Main Content

Linear Prediction

Convert linear predictive coefficients (LPC) to cepstral coefficients, LSF, LSP, RC, and vice versa

Linear prediction estimates the future values of a discrete-time signal as a linear function of previous values. The Autocorrelation LPC block in DSP System Toolbox™ determines the coefficients of an N-step forward linear predictor. A linear predictor is an FIR filter that predicts the next value in a sequence from present and past inputs. For an example that shows this workflow, see Estimate Data Series Using Forward Linear Predictor. Linear prediction has applications in filter design, speech coding, spectral analysis, and system identification.

In digital signal processing, linear prediction is often called linear predictive coding (LPC). DSP System Toolbox offers Simulink® blocks to convert linear prediction coefficients to cepstral coefficients, reflection coefficients (RC), autocorrelation coefficients, line spectral frequencies (LSF), and vice versa.


Autocorrelation LPCDetermine coefficients of Nth-order forward linear predictors
Levinson-DurbinSolve linear system of equations using Levinson-Durbin recursion
LPC to LSF/LSP ConversionConvert linear prediction coefficients to line spectral pairs or line spectral frequencies
LPC to/from Cepstral CoefficientsConvert linear prediction coefficients to cepstral coefficients or cepstral coefficients to linear prediction coefficients
LPC to/from RCConvert linear prediction coefficients to reflection coefficients or reflection coefficients to linear prediction coefficients
LPC/RC to AutocorrelationConvert linear prediction coefficients or reflection coefficients to autocorrelation coefficients
LSF/LSP to LPC ConversionConvert line spectral frequencies or line spectral pairs to linear prediction coefficients