * len(shape) 

elif idx == 'mid': 

idx = tuple(shape // 2) 

elif not hasattr(idx, "__iter__"): 

idx = (idx,) * len(shape) 

 

out[idx] = 1 

return out