AUTOCORRELATION APPLIED TO THE MAGNETIC RECORDING CHANNEL

Abstract

Today I will describe the application of the Autocorrelation function to the Magnetic Recording Channel. I will explain what is an autocorrelated function, how does it behave and where may it be applied in the Magnetic Recording channel. There will be a brief description of Kodak San Diego’s Autocorrelator and how we apply this technology. If I have done my job well at the end of this presentation you will have enough knowledge about autocorrelation to access your own application. Before I start, let me give a brief overview on the application of an Autocorrelator. The Autocorrelator can be used to collect information on signals in a magnetic recording system and display this information graphically as a statistical plot. Autocorrelation, in the time domain, is the counter part to a spectrum analyzer in the frequency domain (Fourier Pair). The information about the signal of interest must be stored for post analysis. This information called a database must then be processed by a computer. The computer passes the database through the autocorrelation algorithm and produces a second database. This second database represents a plot of the autocorrelated function. The next step is to plot the database on a video screen. This plot can be examined for periodicities, randomness, and relational influences on a captured signal. In our application, this signal is an error flag or a dropout flag. We want a statistical picture of the magnitude of errors and their relative frequency. The information gained from Autocorrelation can aid in solutions for: Error Correction Codes Media Evaluation/Qualifications Media Process Defect Identification Mechanical Eccentricities Modulation Code Performances System’s Figure of Merit To use a cliche, “one picture is worth a thousand words,” is exactly the point of the Autocorrelator’s graphical display. It yields information useful to those disciplines which often find difficulty in describing an event in understandable terms.