Adaptive Filter | ADPCM | ADSP | ADSP-2181 | Aliasing | AMR | Anti-Aliasing | ARMA | Autocorrelation | AutoCovariance | Beamforming | Bessel | Blackfin | Butterworth | C6713 | CCS | Chebyshev | CIC Filter | Circular Convolution | Code Composer Studio | Comb Filter | Compression | Convolution | Cross Correlation | DCT | Decimation | Deconvolution | Demodulation | DM642 | DSP Boards | DSP/BIOS | DTMF | Echo Cancellation | Equalization | Equalizer | ETSI | EZLITE (Ez-kit Lite) | FFT | FFTW | FIR Filter | Fixed Point | FSK | G.711 | G.723 | G.729 | Gaussian Noise | Goertzel | GPIO | Hilbert Transform | IFFT | IIR Filter | Interpolation | Invariance | JTAG | Kalman | Laplace Transform | Levinson | LPC | McBSP | MIPS | Modulation | MPEG | Multirate | Notch Filter | Nyquist | OFDM | Oversampling | Pink Noise | Pitch | PLL | Polyphase | QAM | QDMA | Quantization | Quantizer | Radar | Random Noise | Reed Solomon | Remez | Resampling | RTDX | Sampling | Sharc | TI C6711 | Undersampling | Viterbi | Wavelets | White Noise | Wiener Filter | Windowing | XDS510PP | Z Transform

Dear all,

I am confused by the four transforms in Signal & Systems...

The Continuous Time Fourier Transform(CTFT) is most understandable; DFT and 
Fourier Series alone are individually recoginizable and understandable... 
Not sure about how does DTFT kick in...

Anyway, remembering all of these four transforms' formulas are already very 
headache... very easily got confuse one with another...

Even worse, homework and test problems often asks for conversion among these 
four transforms...

Given a signal's CTFT, how do you get DFT for N-point? How does the DFT 
compare to the Fourier Series(looks to me they are all discrete spectrum, 
etc.) so on and so forth, how are they related and how to get one from 
another?

Are there any good resources that clearly demonstrate the relationship and 
conversion among these 4 transforms?

Thanks a lot,

______________________________
New DSP Code Snippets Section now Live.   Learn more about the reward program for contributors here.

Kiki,

The DTFT takes a discrete time domain signal and gives you a continuous, 
periodic frequency domain signal.

The various transforms can be summarized as follows:

CTFT:  continuous <--> continuous
FS:  continuous <--> discrete
DTFT:  discrete <--> continuous
DFT:  discrete <--> discrete

I don't have a good reference for you concerning conversions among these. 
They all are more or less the same (e.g. sinc function time is rectangle in 
frequency, etc.).  The only thing "tricky" is probably the scale factor 
since all of these are orthonormal transforms and maintain the same signal 
power between time and frequency.

Brad


"kiki" <l...@yahoo.com> wrote in message 
news:cn9qg1$qbd$1...@news.Stanford.EDU...
> Dear all,
>
> I am confused by the four transforms in Signal & Systems...
>
> The Continuous Time Fourier Transform(CTFT) is most understandable; DFT 
> and Fourier Series alone are individually recoginizable and 
> understandable... Not sure about how does DTFT kick in...
>
> Anyway, remembering all of these four transforms' formulas are already 
> very headache... very easily got confuse one with another...
>
> Even worse, homework and test problems often asks for conversion among 
> these four transforms...
>
> Given a signal's CTFT, how do you get DFT for N-point? How does the DFT 
> compare to the Fourier Series(looks to me they are all discrete spectrum, 
> etc.) so on and so forth, how are they related and how to get one from 
> another?
>
> Are there any good resources that clearly demonstrate the relationship and 
> conversion among these 4 transforms?
>
> Thanks a lot,
>
> 

______________________________
New DSP Code Snippets Section now Live.   Learn more about the reward program for contributors here.

