A method for predicting missing-error counts at high recording densities

This report describes a prediction method for the missing-error count at high recording densities by means of a simple analysis of error data at a low recording density. The decrease of a signal-peak amplitude is approximately in proportion to the volume ratio of a defect to one bit. Errors can be produced by extraordinary small defects as the recording density is in-ereased, while at lower recording density, errors from these defects can be detected only when the threshold level is set at a much higher level. Upon investigation of this effect the authors have found that the error-count at the higher densities can be predicted on the basis of measurements at low densities. Many experimental investigations were carried out to check the accuracy of the prediction method with a computer-aided disk test system. The predicted error count is in approximate agreement with the experimental resuls, and it is concluded that the prediction method is of considerable practical value.