Equalizer adaptation

Abstract

This invention relates to methods and apparatus for equalizer adaptation for compensating for channel distortion on received data signals. The method comprises, for each bit, forming an adjusted bit signal comprising a weighted contribution from at least one other bit period. The polarity of the adjusted bit signal is determined and the bit is categorized as a hard, i.e. high confidence, bit is the bit is above an upper threshold or below a lower threshold or otherwise is categorized as a soft bit. The weightings are adjusted based on the category of the bit wherein a first adjustment is made it the bit is categorized as a soft bit but a second, different adjustment is made if the bit is categorized as a hard bit. For a soft bit the weightings may be increased for bits which have the same polarity as the bit in question and decreased for bits of opposite polarity. For a hard bit the adjustment may increase the weighting for bits of opposite polarity and decrease the weighting for bits of the same polarity as the bit in question. The adjustments for the hard and soft bits may be weighted differently according to the proportion of soft to hard bits.