Method and apparatus for enhancing frequency domain analysis
AssigneeHewlett-Packard Company
InventorRonald W. Potter
Life of this patent
Estimating the term…
Abstract
A signal processing technique allows accurate interpolation between points of a sampled frequency domain function. A time-based sampling process samples a time waveform of duration T having a sub time interval T'. A signal processor applies a discrete Fourier transform over a time period T to transform the sampled data from the time domain to the frequency domain. The sampled frequency domain data is convolved with one or more convolution kernels to yield a continuous line shape. The result of this convolution permits the spectral composition at arbitrary frequencies to be determined. The disclosed frequency domain interpolation process is characterized by preservation of data in the T' interval of the time domain with an arbitrary but specified degree of accuracy.
Claims35
An instrument for determining the spectral composition of a signal, the instrument comprising:
- a sampler for sampling the signal in the time domain over a record length T to produce sampled time domain data;
- DFT means for producing sampled frequency domain data corresponding to the sampled time domain data, the frequency domain data being sampled at frequency intervals of 2/T;
- and
- processing means for processing the sampled frequency domain data to produce interpolated frequency domain data representing a spectral component of the signal at a frequency other than a multiple of 1/T, the interpolated frequency domain data corresponding to the sampled time domain data to within a specified degree of accuracy.
A signal processing method comprising the steps:
- obtaining a plurality of time sampled data points over a record length T';
- producing a sampled frequency spectrum by applying a transform to a set of time domain data that includes the plurality of time sampled data points;
- and
- convolving the sampled frequency spectrum with an interpolation kernel centered about a new frequency to interpolate and resample at said new frequency.
A signal processing method comprising the steps:
- providing a sampled frequency spectrum, the samples being spaced at frequency intervals of 1/T;
- evaluating the convolution of the sampled frequency spectrum with an interpolation kernel at new frequencies spaced less than or equal to 1/T' to interpolate and resample at said new frequencies, said interpolation preserving a protected portion T' of a time domain sequence T corresponding to the sampled frequency spectrum.
A method of producing interpolated values that contain accurate information about the time domain of an analog signal, the method comprising:
- providing a sampled frequency spectrum corresponding to the spectral content of the analog signal at a plurality of discrete frequencies;
- and
- interpolating between said discrete frequencies to obtain frequency domain data at a new frequency, the interpolating comprising the substep of convolving the sampled frequency spectrum with an interpolation kernel centered about the new frequency.
A method of frequency domain analysis comprising:
- sampling an input signal to obtain time domain data;
- processing a record of length T of said time domain data to yield a set of corresponding frequency domain data, said frequency domain data representing spectral components of the input signal at a plurality of spaced frequency points, the spacing of said frequency points depending on the record length T;
- and
- interpolating to obtain frequency domain data for at least one point in between said spaced frequency points, said interpolation preserving the accuracy of a protected interval T' of the time domain record to within a specified accuracy.
A method of frequency domain analysis comprising:
- sampling an input signal to obtain original time domain data;
- processing a record of length T of aid original time domain data to yield a set of corresponding frequency domain data, said frequency domain data representing spectral components of the input signal at a plurality of spaced frequency points;
- and
- interpolating to obtain frequency domain data or at least one point in between said spaced frequency points, the frequency domain data including said interpolated point comprising an enhanced set of frequency domain data, wherein time domain data corresponding to said enhanced set of frequency domain data matches the original time domain data without corruption.
A method of frequency domain analysis comprising:
- sampling an input signal to obtain time domain data;
- processing a record of length T of said time domain data to yield a set of corresponding frequency domain data, said frequency domain data representing spectral components of the input signal at a plurality of spaced frequency points;
- and
- interpolating to obtain frequency domain data for at least one point in between said spaced frequency points, said interpolation including convolving data from said set of corresponding frequency domain data with an interpolation kernel.
In a method of frequency domain analysis of a signal that includes processing a time domain record of length T to obtain a set of frequency domain data corresponding thereto, an improvement comprising processing the frequency domain data to resolve additional frequency domain components of the signal with a resolution not limited by the length T of the time domain record.
The method of claim 28 in which said processing includes interpolating to obtain a continuous spectrum of frequency domain data.
In a method of processing of a time record of length T to obtain a corresponding set of frequency domain data, an improvement comprising interpolating the frequency domain data to yield, additional frequency domain data, said interpolating decoupling resolution of the additional frequency domain data from the length T of the time record.
A method of processing a signal so as to preserve attributes in a complementary Fourier domain comprising:
- obtaining an initial set of sampled data, said set including samples representing a signal's attributes at a plurality of spaced apart points in a first Fourier domain;
- producing an enhanced set of data by convolving said initial set of sampled data with an interpolation kernel so as to yield data at one or more points in addition to said plurality of spaced apart points;
- said interpolation kernel being of such a type that a transformation of the enhanced set of data into a second, complementary Fourier domain substantially matches a transformation of the initial set of data into said second domain.
A spectrum analysis instrument comprising:
- an input port for receiving an input signal;
- a sample coupled to the input port for producing an initial set of sampled data, said initial set of sampled data comprising samples representing attributes of the input signal at a plurality of spaced apart points in a first Fourier domain;
- a display;
- and
- a microprocessor interposed between the sampler and the display, the microprocessor including means for producing an enhanced set of data by convolving said initial set of sampled data with an interpolation kernel so as to yield data at one or more points in addition to said plurality of spaced apart points, said interpolation kernel being of such a type that a transformation of the enhanced set of data into a second, complementary Fourier domain substantially matches a transformation of the initial set o data into said second domain.
A spectrum analysis instrument comprising:
- an input port for for receiving an input signal;
- a sampler coupled to the input port for producing a time domain record of length T corresponding to said input signal;
- a display;
- and
- a microprocessor interposed between the sampler and the display, the microprocessor including means for processing the time domain record to obtain a set of frequency domain data corresponding thereto, the microprocessor further including means for processing the frequency domain data to resolve additional frequency domain components of the signal with a resolution not limited by the length T of the time domain record.
A spectrum analysis instrument comprising:
- an input port for receiving an input signal;
- a sampler coupled to the input port for producing a time record of length T corresponding to said input signal;
- a display;
- and
- a microprocessor interposed between the sampler and the display, the microprocessor including means for processing the time record of length T to obtain a corresponding set of frequency domain data, the microprocessor further including means for interpolating the frequency domain data to yield additional frequency domain data, said interpolating decoupling resolution of the additional frequency domain data from the length T of the time record.
Dependencies are inferred from the claim text.