tag:blogger.com,1999:blog-3326773214329183284.post2651292026320420134..comments2014-09-14T10:36:04.381-04:00Comments on Ham Radio Blog by AG1LE: Experiment: Decoding multiple Morse code signals automatically on a noisy RF bandMauri Niininenhttps://plus.google.com/100950409777337002358noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-3326773214329183284.post-27093125211619149582012-12-25T17:25:35.020-05:002012-12-25T17:25:35.020-05:00Hi David
Thanks for the tip. I did this experime...Hi David <br />Thanks for the tip. I did this experiment on very simple matched filter code using Octave (Matlab compatible open source language). Since I have well defined test audio files this approach works well. In real life the phase, strength and even frequency of morse code signals vary, not to mention speed variances. I have not tried DFT in Octave simulation but in the FLDIGI implementation there are two filtering options - one using frequency domain matched filter and another one using FFT filter (see http://www.w1hkj.com/FldigiHelp-3.21/configCW.html). In lower signal to noise situations matched filter seems to work a bit better. <br /><br />- 73 de Mauri, AG1LE Mauri, AG1LEhttp://www.blogger.com/profile/16415319751367496314noreply@blogger.comtag:blogger.com,1999:blog-3326773214329183284.post-22184655592829637342012-12-25T13:46:17.618-05:002012-12-25T13:46:17.618-05:00Hi Mauri -- very interesting directions you are go...Hi Mauri -- very interesting directions you are going. On your matched filter, you are assuming a sinewave burst, but depending on starting phase of received audio, it might be an arbitrary shift of the sine wave. So a better possible approach is the RSS of cross correlation against both a sinewave and a cosine wave (same as DFT). That keeps it detecting well, regardless of relative phase shift between signal and template.<br /><br />- 73 de Dave, N7AIGDavid McClainhttp://www.blogger.com/profile/16702445640300113747noreply@blogger.comtag:blogger.com,1999:blog-3326773214329183284.post-82894348043385611522012-07-31T09:49:18.178-04:002012-07-31T09:49:18.178-04:00Nice work, Mauri! Reminds me of my initial 1990&#...Nice work, Mauri! Reminds me of my initial 1990's PhD research renewed recently with YT7PWR help. I work with Morse Runner simulator of HF contests. Life is easy until QRM is enabled. You might find more at Facebook or Googlecode.<br /><br />73 de Mario, S56A, N1YUMarijan Miletichttp://www.blogger.com/profile/05361739967163736128noreply@blogger.com