Generate sweptfrequency cosine (chirp) signal
Sources
dspsrcs4
The Chirp block outputs a sweptfrequency cosine (chirp) signal with unity amplitude and continuous phase. To specify the desired output chirp signal, you must define its instantaneous frequency function, also known as the output frequency sweep. The frequency sweep can be linear, quadratic, or logarithmic, and repeats once every Sweep time by default. See other sections of this reference page for more details about the block.
Variables Used in This Reference Page
f_{0}  Initial frequency parameter (Hz) 
f_{i}(t_{g})  Target frequency parameter (Hz) 
t_{g}  Target time parameter (seconds) 
T_{sw}  Sweep time parameter (seconds) 
$${\varphi}_{0}$$  Initial phase parameter (radians) 
$$\psi (t)$$  Phase of the chirp signal (radians) 
f_{i}(t)  Userspecified output instantaneous frequency function (Hz); userspecified sweep 
f_{i(actual)}(t)  Actual output instantaneous frequency function (Hz); actual output sweep 
y_{chirp}(t)  Output chirp function 
The basic shape of the output instantaneous frequency sweep, f_{i}(t), is set by the Frequency sweep and Sweep mode parameters, described in the following table.
Parameters for Setting Sweep Shape  Possible Setting  Parameter Description 

Frequency sweep  Linear Quadratic Logarithmic Swept cosine  Determines whether the sweep frequencies vary linearly, quadratically, or logarithmically. Linear and swept cosine sweeps both vary linearly. 
Sweep mode  Unidirectional Bidirectional  Determines whether the sweep is unidirectional or bidirectional. For details, see Unidirectional and Bidirectional Sweep Modes 
The following diagram illustrates the possible shapes of the frequency sweep that you can obtain by setting the Frequency sweep and Sweep mode parameters.
For information on how to set the frequency values in your sweep, see Setting Instantaneous Frequency Sweep Values.
The Sweep mode parameter determines whether your sweep is unidirectional or bidirectional, which affects the shape of your output frequency sweep (see Shaping the Frequency Sweep by Setting Frequency Sweep and Sweep Mode). The following table describes the characteristics of unidirectional and bidirectional sweeps.
Sweep Mode Parameter Settings  Sweep Characteristics 





The following diagram illustrates a linear sweep in both sweep modes. For information on setting the frequency values in your sweep, see Setting Instantaneous Frequency Sweep Values.
Set the following parameters to tune the frequency values of your output frequency sweep.
Initial frequency (Hz), f_{0}
Target frequency (Hz), f_{i}(t_{g})
Target time (seconds), t_{g}
The following table summarizes the sweep values at specific times for all Frequency sweep settings. For information on the formulas used to compute sweep values at other times, see Block Computation Methods.
Instantaneous Frequency Sweep Values
Frequency Sweep  Sweep Value at t = 0  Sweep Value at t = t_{g}  Time when Sweep Value Is Target Frequency, f_{i}(t_{g}) 

Linear  f_{0}  f_{i}(t_{g})  t_{g} 
Quadratic  f_{0}  f_{i}(t_{g})  t_{g} 
Logarithmic  f_{0}  f_{i}(t_{g})  t_{g} 
Swept cosine  f_{0}  2f_{i}(t_{g})  f_{0}  t_{g}/2 
The Chirp block uses one of two formulas to compute the block output, depending on the Frequency Sweep parameter setting. For details, see the following sections:
The following table shows the equations used by the block to
compute the userspecified output frequency sweep, f_{i}(t),
the block output, y_{chirp}(t),
and the actual output frequency sweep, f_{i(actual)}(t).
The only time the userspecified sweep is not the actual output sweep
is when the Frequency sweep parameter is set
to Swept cosine
.
Note The following equations apply only to unidirectional sweeps in which f_{i}(0) < f_{i}(t_{g}). To derive equations for other cases, you might find it helpful to examine the following table and the diagram in Shaping the Frequency Sweep by Setting Frequency Sweep and Sweep Mode. 
The table below contains the following variables:
f_{i}(t) — the userspecified frequency sweep
f_{i(actual)}(t) — the actual output frequency sweep, usually equal to f_{i}(t)
y(t) — the Chirp block output
$$\psi (t)$$ — the phase of the chirp signal, where $$\psi (0)=0$$, and $$2\pi {f}_{i}(t)$$ is the derivative of the phase
$${f}_{i}(t)=\frac{1}{2\pi}\cdot \frac{d\psi (t)}{dt}$$
$${\varphi}_{0}$$ — the Initial phase parameter value, where $${y}_{chirp}(0)=\mathrm{cos}({\varphi}_{0})$$
Equations Used by the Chirp Block for Unidirectional Positive Sweeps
Frequency Sweep  Block Output Chirp Signal  UserSpecified Frequency Sweep, f_{i}(t)  $$\beta $$  Actual Frequency Sweep, f_{i(actual)}(t) 

