On Oct 31, 2010, at 6:13 PM, Jeff Cochrane - VK4BOF wrote: > I know nothing of CW and it's esoterics but it seems to me that maybe this > APF thingy maybe of some use to digital signals too, as in peaking up a > (very) weak PSK31 signal or something of that nature. > Would that be the case?
IMHO, no. Any PSK31 demodulator worth its salt should already have a matched filter (filter response matched to the keying sidebands of a PSK31 signal). Applying any shaper filter will just raise the bit error rate. The same is true if you use an APF that is sharper than what is determined by the rise and fall times of a CW keying waveform (i.e., increasing the Q of the AFP beyond a certain point will reduce the SNR of a CW signal). In terms of other digital signals, you definitely don't want to apply any filtering ahead of an MFSK signal (MFSK16, DominoEX, etc), since the filter for each tone in those modes are already optimal (sin(x)/s) in shape because everybody uses FFTs to demodulate amateur MFSK. In the RTTY world, "twin peak filters" available in many rigs often give better results only because many software demodulators do not come with matched filters. If you apply a twin peak filter in front of an RTTY demodulator that already has matched filtering that is matched to the RTTY baud rate, you will also degrade print on RTTY signals that have poor SNR and no adjacent channel QRM. Try that with RITTY (an MS-DOS program by K6STI) and cocoaModem, for example -- those are known to use matched filtering. 73 Chen, W7AY ______________________________________________________________ Elecraft mailing list Home: http://mailman.qth.net/mailman/listinfo/elecraft Help: http://mailman.qth.net/mmfaq.htm Post: mailto:Elecraft@mailman.qth.net This list hosted by: http://www.qsl.net Please help support this email list: http://www.qsl.net/donate.html
