Ex Parte Romesburg

Board of Patent Appeals and Interferences

Jan 29, 2004

09131167 (B.P.A.I. Jan. 29, 2004)

1 Application for patent filed August 7, 1998, entitled
"Methods and Apparatus for Mitigating the Effects of Microphone
Overload in Echo Cancelation Systems."
- 1 -
The opinion in support of the decision being
entered today was not written for publication
and is not binding precedent of the Board.
_______________
Paper No. 26
UNITED STATES PATENT AND TRADEMARK OFFICE

BEFORE THE BOARD OF PATENT APPEALS
AND INTERFERENCES

Ex parte ERIC DOUGLAS ROMESBURG

Appeal No. 2002-1197
Application 09/131,1671

ON BRIEF

Before BARRETT, DIXON, and GROSS, Administrative Patent Judges.
BARRETT, Administrative Patent Judge.
DECISION ON APPEAL
This is a decision on appeal under 35 U.S.C. § 134(a) from
the final rejection of claims 1-41.
We affirm-in-part and enter a new ground of rejection.
Appeal No. 2002-1197
Application 09/131,167
- 2 -
BACKGROUND
The invention relates to echo cancelation in a
bi-directional communications link, such as a telephone system.
In such systems, speech of a near-end user is detected by a
near-end microphone for transmission over a communications link
to a far-end loudspeaker and speech of a far-end user is detected
by a far-end microphone and transmitted to a near-end loudspeaker
for reproduction and presentation to the near-end user. At
either end of the communications link, loudspeaker output
detected by a proximate microphone may inadvertently be
transmitted back over the communications link, resulting in
unacceptable feedback or "echo." This is especially a problem in
portable telephones of small size having a close distance between
the loudspeaker and microphone and with hands-free speaker
phones. Conventionally, echo suppression has been accomplished
with echo canceling circuits which employ adaptive filters.
According to appellant, practical experience has demonstrated
that adaptive filtering does not work well when the source signal
(i.e., the near-end microphone signal) becomes saturated. The
present invention overcomes the problem by updating the filter
coefficients of the adaptive filter only when the magnitude of
the source signal is within a predetermined range of possible
values less than the saturation level.
Appeal No. 2002-1197
Application 09/131,167
- 3 -
Claim 17 is reproduced below.
17. A communications device, comprising:
an adaptive system configured to estimate a component
of a source signal,
wherein variables of said adaptive system are updated
only when a prevailing value of the source signal is within
a predetermined range of possible source signal values.
The examiner relies on the following references:
Haneda et al. (Haneda) 5,721,772 February 24, 1998
Koski 6,192,126 February 20, 2001
(filed May 28, 1998)
Claims 1-41 stand rejected under 35 U.S.C. § 103(a) as being
unpatentable over Koski and Haneda. The examiner finds that
Koski discloses the claimed subject matter except for the echo
estimate. The examiner finds that Haneda teaches providing an
echo estimate. The examiner concludes that it would have been
obvious to modify Koski to provide an echo estimate, in view of
the teachings of Haneda, to allow the echo component of the
source signal to be subtracted out.
We refer to the final rejection (Paper No. 12) (pages
referred to as "FR__") and the examiner's answer (Paper No. 18)
(pages referred to as "EA__") for a statement of the examiner's
rejection, and to the brief (Paper No. 17) (pages referred to as
"Br__") and reply brief (Paper No. 20) (pages referred to as
"RBr__") for a statement of appellant's arguments thereagainst.
Appeal No. 2002-1197
Application 09/131,167
- 4 -
OPINION
Grouping of claims
Appellant argues several groups of claims separately (Br7):
(1) claims 1-4, 8-28, and 32-37 stand or fall together;
(2) claims 5, 7, 29, and 31 stand or fall together; (3) claim 6
stands alone; (4) claim 30 stands alone; (5) claim 38 stands
alone; and (6) claims 39-41 stand or fall together.
New ground of rejection pursuant to 37 CFR § 1.196(b)
Claims 1-24 are rejected under 35 U.S.C. § 112, first and/or
second paragraph, because claims 1, 13, and 17 contain purely
functional language that either lacks enabling disclosure for the
scope of the claim or is indefinite.
Claim 1 recites "wherein filter coefficients of said
adaptive filter are updated only when a prevailing value of the
output signal of said microphone is within a predetermined range
of possible output signal values." Claim 13 recites an almost
