Misplaced Pages

Normalized frequency (signal processing): Difference between revisions

Article snapshot taken from Wikipedia with creative commons attribution-sharealike license. Give it a read and then ask your questions in the chat. We can research this topic together.
Browse history interactively← Previous editNext edit →Content deleted Content addedVisualWikitext
Revision as of 11:43, 9 February 2023 editCitation bot (talk | contribs)Bots5,406,762 edits Add: pages. Removed parameters. Some additions/deletions were parameter name changes. | Use this bot. Report bugs. | Suggested by Anas1712 | #UCB_webform 2806/3813← Previous edit Revision as of 09:51, 29 December 2023 edit undoAndry1992 (talk | contribs)9 editsm →Examples of normalizationTag: Visual editNext edit →
Line 3: Line 3:
{{mergewith|Digital frequency|date=September 2012}} --> {{mergewith|Digital frequency|date=September 2012}} -->


In ] (DSP), a '''normalized frequency''' is a ratio of a variable ] ({{math|''f''}}) and a constant frequency associated with a system (such as a '']'', {{math|''f''<sub>s</sub>}}). Some software applications require normalized inputs and produce normalized outputs, which can be re-scaled to physical units when necessary. Mathematical derivations are usually done in normalized units, relevant to a wide range of applications. In ] (DSP), a '''normalized frequency''' is a ratio of a variable ] (<math>f</math>) and a constant frequency associated with a system (such as a '']'', <math>f_s</math>). Some software applications require normalized inputs and produce normalized outputs, which can be re-scaled to physical units when necessary. Mathematical derivations are usually done in normalized units, relevant to a wide range of applications.


== Examples of normalization == == Examples of normalization ==
A typical choice of characteristic frequency is the '']'' ({{math|''f''<sub>s</sub>}}) that is used to create the digital signal from a continuous one. The normalized quantity, {{math|1=''f''{{′}} = ''f'' / ''f''<sub>s</sub>}}, has the unit ''cycle per sample'' regardless of whether the original signal is a function of time or distance. For example, when {{mvar|f}} is expressed in ] (''cycles per second''), {{math|''f''<sub>s</sub>}} is expressed in ''samples per second''.<ref name=Carlson/> A typical choice of characteristic frequency is the '']'' (<math>f_s</math>) that is used to create the digital signal from a continuous one. The normalized quantity,<math>f' = \frac{f}{f_s}</math>, has the unit ''cycle per sample'' regardless of whether the original signal is a function of time or distance. For example, when <math>f</math> is expressed in ] (''cycles per second''), <math>f_s</math> is expressed in ''samples per second''.<ref name=Carlson/>


Some programs (such as ] toolboxes) that design filters with real-valued coefficients prefer the ] ({{math|''f''<sub>s</sub>/2}}) as the frequency reference, which changes the numeric range that represents frequencies of interest from {{math|}} ''cycle/sample'' to {{math|}} ''half-cycle/sample''. Therefore, the normalized frequency unit is obviously important when converting normalized results into physical units. Some programs (such as ] toolboxes) that design filters with real-valued coefficients prefer the ] (<math>f_s/2</math>) as the frequency reference, which changes the numeric range that represents frequencies of interest from {{math|}} ''cycle/sample'' to {{math|}} ''half-cycle/sample''. Therefore, the normalized frequency unit is obviously important when converting normalized results into physical units.


] ]
A common practice is to sample the frequency spectrum of the sampled data at frequency intervals of {{math|''f''<sub>s</sub>/''N''}}, for some arbitrary integer {{mvar|N}} (see {{slink|Discrete-time_Fourier_transform|Sampling_the_DTFT|nopage=y}}). The samples (sometimes called frequency ''bins'') are numbered consecutively, corresponding to a frequency normalization by {{math|''f''<sub>s</sub>/''N''}}.<ref name=Harris/>{{rp|p.56 eq.(16)}} The normalized Nyquist frequency is {{math|''N''/2}} with the unit {{sfrac|1|N}}<sup>th</sup> ''cycle/sample''. A common practice is to sample the frequency spectrum of the sampled data at frequency intervals of <math>\frac{f_s}{N}</math>, for some arbitrary integer {{mvar|N}} (see {{slink|Discrete-time_Fourier_transform|Sampling_the_DTFT|nopage=y}}). The samples (sometimes called frequency ''bins'') are numbered consecutively, corresponding to a frequency normalization by <math>\frac{f_s}{N}</math>.<ref name=Harris/>{{rp|p.56 eq.(16)}} The normalized Nyquist frequency is <math>\frac{N}{2}</math> with the unit {{sfrac|1|N}}<sup>th</sup> ''cycle/sample''.


