True linear prediction by use of a theoretical impulse response
J. Kauppinen, Pekka Saarinen
Abstract
J. Kauppinen, Pekka Saarinen
Abstract
A functional form is derived for an impulse response in a time-domain linear prediction based on an autoregressive model. For any impulse-response length, this theoretical impulse response allows us to predict correctly a given time-domain signal consisting of discrete cosine waves. A method is also given for finding the wave frequencies, which are needed in the construction of the impulse response. The properties of this theoretical impulse response are compared with those of the impulse responses computed by Burg’s algorithm or by fitting of the autocorrelation function of the original time-domain signal.
OpenAlex reports 7 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
A functional form is derived for an impulse response in a time-domain linear prediction based on an autoregressive model. For any impulse-response length, this theoretical impulse response allows us to predict correctly a given time-domain signal consisting of discrete cosine waves. A method is also given for finding the wave frequencies, which are needed in the construction of the impulse response. The properties of this theoretical impulse response are compared with those of the impulse responses computed by Burg’s algorithm or by fitting of the autocorrelation function of the original time-domain signal.
Key concepts: Impulse response, Impulse invariance, Time domain, Impulse (physics), Autocorrelation, Infinite impulse response, Autoregressive model, Linear filter