chelena
Newbie level 3
In adaptive filters for noice cancellation, one thing is not clear to me.
There is a Desired Signal + Noise (d[k]=s[k]+n[k]) in the primary sensor (mic1). Adaptive filter solution is viable when there is an additional reference noise input
x[k] available that is correlated with the original corrupting noise n[k] in the reference sensor (mic2). The filter filters the reference noise x[k] and produces an estimate of the actual noise n[k] and subtracts it from the primary input s[k]+n[k] to compute an estimate of the signal s[k].
until this its fine. But "For adaptive noise cancelling generally little or no priori knowledge of s[k],n[k], and x[k] or their relationship is needed."
If we don't have the knowledge of the signals how can we adapt the filter? Isn't it contradicting with the above idea? Or else I don't understand the basic concept of a priori knowledge of a signal. Please I need very urgent support!!
Thanks
There is a Desired Signal + Noise (d[k]=s[k]+n[k]) in the primary sensor (mic1). Adaptive filter solution is viable when there is an additional reference noise input
x[k] available that is correlated with the original corrupting noise n[k] in the reference sensor (mic2). The filter filters the reference noise x[k] and produces an estimate of the actual noise n[k] and subtracts it from the primary input s[k]+n[k] to compute an estimate of the signal s[k].
until this its fine. But "For adaptive noise cancelling generally little or no priori knowledge of s[k],n[k], and x[k] or their relationship is needed."
If we don't have the knowledge of the signals how can we adapt the filter? Isn't it contradicting with the above idea? Or else I don't understand the basic concept of a priori knowledge of a signal. Please I need very urgent support!!
Thanks