Appeal No. 1997-3914
Application No. 08/384,457
speakers for a given word into a parallel node model
60. (Column 11, lines 37-42.) (Emphasis added.)
It is this intermediate step of Gillick that meets the claim
language of having “one set of reference pattern data for each
of said plurality of existing speakers”. We read (b) of claim
9, the analyzing step, as part of the “concatenating” which
places the multi-speaker information into a single reference
pattern. In this manner, 100 speakers in Gillick would have
produced 100 node models 20A from 100 different speakers for a
given word. Then, these 100 patterns (node models 20A) would
be concatenated as part of the analyzing step into the single
reference pattern acknowledged by Appellant supra. Thus,
Appellant’s step (a) language is met by Gillick.
With respect to step (b) of claim 9, Appellant argues,
“because Gillick does not teach or suggest generating the
plurality of sets of reference pattern data, one set for each
of the existing speakers, as required by step (a) of Claim 9,
Gillick necessarily fails to teach or suggest analyzing the
sets of reference pattern data.” (Brief-page 6.)
However, as we have noted supra, Gillick does teach
generating the plurality of sets of reference pattern data,
6
Page: Previous 1 2 3 4 5 6 7 8 9 10 11 Next
Last modified: November 3, 2007