], denoted by {{mvar|ω}} and with the unit '']'', can be similarly normalized. When {{mvar|ω}} is normalized with reference to the sampling rate as {{math|1=''ω''′ = ''ω'' / ''f''<sub>s</sub>}}, the normalized Nyquist angular frequency is ''π radians/sample''. ], denoted by {{mvar|ω}} and with the unit '']'', can be similarly normalized. When {{mvar|ω}} is normalized with reference to the sampling rate as <math>\omega' = \frac{\omega}{f_s}</math>, the normalized Nyquist angular frequency is ''π radians/sample''.


The following table shows examples of normalized frequency for {{mvar|f}}&nbsp;=&nbsp;1&nbsp;kHz, &nbsp;{{math|''f''<sub>s</sub>}}&nbsp;=&nbsp;44100&nbsp;''samples/second'' (often denoted by ]), and 4 normalization conventions: The following table shows examples of normalized frequency for <math>f = 1\text{kHz}, f_s = 44100 \text{ samples/second}</math> (often denoted by ]), and 4 normalization conventions:


{| class="wikitable" {| class="wikitable"
Line 24: Line 24:
!'''Reverse''' !'''Reverse'''
|- |-
|{{math|1=''f''{{′}} = ''f'' / ''f''<sub>s</sub>}} |<math>f' = \frac{f}{f_s}</math>
|&nbsp;&nbsp;{{math||size=150%}}&nbsp;''cycle/sample'' |&nbsp;&nbsp;{{math||size=150%}}&nbsp;''cycle/sample''
|1000 / 44100 = 0.02268 |1000 / 44100 = 0.02268
|{{math|1=''f'' = ''f''{{′}} × ''f''<sub>s</sub>}} |<math>f = f' \cdot f_s</math>
|- |-
|{{math|1=''f''{{′}} = ''f'' / (''f''<sub>s</sub>/2)}} |<math>f' = \frac{f}{f_s / 2}</math>
|&nbsp;&nbsp;&nbsp;''half-cycle/sample'' |&nbsp;&nbsp;&nbsp;''half-cycle/sample''
|1000 / 22050 = 0.04535 |1000 / 22050 = 0.04535
|{{math|1=''f'' = ''f''{{′}} × ''f''<sub>s</sub> / ''2''}} |<math>f = f' \cdot \frac{f_s}{2}</math>
|- |-
|{{math|1=''f''{{′}} = ''f'' / (''f''<sub>s</sub>/''N'')}} |<math>f' = \frac{f}{f_s / N}</math>
|&nbsp;&nbsp;{{math||size=150%}}&nbsp;''bins'' |&nbsp;&nbsp;{{math||size=150%}}&nbsp;''bins''
|1000 × {{mvar|N}} / 44100 = 0.02268 {{mvar|N}} |1000 × {{mvar|N}} / 44100 = 0.02268 {{mvar|N}}
|{{math|1=''f'' = ''f''{{′}} × ''f''<sub>s</sub> / ''N''}} |<math>f = f ' \cdot \frac{f_s}{N}</math>
|- |-
|<math>\omega' = \frac{\omega}{f_s}</math>
|{{math|1=''ω''{{′}} = ''ω'' / ''f''<sub>s</sub>}}
|&nbsp;&nbsp;&nbsp;''radians/sample'' |&nbsp;&nbsp;&nbsp;''radians/sample''
|1000 × 2π / 44100 = 0.14250 |1000 × 2π / 44100 = 0.14250
|{{math|1=''ω'' = ''ω''{{′}} × ''f''<sub>s</sub>}} |<math>\omega = \omega' \cdot f_s</math>
|} |}


Revision as of 09:51, 29 December 2023

Frequency divided by a characteristic frequency

In digital signal processing (DSP), a normalized frequency is a ratio of a variable frequency ( f {\displaystyle f} ) and a constant frequency associated with a system (such as a sampling rate, f s {\displaystyle f_{s}} ). Some software applications require normalized inputs and produce normalized outputs, which can be re-scaled to physical units when necessary. Mathematical derivations are usually done in normalized units, relevant to a wide range of applications.