 $$y(t)=\mathrm{cos}(\psi (t)+{\varphi}_{0})$$  $${f}_{i}(t)={f}_{0}+\beta t$$  $$\beta =\frac{{f}_{i}({t}_{g}){f}_{0}}{{t}_{g}}$$  $${f}_{i}{{}_{(actual)}}^{(t)}={f}_{i}{}^{(t)}$$ 
 Same as  $${f}_{i}(t)={f}_{0}+\beta {t}^{2}$$  $$\beta =\frac{{f}_{i}({t}_{g}){f}_{0}}{{t}_{g}^{2}}$$  $${f}_{i}{{}_{(actual)}}^{(t)}={f}_{i}{}^{(t)}$$ 
 Same as  $${F}_{i}(t)={f}_{0}{\left(\frac{{f}_{i}\left({t}_{g}\right)}{{f}_{0}}\right)}^{\frac{t}{{t}_{g}}}$$ Where f_{i}(t_{g}) > f_{0}> 0  N/A  $${f}_{i(actual)}(t)={f}_{i}(t)$$ 
 $$y(t)=\mathrm{cos}(2\pi {f}_{i}(t)t+{\varphi}_{0})$$  Same as  Same as  $${f}_{i(actual)}(t)={f}_{i}(t)+\beta t$$ 
The derivative of the phase of a chirp function gives the instantaneous
frequency of the chirp function. The Chirp block uses this principle
to calculate the chirp output when the Frequency Sweep parameter
is set to Linear
, Quadratic
,
or Logarithmic
.
$${y}_{chirp}(t)=\mathrm{cos}(\psi (t)+{\varphi}_{0})$$  Linear, quadratic, or logarithmic chirp signal with phase $$\psi (t)$$ 
$${f}_{i}(t)=\frac{1}{2\pi}\cdot \frac{d\psi (t)}{dt}$$  Phase derivative is instantaneous frequency 
For instance, if you want a chirp signal with a linear instantaneous
frequency sweep, you should set the Frequency Sweep parameter
to Linear
, and tune the linear sweep values
by setting other parameters appropriately. The block outputs a chirp
signal, the phase derivative of which is the specified linear sweep.
This ensures that the instantaneous frequency of the output is the
linear sweep you desired. For equations describing the linear, quadratic,
and logarithmic sweeps, see Equations for Output Computation.
To generate the swept cosine chirp signal, the block sets the swept cosine chirp output as follows.
$${y}_{chirp}(t)=\mathrm{cos}(\psi (t)+{\varphi}_{0})=\mathrm{cos}(2\pi {f}_{i}(t)t+{\varphi}_{0})$$  Swept cosine chirp output (Instantaneous frequency equation, shown above, does not hold.) 
Note that the instantaneous frequency equation, shown above, does not hold for the swept cosine chirp, so the userdefined frequency sweep, f_{i}(t), is not the actual output frequency sweep, f_{i(actual)}(t), of the swept cosine chirp. Thus, the swept cosine output might not behave as you expect. To learn more about swept cosine chirp behavior, see Cautions Regarding the Swept Cosine Sweep and Equations for Output Computation.
When you want a linearly swept chirp signal, we recommend you use a linear frequency sweep. Though a swept cosine frequency sweep also yields a linearly swept chirp signal, the output might have unexpected frequency content. For details, see the following two sections.
The swept cosine sweep value at the Target time is not necessarily the Target frequency. This is because the userspecified sweep is not the actual frequency sweep of the swept cosine output, as noted in Output Computation Method for Swept Cosine Frequency Sweep. See the table Instantaneous Frequency Sweep Values for the actual value of the swept cosine sweep at the Target time.
In Swept cosine mode, you should not set the parameters so that 1/Tsw is very large compared to the values of the Initial frequency and Target frequency parameters. In such cases, the actual frequency content of the swept cosine sweep might be closer to 1/Tsw, far exceeding the Initial frequency and Target frequency parameter values.
The type of output instantaneous frequency sweep, f_{i}(t): Linear
, Logarithmic
, Quadratic
,
or Swept cosine
.
The directionality of the chirp signal: Unidirectional
or Bidirectional
.
For Linear
, Quadratic
,
and Swept cosine
sweeps, the initial frequency, f_{0},
of the output chirp signal. For Logarithmic
sweeps, Initial
frequency is one less than the actual initial frequency
of the sweep. Also, when the sweep is Logarithmic
,
you must set the Initial frequency to be less
than the Target frequency. Tunable.
For Linear
, Quadratic
,
and Logarithmic
sweeps, the instantaneous
frequency, f_{i}(t_{g}),
of the output at the Target time, t_{g}.
For a Swept cosine
sweep, Target
frequency is the instantaneous frequency of the output
at half the Target time, t_{g}/2.
When Frequency sweep is Logarithmic
,
you must set the Target frequency to be greater
than the Initial frequency. Tunable.
For Linear
, Quadratic
,
and Logarithmic
sweeps, the time, t_{g},
at which the Target frequency, f_{i}(t_{g}),
is reached by the sweep. For a Swept cosine
sweep, Target
time is the time at which the sweep reaches 2f_{i}(t_{g})
 f_{0}. You must set Target
time to be no greater than Sweep
time , $${T}_{sw}\ge {t}_{g}$$. Tunable.
In Unidirectional
Sweep
mode, the Sweep time, T_{sw},
is the period of the output frequency sweep. In Bidirectional
Sweep
mode, the Sweep time is half the period
of the output frequency sweep. You must set Sweep time to
be no less than Target time, $${T}_{sw}\ge {t}_{g}$$. Tunable.
The phase, $${\varphi}_{0}$$, of the cosine output at t=0; $${y}_{chirp}(t)=\mathrm{cos}({\varphi}_{0})$$. Tunable.
The sample period, T_{s}, of the output. The output frame period is M_{o}*T_{s}.
The number of samples, M_{o}, to buffer into each output frame.
The data type of the output, singleprecision or doubleprecision.
The first few examples demonstrate how to use the Chirp block's main parameters, how to view the output in the time domain, and how to view the output spectrogram:
Examples 4 and 5 illustrate Chirp block settings that might produce unexpected outputs:
Often times, you might want a unidirectional sweep for which you know the initial and final frequency values. You can specify the final frequency of a unidirectional sweep by setting Target time equal to Sweep time, in which case the Target frequency becomes the final frequency in the sweep. The following model demonstrates this method.
This technique might not work for swept cosine sweeps. For details, see Cautions Regarding the Swept Cosine Sweep.
Open the Example 1 model by typing ex_chirp_ref at the MATLAB^{®} command line. You can also rebuild the model yourself; see the following list for model parameter settings (leave unlisted parameters in their default states).
Since Target time is set to equal Sweep
time (1
second), the Target
frequency (25
Hz) is the final frequency
of the unidirectional sweep. Run your model to see the time domain
output:
Type the following command to view the chirp output spectrogram:
spectrogram(dsp_examples_yout,hamming(128),... 110,[0:.01:40],400)
Chirp Block Parameters for Example 1  
Frequency sweep 