Brad Griffis wrote:
> Kiki,
> 
> The DTFT takes a discrete time domain signal and gives you a continuous, 
> periodic frequency domain signal.
> 
> The various transforms can be summarized as follows:
> 
> CTFT:  continuous <--> continuous
> FS:  continuous <--> discrete
> DTFT:  discrete <--> continuous
> DFT:  discrete <--> discrete
>


Writing things down as it it were always time and frequency there are
four kinds of FTs. Two have unbounded time and two have periodic time
as unbounded and periodic are a yes/no pair. Also two have continuous
time and two have discrete time which is another yes/no pair. So you
get the list above which should also show whether time is unbounded or
periodic. The real interest however is in the kind of frequencies that
go with the various kinds of time. If time is periodic then the
frequencies can only take on discrete values and if the time is
unbounded then the frequencies can take continuous values. If the time
is discrete then the frequencies will be periodic but if time is
continuous then the frequencies can be unbounded. So the complete
table would be

               TIME                   FREQUENCY
CTFT:  unbounded continuous <--> continuous unbounded
FS:    periodic  continuous <--> discrete   unbounded
DTFT:  unbounded discrete   <--> continuous periodic
DFT:   periodic  discrete   <--> discrete   periodic

CTFT is the classical Fourier Transform.
FS, or Fourier Series, is the Fourier Transform of rotation angles.
DTFT, or Fourier Sequences, is the Fourier Transform of sampled time.
DFT is the Discrete Fourier Transform of numerical computation.

The names Fourier Series and Fourier Sequences are not so standard
that they can assumed known by everyone. Usually it is just FT for
CTFT which has been used because of the lack of a good name for
Fourier Sequences.

Notice that unbounded and continuous go together just as periodic
and discrete go together whether time or frequency.

If you multiply by a Dirac Comb then you get discrete time and if
you convolve with a Dirac Comb then you get periodic time. If you
do both then your get discrete periodic time. A Dirac Comb is either
a good heuristic or a very subtle discussion of distribution theory
depending upon whether is shows up in Inroduction to Engineering
Calculus or Advanced Measure Theory.


> I don't have a good reference for you concerning conversions among these. 
> They all are more or less the same (e.g. sinc function time is rectangle in 
> frequency, etc.).  The only thing "tricky" is probably the scale factor 
> since all of these are orthonormal transforms and maintain the same signal 
> power between time and frequency.
> 
> Brad
> 
> 
> "kiki" <l...@yahoo.com> wrote in message 
> news:cn9qg1$qbd$1...@news.Stanford.EDU...
> 
>>Dear all,
>>
>>I am confused by the four transforms in Signal & Systems...
>>
>>The Continuous Time Fourier Transform(CTFT) is most understandable; DFT 
>>and Fourier Series alone are individually recoginizable and 
>>understandable... Not sure about how does DTFT kick in...
>>
>>Anyway, remembering all of these four transforms' formulas are already 
>>very headache... very easily got confuse one with another...
>>
>>Even worse, homework and test problems often asks for conversion among 
>>these four transforms...
>>
>>Given a signal's CTFT, how do you get DFT for N-point? How does the DFT 
>>compare to the Fourier Series(looks to me they are all discrete spectrum, 
>>etc.) so on and so forth, how are they related and how to get one from 
>>another?
>>
>>Are there any good resources that clearly demonstrate the relationship and 
>>conversion among these 4 transforms?
>>
>>Thanks a lot,
>>
>>
> 
> 
> 

______________________________
New DSP Code Snippets Section now Live.   Learn more about the reward program for contributors here.

Gordon Sande wrote:

You can also view the DFT as a filter bank, as the Z transform evaluated 
at a set of points on the unit circle, or an orthogonal matrix operation.


