Revision as of 12:18, 19 January 2023 editBob K (talk | contribs)Extended confirmed users6,614 editsm signal → signal or filter bandwidthTag: Visual edit← Previous edit | Revision as of 17:15, 3 February 2023 edit undoBob K (talk | contribs)Extended confirmed users6,614 edits add a normalization optionTag: Visual edit: SwitchedNext edit → | ||
Line 3: | Line 3: | ||
{{mergewith|Digital frequency|date=September 2012}} --> | {{mergewith|Digital frequency|date=September 2012}} --> | ||
In ] (DSP), a '''normalized frequency''' is a ] |
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. | ||
=== 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 |
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 {{math|''f''}} is expressed in ] (''cycles per second''), {{math|''f''<sub>s</sub>}} is expressed in ''samples per second''.<ref>{{cite book |last=Carlson |first=Gordon E. |title=Signal and Linear System Analysis|publisher=©Houghton Mifflin Co |year=1992 |isbn=8170232384 |location=Boston, MA |pages=469, 490}}</ref> | ||
Some programs (such as ] toolboxes) that design filters with real-valued coefficients prefer the ] ({{math|''f''<sub>s</sub>/2}}) as the |
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. | ||
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 ''N'' (see {{Section link|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''}}. The normalized Nyquist frequency is {{math|''N''/2}} with the unit {{sfrac|1|N}}<sup>th</sup> ''cycle/sample''. | |||
], denoted by {{math|''ω''}} and with the unit '']'', can be similarly normalized. When {{math|''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 {{math|''ω''}} and with the unit '']'', can be similarly normalized. When {{math|''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''. | ||
The following table shows examples of normalized frequencies for a 1 kHz signal (or filter bandwidth), a sampling rate {{math|''f''<sub>s</sub>}} = 44100 ''samples/second'' (often denoted by ]), and |
The following table shows examples of normalized frequencies for a 1 kHz signal (or filter bandwidth), a sampling rate {{math|''f''<sub>s</sub>}} = 44100 ''samples/second'' (often denoted by ]), and 4 normalization options. | ||
{| class="wikitable" | {| class="wikitable" | ||
Line 17: | Line 20: | ||
!'''Quantity''' | !'''Quantity''' | ||
!'''Numeric range''' | !'''Numeric range''' | ||
!''' |
!'''Calculation''' | ||
!'''Value''' | |||
|- | |- | ||
|{{math|1=''f'' / ''f''<sub>s</sub>}} | |{{math|1=''f'' / ''f''<sub>s</sub>}} | ||
| {{math||size=150%}} | | {{math||size=150%}} ''cycle/sample'' | ||
|1000 |
|1000 / 44100 = 0.02268 | ||
|0.02268 cycle/sample | |||
|- | |- | ||
|{{math|1=''f'' / (''f''<sub>s</sub>/2) |
|{{math|1=''f'' / (''f''<sub>s</sub>/2)}} | ||
| | | ''half-cycle/sample'' | ||
|1000 / 22050 = 0.04535 | |||
|2000 half-cycles/second / 44100 samples/second | |||
|- | |||
|0.04535 half-cycle/sample | |||
|{{math|1=''f'' / (''f''<sub>s</sub>/''N'')}} | |||
| {{math||size=150%}} ''bins'' | |||
|1000 × N / 44100 = 0.02268 N | |||
|- | |- | ||
|{{math|''ω'' / ''f''<sub>s</sub>}} | |{{math|''ω'' / ''f''<sub>s</sub>}} | ||
| | | ''radians/sample'' | ||
| |
|1000 × 2π / 44100 = 0.14250 | ||
|0.14250 radian/sample | |||
|} | |} | ||
==See also== | ==See also== |
Revision as of 17:15, 3 February 2023
Frequency divided by a characteristic frequencyIn digital signal processing (DSP), a normalized frequency is a ratio of a variable frequency (f) and a constant frequency associated with a system (such as a sampling rate, fs). 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 (fs) that is used to create the digital signal from a continuous one. The normalized quantity, f′ = f / fs, has the unit cycle per sample regardless of whether the original signal is a function of time or distance. For example, when f is expressed in Hz (cycles per second), fs is expressed in samples per second.
Some programs (such as MATLAB toolboxes) that design filters with real-valued coefficients prefer the Nyquist frequency (fs/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.
A common practice is to sample the frequency spectrum of the sampled data at frequency intervals of fs/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 fs/N. The normalized Nyquist frequency is 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 / fs, the normalized Nyquist angular frequency is π radians/sample.
The following table shows examples of normalized frequencies for a 1 kHz signal (or filter bandwidth), a sampling rate fs = 44100 samples/second (often denoted by 44.1 kHz), and 4 normalization options.
Quantity | Numeric range | Calculation |
---|---|---|
f / fs | cycle/sample | 1000 / 44100 = 0.02268 |
f / (fs/2) | half-cycle/sample | 1000 / 22050 = 0.04535 |
f / (fs/N) | bins | 1000 × N / 44100 = 0.02268 N |
ω / fs | radians/sample | 1000 × 2π / 44100 = 0.14250 |
See also
Citations
- Carlson, Gordon E. (1992). Signal and Linear System Analysis. Boston, MA: ©Houghton Mifflin Co. pp. 469, 490. ISBN 8170232384.