Dirk, I have one more question what is the difference between "the difference output or the predictive filter output"?
problem with reference signal in LMS algoritm
Started by ●November 18, 2005
Reply by ●November 18, 20052005-11-18
Reply by ●November 18, 20052005-11-18
Predictive filter output is output of the adaptive FIR (I realize that other architectures are possible) that is filtering the primary input. Difference output is the difference between the reference output and the predictive filter output. LMS filtering can be applied in either mode, but for a given application (which includes signal properties), one particular mode is generally appropriate. You need to think about what you are trying to accomplish with the filtering to select the mode, and to define any delays used, and where any delays are placed, ... Dirk
Reply by ●November 18, 20052005-11-18
And what shall I do if the primary signal which I want to clear haven't got period? My primary signal is only recorded speech with ading white noise - it has got 40000 samples, and I treid to make delay 10, 100, 1000 samples and it didn't work, how to easy find the delay? Is there any way to find it?
Reply by ●November 18, 20052005-11-18
"ralph" <rklimski@o2.pl> wrote in news:1132333722.394654.96050 @g49g2000cwa.googlegroups.com:> And what shall I do if the primary signal which I want to clear haven't > got period? My primary signal is only recorded speech with ading white > noise - it has got 40000 samples, and I treid to make delay 10, 100, > 1000 samples and it didn't work, how to easy find the delay? Is there > any way to find it? > >Try a delay of 1 sample. If you use many samples, the speech is decorrelated to itself. -- Al Clark Danville Signal Processing, Inc. -------------------------------------------------------------------- Purveyors of Fine DSP Hardware and other Cool Stuff Available at http://www.danvillesignal.com
Reply by ●November 18, 20052005-11-18
"ralph" <rklimski@o2.pl> wrote in news:1132333722.394654.96050 @g49g2000cwa.googlegroups.com:> And what shall I do if the primary signal which I want to clear haven't > got period? My primary signal is only recorded speech with ading white > noise - it has got 40000 samples, and I treid to make delay 10, 100, > 1000 samples and it didn't work, how to easy find the delay? Is there > any way to find it? > >The type of noise reduction you are discussing is well known. It is not a great method. There was a ham radio article on exactly this method in the early 1990s. Search on keyword: W9GR -- Al Clark Danville Signal Processing, Inc. -------------------------------------------------------------------- Purveyors of Fine DSP Hardware and other Cool Stuff Available at http://www.danvillesignal.com
Reply by ●November 18, 20052005-11-18
If I delayed of 1 sample the output was just quiet, the same result I have if I delayed of 2 sample all samples was reset to zero. I tried to find something about w9gr. Thanks for all information, and if you know something more about adaptive noise cancellation (other algoritms or other metods) please e-mail me: rklimski@o2.pl or please write on this news. Ralph Klimski
Reply by ●November 18, 20052005-11-18
"ralph" <rklimski@o2.pl> wrote in news:1132348580.690793.76180 @g43g2000cwa.googlegroups.com:> If I delayed of 1 sample the output was just quiet, the same result I > have if I delayed of 2 sample all samples was reset to zero. I tried to > find something about w9gr. > Thanks for all information, and if you know something more about > adaptive noise cancellation (other algoritms or other metods) please > e-mail me: rklimski@o2.pl or please write on this news. > Ralph Klimski > >I know a lot about this topic but we are entering into stuff that we don't share. Here is a link describing the W9GR ANC: http://www.johanforrer.net/ASP/index.html -- Al Clark Danville Signal Processing, Inc. -------------------------------------------------------------------- Purveyors of Fine DSP Hardware and other Cool Stuff Available at http://www.danvillesignal.com