> 
> 
> Brad Griffis wrote:
> 
>> Kiki,
>>
>> The DTFT takes a discrete time domain signal and gives you a 
>> continuous, periodic frequency domain signal.
>>
>> The various transforms can be summarized as follows:
>>
>> CTFT:  continuous <--> continuous
>> FS:  continuous <--> discrete
>> DTFT:  discrete <--> continuous
>> DFT:  discrete <--> discrete
>>
> 
> 
> Writing things down as it it were always time and frequency there are
> four kinds of FTs. Two have unbounded time and two have periodic time
> as unbounded and periodic are a yes/no pair. Also two have continuous
> time and two have discrete time which is another yes/no pair. So you
> get the list above which should also show whether time is unbounded or
> periodic. The real interest however is in the kind of frequencies that
> go with the various kinds of time. If time is periodic then the
> frequencies can only take on discrete values and if the time is
> unbounded then the frequencies can take continuous values. If the time
> is discrete then the frequencies will be periodic but if time is
> continuous then the frequencies can be unbounded. So the complete
> table would be
> 
>               TIME                   FREQUENCY
> CTFT:  unbounded continuous <--> continuous unbounded
> FS:    periodic  continuous <--> discrete   unbounded
> DTFT:  unbounded discrete   <--> continuous periodic
> DFT:   periodic  discrete   <--> discrete   periodic
> 
> CTFT is the classical Fourier Transform.
> FS, or Fourier Series, is the Fourier Transform of rotation angles.
> DTFT, or Fourier Sequences, is the Fourier Transform of sampled time.
> DFT is the Discrete Fourier Transform of numerical computation.
> 
> The names Fourier Series and Fourier Sequences are not so standard
> that they can assumed known by everyone. Usually it is just FT for
> CTFT which has been used because of the lack of a good name for
> Fourier Sequences.
> 
> Notice that unbounded and continuous go together just as periodic
> and discrete go together whether time or frequency.
> 
> If you multiply by a Dirac Comb then you get discrete time and if
> you convolve with a Dirac Comb then you get periodic time. If you
> do both then your get discrete periodic time. A Dirac Comb is either
> a good heuristic or a very subtle discussion of distribution theory
> depending upon whether is shows up in Inroduction to Engineering
> Calculus or Advanced Measure Theory.
> 
> 
>> I don't have a good reference for you concerning conversions among 
>> these. They all are more or less the same (e.g. sinc function time is 
>> rectangle in frequency, etc.).  The only thing "tricky" is probably 
>> the scale factor since all of these are orthonormal transforms and 
>> maintain the same signal power between time and frequency.
>>
>> Brad
>>
>>
>> "kiki" <l...@yahoo.com> wrote in message 
>> news:cn9qg1$qbd$1...@news.Stanford.EDU...
>>
>>> Dear all,
>>>
>>> I am confused by the four transforms in Signal & Systems...
>>>
>>> The Continuous Time Fourier Transform(CTFT) is most understandable; 
>>> DFT and Fourier Series alone are individually recoginizable and 
>>> understandable... Not sure about how does DTFT kick in...
>>>
>>> Anyway, remembering all of these four transforms' formulas are 
>>> already very headache... very easily got confuse one with another...
>>>
>>> Even worse, homework and test problems often asks for conversion 
>>> among these four transforms...
>>>
>>> Given a signal's CTFT, how do you get DFT for N-point? How does the 
>>> DFT compare to the Fourier Series(looks to me they are all discrete 
>>> spectrum, etc.) so on and so forth, how are they related and how to 
>>> get one from another?
>>>
>>> Are there any good resources that clearly demonstrate the 
>>> relationship and conversion among these 4 transforms?
>>>
>>> Thanks a lot,
>>>
>>>
>>
>>
>>

______________________________
New DSP Code Snippets Section now Live.   Learn more about the reward program for contributors here.

Folks,

I have found these issues to be clearly discussed in:

   1.  "Digital Spectral Analysis with Applications", by S. Lawrence 
Marple, Prentice-Hall, 1987, Chapter 2.

   2.  "Digital Signal Processing, Principles, Algorithms, and 
Applications, 3'rd Edition" by John Proakis and Dimitris Manolakis, 
Prentice-Hall, 1996, Chapters 4 and 5.

The Marple book has the shorter explanation and covers the scale factors 
relating the discrete forms to the continuous.  Unfortunately, it is out 
of print (except perhaps on ebay!).  The Proakis book spreads out the 
coverage of this material, and although somewhat pricey (I think), 
discusses many other very useful topics.

