Andy McFadden replied: >Michael J. Mahon wrote: >> My experience with "published" cassettes is that many of them had >> non-standard tape azimuth alignment, causing significant high frequency >> rolloff when played on a properly aligned deck. >> >> The solution was to get out a small screwdriver and individually adjust >> the alignment for each tape, peaking the high frequencies. > >The amplitude of the 2KHz '0' cycles tends to be less than half the >amplitude of the 1KHz '1' cycles on Softape cassettes. It's not uncommon >to get batches of 0s that have been pushed up above the zero line, which is >why I'm providing an alternate algorithm that measures distances between >peaks instead of zero crossings. (Of course, figuring out what a "peak" >is can be a lot of fun on some of these things.) The high-frequency >effects get even worse for the "short 0" half-cycle that's supposed to >start things off... if you miss that, you catch the second half of the >cycle (which is a "regular" 0), and can end up off by half a cycle, >which doesn't work out very well. Even inexpensive tape decks and cassettes should have better response at 2kHz than that, so I continue to suspect that they were recorded with bad azimuth adjustment. The effect of azimuth misalignment is very rapid rolloff in high frequency response. The fact that the change in amplitude is causing your sound card to miss zero crossings suggests that a high-pass filter could help sugstantially. I'm assuming that you have plenty of audio level to play with, so an attenuation of about 20:1 wouldn't be a problem... Then this simple high-pass will boost high frequencies by 6dB/octave, which should bring the 1kHz and 2kHz levels approximately equal: || Cassette out >------||-------+-------> Sound card in || | / .01mf \ / 1k \ | V Gnd. This will also effectively remove the zero offset (since it is a differentiator) which should make the zero crossing times much more dependable. Of course, this kind of correction should only be needed on "problem tapes". -michael Check out parallel computing for 8-bit Apples on my Home page: http://members.aol.com/MJMahon/