Appeal No. 1998-2713
Application No. 08/583,295
channel IC . . . . Abbott discloses that the
adaptive update algorithm can be performed during
normal operation while reading the recorded user
data, or during a “training” mode where the filter
is adapted by reading a known test pattern from the
disk (Abbott, col. 22, lines 26+). Either way, the
adaptive algorithm operates by adjusting the filter
coefficients in “real time” using a single error
value generated with each data sample read from the
disk . . . .
To overcome these drawbacks, the appellant has
[sic, appellants have] disclosed a calibration
method that is not real-time adaptive. Essentially,
the present invention operates by measuring several
error values, and specifically accumulating several
sample error values, over a range of filter
parameter settings, and then programming the filter
according to a predetermined criteria based on the
measured error values, such as the parameter setting
that generates the minimum error value.
We agree with the examiner (paper number 11, pages 2 and
3) that the excised portions of claim 1 are found in Abbott.
On the other hand, we agree with appellants’ argument that
Abbott does not program the filter “with at least one
component setting responsive to the measured error values”
(brief, page 6) (emphasis added). Abbott expressly states
(column 24, lines 14 through 17) that “[d]uring the filter
training mode, the error value on the path 215 causes the
recursive adaptation circuit 222 to adjust the filter
coefficients to minimize the squared-error value” (emphasis
4
Page: Previous 1 2 3 4 5 6 7 8 9 10 Next
Last modified: November 3, 2007