jj@alice.UUCP (07/27/85)
Hamilton is half right, in that interpolation (in the sense that CD players use it for oversampling) had nothing at all to do with error correction, BUT, I disagree is a few places. The first is that he says that "all filters have phase shift". This statement is, strictly speaking, true, but there is a whole class of filters (symetric and anti-symetric finite impulse response filters) that have a TRANSIT DELAY. While this is equivelant to a phase shift that changes linearly with frequency, it does NOT mean that there is any effective phase shift in the signal. In fact, these (symmetric FIR filters) filters are what are used inthe "interpolation process", not any process that's like what one used to do with a trig table. The comments about switching noise are properly called aliasing, and what interpolation does is remove the aliasing at the low frequency, and then use a simple, and well behaved filter, to remove the high-frequency aliasing that is generated instead. Rabiner and Shaeffer have a good textbook on such things. -- TEDDY BEARS MAY BECOME EXTINCT! HELP AN ENDANGERED SPECIES! "It was great when it all began, ..." (ihnp4/allegra)!alice!jj