Both books are well-worth having on your shelf.


Brad Griffis wrote:

> Kiki,
> 
> The DTFT takes a discrete time domain signal and gives you a continuous, 
> periodic frequency domain signal.
> 
> The various transforms can be summarized as follows:
> 
> CTFT:  continuous <--> continuous
> FS:  continuous <--> discrete
> DTFT:  discrete <--> continuous
> DFT:  discrete <--> discrete
> 
> I don't have a good reference for you concerning conversions among these. 
> They all are more or less the same (e.g. sinc function time is rectangle in 
> frequency, etc.).  The only thing "tricky" is probably the scale factor 
> since all of these are orthonormal transforms and maintain the same signal 
> power between time and frequency.
> 
> Brad
> 
> 
> "kiki" <l...@yahoo.com> wrote in message 
> news:cn9qg1$qbd$1...@news.Stanford.EDU...
> 
>>Dear all,
>>
>>I am confused by the four transforms in Signal & Systems...
>>
>>The Continuous Time Fourier Transform(CTFT) is most understandable; DFT 
>>and Fourier Series alone are individually recoginizable and 
>>understandable... Not sure about how does DTFT kick in...
>>
>>Anyway, remembering all of these four transforms' formulas are already 
>>very headache... very easily got confuse one with another...
>>
>>Even worse, homework and test problems often asks for conversion among 
>>these four transforms...
>>
>>Given a signal's CTFT, how do you get DFT for N-point? How does the DFT 
>>compare to the Fourier Series(looks to me they are all discrete spectrum, 
>>etc.) so on and so forth, how are they related and how to get one from 
>>another?
>>
>>Are there any good resources that clearly demonstrate the relationship and 
>>conversion among these 4 transforms?
>>
>>Thanks a lot,
>>
>>
> 
> 
> 

______________________________
New DSP Code Snippets Section now Live.   Learn more about the reward program for contributors here.

"kiki" <l...@yahoo.com> wrote in message news:<cn9qg1$qbd$1...@news.Stanford.EDU>...
> Dear all,
> 
> I am confused by the four transforms in Signal & Systems...
> 
> The Continuous Time Fourier Transform(CTFT) is most understandable; DFT and 
> Fourier Series alone are individually recoginizable and understandable... 
> Not sure about how does DTFT kick in...
> 

I know several people have said out the point. Here's another
perspective.
Depending upon a signal being continuos/discrete in time and
periodic/aperiodic in nature, you have CTFS/CTFT/DTFS/DTFT which
represents the corresponding frequency characteristics (thumb rule:
Series<=> periodicity & transform <=> aperiodic).
Of specific importance is DTFT which is a continuous function of
frequency. Inorder to make it available for practical processing, one
does sampling of DTFT which is nothing but DFT.
-KK

______________________________
New DSP Code Snippets Section now Live.   Learn more about the reward program for contributors here.

Check http://www.ee.vt.edu/~ee4624ss/week4.pdf




"kiki" <l...@yahoo.com> wrote in message news:<cn9qg1$qbd$1...@news.Stanford.EDU>...
> Dear all,
> 
> I am confused by the four transforms in Signal & Systems...
> 
> The Continuous Time Fourier Transform(CTFT) is most understandable; DFT and 
> Fourier Series alone are individually recoginizable and understandable... 
> Not sure about how does DTFT kick in...
> 
> Anyway, remembering all of these four transforms' formulas are already very 
> headache... very easily got confuse one with another...
> 
> Even worse, homework and test problems often asks for conversion among these 
> four transforms...
> 
> Given a signal's CTFT, how do you get DFT for N-point? How does the DFT 
> compare to the Fourier Series(looks to me they are all discrete spectrum, 
> etc.) so on and so forth, how are they related and how to get one from 
> another?
> 
> Are there any good resources that clearly demonstrate the relationship and 
> conversion among these 4 transforms?
> 
> Thanks a lot,

______________________________
New DSP Code Snippets Section now Live.   Learn more about the reward program for contributors here.