can be described by a function of the frequency variable ‘ω’ or H(ω), which is the Fourier transform of the unit sample response h(n) of the system.
•H(ω) allows us to determine the
steady-state response of the system to any arbitrary weighted linear combination of sinusoids or complex exponentials. Frequency domain analysis of LTI systems
•Fourier series decomposition of
periodic sequences can be represented as weighted sum of harmonically related exponentials.
•Aperiodic sequences can be
expressed as superposition of infinitesimal size complex exponentials. System response to complex exponentials •Response of any LTI system to arbitrary input signal x(n) is ….. Convolution sum
• Let system input be
Then
• Thus, response y(n) is also a complex
exponential with same frequency as input. • is called eigenfunction and H(ω) is called eigenvalue at that frequency. System response to complex exponentials -Example System response to complex exponentials • In general, H(ω) is complex valued function of ω. So, = ). • Since H(ω) is Fourier transform of h(k), H(ω) is periodic with period 2π. System response to complex exponentials -Example System response to complex exponentials-Example System response to complex exponentials-Example System response to complex exponentials-Example System response to complex exponentials-Example System response to complex exponentials-Example Frequency domain characteristics of LTI systems- summary • If the input to an LTI system consists of linear combination of sinusoids of form
where and are the amplitudes and phases of the sinusoidal
components, the system response is
• A LTI system is functions as a filter to sinusoids of different
frequencies, passing some frequency components to the output and suppressing or preventing others. Filter design involves determining LTI system parameters for achieving a desired Steady state and transient response to sinusoidal input signals
• If exponential or sinusoidal input signals
are applied to a LTI system at n=-ꝏ, the response that is observed is the steady- state response. • If the input signal is applied to the LTI system at some finite time instant say t=0, the response consist of two terms – the transient response and the steady state response. Steady state and transient response to sinusoidal input signals Steady state response to periodic input signals Steady state response to aperiodic input signals Steady state response to aperiodic input signals- Example The discrete Fourier transform (DFT) • A digital computer can work only with discrete data (sequence of numbers) • To enable frequency analysis, we convert the time domain sequence to its frequency domain representation i.e. the Fourier transform X(ω) of x(n). • But X(ω) is a continuous function. • DFT involves representation of a aperiodic sequence x(n) by samples of its spectrum X(ω). • So, we need to relate x(n) to samples of X(ω). The discrete Fourier transform (DFT)
• DFT involves representation of a aperiodic
sequence x(n) by samples of its spectrum X(ω). • So, we need to relate x(n) to samples of X(ω). • Samples of X(ω) do not uniquely represent the original sequence x(n) when x(n) has infinite duration. • Samples of X(ω) uniquely represent x(n) when x(n) is of finite duration. The discrete Fourier transform (DFT)-example The discrete Fourier transform (DFT)-example The discrete Fourier transform (DFT) The discrete Fourier transform (DFT)-example The discrete Fourier transform (DFT)-example The discrete Fourier transform (DFT)-example The discrete Fourier transform (DFT)-zero padding DFT as a linear transformation DFT – complexity of computation DFT - complexity of computation DFT - Example Properties of DFT Properties of DFT Properties of DFT Properties of DFT Properties of DFT Properties of DFT Product of two DFTs and circular convolution Circular convolution-example Circular convolution-example contd. Circular convolution-example contd. Circular convolution-example contd. Circular convolution-example 2 More properties of DFT DFT – complexity of computation
Direct computation of DFT is inefficient as it does not exploit the
symmetry and the periodicity property of WN Efficient algorithms for computation of DFT- FFT Efficient algorithms for computation of DFT- FFT Efficient algorithms for computation of DFT- FFT Efficient algorithms for computation of DFT- FFT Efficient algorithms for computation of DFT- FFT FFT - example Efficient algorithms for computation of DFT- FFT Radix 2 FFT algorithms Radix 2 FFT algorithms