Sweep mode 

Initial frequency  0 
Target frequency  25 
Target time  1 
Sweep time  1 
Initial phase  0 
Sample time  1/400 
Samples per frame  400 
Vector Scope Block Parameters for Example 1  
Input domain 

Time display span  6 
Signal To Workspace Block Parameters for Example 1  
Variable name 

Configuration Dialog Parameters for Example 1  
Stop time  5 
Change the Sweep mode parameter in the
Example 1 model to Bidirectional
, and leave
all other parameters the same to view the following bidirectional
chirp. Note that in the bidirectional sweep, the period of the sweep
is twice the Sweep time (2 seconds), whereas
it was one Sweep time (1 second) for the unidirectional
sweep in Example 1.
Open the Example 2 model by typing ex_chirp_ref2
at the MATLAB command
line.
Run your model to see the time domain output:
Type the following command to view the chirp output spectrogram:
spectrogram(dsp_examples_yout,hamming(128),... 110,[0:.01:40],400)
Setting Sweep time to 1.5
and
leaving the rest of the parameters as in the Example 1 model gives
the following output. The sweep still reaches the Target
frequency (25 Hz) at the Target time (1
second), but since Sweep time is greater than Target
time, the sweep continues on its linear path until one Sweep
time (1.5 seconds) is traversed.
Unexpected behavior might arise when you set Sweep time greater than Target time; see Example 4: Output Sweep with Negative Frequencies for details.
Open the Example 3 model by typing ex_chirp_ref3
at the MATLAB command
line.
Run your model to see the time domain output:
Type the following command to view the chirp output spectrogram:
spectrogram(dsp_examples_yout,hamming(128),... 110,[0:.01:40],400)
Modify the Example 1 model by changing Sweep time to 1.5
, Initial
frequency to 25
, and Target
frequency to 0
. The output
chirp of this example might not behave as you expect because
the sweep contains negative frequencies between 1 and 1.5 seconds.
The sweep reaches the Target frequency of 0 Hz
at one second, then continues on its negative slope, taking on negative
frequency values until it traverses one Sweep time (1.5
seconds).
Open the Example 4 model by typing ex_chirp_ref4
at the MATLAB command
line.
Run your model to see the time domain output:
Modify the Example 1 model by changing the Target
frequency parameter to 275
. The
output chirp of this model might not behave as you expect because
the sweep contains frequencies greater than half the sampling frequency
(200
Hz), which causes aliasing. If you unexpectedly
get a chirp output with a spectrogram resembling the one following,
your chirp's sweep might contain frequencies greater than half the
sampling frequency.
Open the Example 5 model by typing ex_chirp_ref5
at the MATLAB command
line.
Run your model to see the time domain output:
Type the following command to view the chirp output spectrogram:
spectrogram(dsp_examples_yout,hamming(64),... 60,256,400)
Doubleprecision floating point
Singleprecision floating point
Signal From Workspace  DSP System Toolbox 
Signal Generator  Simulink 
Sine Wave  DSP System Toolbox 
chirp  Signal Processing Toolbox 
spectrogram  Signal Processing Toolbox 