Examples of normalization

A typical choice of characteristic frequency is the sampling rate ( f s {\displaystyle f_{s}} ) that is used to create the digital signal from a continuous one. The normalized quantity, f ′ = f f s {\displaystyle f'={\frac {f}{f_{s}}}} , has the unit cycle per sample regardless of whether the original signal is a function of time or distance. For example, when f {\displaystyle f} is expressed in Hz (cycles per second), f s {\displaystyle f_{s}} is expressed in samples per second.

Some programs (such as MATLAB toolboxes) that design filters with real-valued coefficients prefer the Nyquist frequency ( f s / 2 {\displaystyle f_{s}/2} ) as the frequency reference, which changes the numeric range that represents frequencies of interest from cycle/sample to half-cycle/sample. Therefore, the normalized frequency unit is obviously important when converting normalized results into physical units.

Example of plotting samples of a frequency distribution in the unit "bins", which are integer values. A scale factor of 0.7812 converts a bin number into the corresponding physical unit (hertz).

A common practice is to sample the frequency spectrum of the sampled data at frequency intervals of f s N {\displaystyle {\frac {f_{s}}{N}}} , for some arbitrary integer N (see § Sampling the DTFT). The samples (sometimes called frequency bins) are numbered consecutively, corresponding to a frequency normalization by f s N {\displaystyle {\frac {f_{s}}{N}}} . The normalized Nyquist frequency is N 2 {\displaystyle {\frac {N}{2}}} with the unit ⁠1/N⁠ cycle/sample.

Angular frequency, denoted by ω and with the unit radians per second, can be similarly normalized. When ω is normalized with reference to the sampling rate as ω ′ = ω f s {\displaystyle \omega '={\frac {\omega }{f_{s}}}} , the normalized Nyquist angular frequency is π radians/sample.

The following table shows examples of normalized frequency for f = 1 kHz , f s = 44100  samples/second {\displaystyle f=1{\text{kHz}},f_{s}=44100{\text{ samples/second}}} (often denoted by 44.1 kHz), and 4 normalization conventions:

Quantity Numeric range Calculation Reverse
f ′ = f f s {\displaystyle f'={\frac {f}{f_{s}}}}    cycle/sample 1000 / 44100 = 0.02268 f = f ′ ⋅ f s {\displaystyle f=f'\cdot f_{s}}
f ′ = f f s / 2 {\displaystyle f'={\frac {f}{f_{s}/2}}}    half-cycle/sample 1000 / 22050 = 0.04535 f = f ′ ⋅ f s 2 {\displaystyle f=f'\cdot {\frac {f_{s}}{2}}}
f ′ = f f s / N {\displaystyle f'={\frac {f}{f_{s}/N}}}    bins 1000 × N / 44100 = 0.02268 N f = f ′ ⋅ f s N {\displaystyle f=f'\cdot {\frac {f_{s}}{N}}}
ω ′ = ω f s {\displaystyle \omega '={\frac {\omega }{f_{s}}}}    radians/sample 1000 × 2π / 44100 = 0.14250 ω = ω ′ ⋅ f s {\displaystyle \omega =\omega '\cdot f_{s}}

See also

References

  1. Carlson, Gordon E. (1992). Signal and Linear System Analysis. Boston, MA: ©Houghton Mifflin Co. pp. 469, 490. ISBN 8170232384.
  2. Harris, Fredric J. (Jan 1978). "On the use of Windows for Harmonic Analysis with the Discrete Fourier Transform" (PDF). Proceedings of the IEEE. 66 (1): 51–83. Bibcode:1978IEEEP..66...51H. CiteSeerX 10.1.1.649.9880. doi:10.1109/PROC.1978.10837. S2CID 426548.
Categories: