Appeal No. 2006-1690
Application No. 10/154,185
double-talk and suspends replacement of the non-adaptive filter weights before
portions of the first signal are cancelled by the non-adaptive filter as claimed.
Although Park does state in col. 5, line 62 that "it is not necessary to
detect double-talk" in connection with the echo canceler's improved performance
during double-talk, Park's disclosure nevertheless contains other statements that
call for detecting double-talk. For example, in col. 3, lines 39-42, Park states that
"[a] mode controller 25 receives the first and second voice signals U and S, and
the first and second echo-canceled signals e1 and e2, to detect double-talk"
[emphasis added]. Also, in col. 4, lines 38-45, Park emphasizes the importance
of detecting double-talk accurately to effectively cancel echoes during double-
talk. Specifically, Park notes:
[T]he echo canceler 20 performs echo canceling even
in the case of double-talk. It is very important that the
echo canceler 20 detects double talk accurately, since
the adaptive filter 21 quickly generates its output
during double-talk. In general, double-talk is detected
by using an energy level based on the assumption
that a loss caused through an echo path is at least 6
dB [emphasis added].
Moreover, in col. 6, lines 5-10, Park notes:
In practice, it is preferable that the adaptive filter 21
operates inadaptively during double-talk…For this
purpose, double-talk using its energy level is detected
[emphasis added].
Park's repeated references to detecting double-talk reasonably suggest to
the skilled artisan that double-talk is detected notwithstanding the statement in
col. 5, line 62.
7
Page: Previous 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Next
Last modified: November 3, 2007