identical limitation. Claim 17 recites "wherein variables of
said adaptive system are updated only when a prevailing value of
the source signal is within a predetermined range of possible
source signal values." These limitations are purely functional
because no structure or "means" has been recited to support the
function. In Halliburton Oil Well Cementing Co. v. Walker,
329 U.S. 1, 71 USPQ 175 (1946), the Supreme Court held invalid an
Appeal No. 2002-1197
Application 09/131,167
- 5 -
apparatus claim on the ground that it used a "means-plus-
function" term which was considered to be purely functional.
Such a claim was improper because the means term with a stated
function merely described a particular end result, did not set
forth any specific structure, and would encompass any and all
structures for achieving that result, including those which were
not what the applicant had invented. In Greenberg v. Ethicon
Endo-Surgery Inc., 91 F.3d 1580, 1584, 39 USPQ2d 1783, 1785 (Fed.
Cir. 1996), the Court of Appeals for the Federal Circuit noted
that Congress enacted 35 U.S.C. § 112, sixth paragraph
(originally third paragraph), to overrule that holding. In place
of the Halliburton rule, Congress adopted a compromise solution,
one that had support in the pre-Halliburton case law: Congress
permitted the use of means-plus-function language in claims, but
it limited the breadth of such claim language by restricting its
scope to the structure disclosed in the specification and
equivalents thereof. Thus, apparatus claims must either recite
structure to perform a function or must recite means-plus-
function to be interpreted under § 112, sixth paragraph; they
cannot recite purely functional language. Although the statutory
basis of the rejection is not clear, § 112, first paragraph, lack
of enablement is appropriate because the scope of the enabling
disclosure is not commensurate with the scope of the claim
because the specification does not describe all structures for
Appeal No. 2002-1197
Application 09/131,167
- 6 -
performing the function (this appears to be the rationale in
Halliburton). Also, § 112, second paragraph, indefiniteness is
appropriate because the purely functional limitations in the
"wherein" clauses are not permitted under the Halliburton
rationale and, hence, are not definite claim limitations.
Obviousness
Claims 1-4, 8-28, and 32-37
The issue with respect to the independent claims in this
group is whether the combination of Koski and Haneda discloses or
suggests that the "filter coefficients of said adaptive filter
are updated only when a prevailing value of the output signal of
said microphone is within a predetermined range of possible
output signal values" (claim 1 and an almost identical limitation
in claim 13), or "variables of said adaptive system are updated
only when a prevailing value of the source signal is within a
predetermined range of possible source signal values" (claim 17),
or "updating filter coefficients of the adaptive filter . . .
only when a prevailing value of the communications signal is
within a predetermined range of possible signal values"
(claim 25). Initially, we must interpret these limitations
without reading in limitations from the specification. The
limitation of "within a predetermined range of possible signal
values" is very broad and does not exclude the "predetermined
Appeal No. 2002-1197
Application 09/131,167
- 7 -
range" from being the full range of possible signal values.
Thus, updating the filter coefficients "only when a prevailing
value of the output signal of said microphone is within a
predetermined range of possible output signal values" (claim 1)
does not preclude always updating the filter coefficients where
the "predetermined range" is the entire range. The claims do not
define a "predetermined range" as being less than the total
possible range; compare claim 2, which recites only updating when
a value of the output signal is "below a predetermined threshold
value" and claim 38, which recites only updating when a value of
the output signal is "less than a saturation constant." Although
the word "only" implies that there are some values of output
signal for which the filter coefficients are not updated, we
decline to read this as a limitation on what is covered by a
"predetermined range." See In re Zletz, 893 F.2d 319, 321,
13 USPQ2d 1320, 1322 (Fed. Cir. 1989)("[D]uring patent
prosecution when claims can be amended, ambiguities should be
recognized, scope and breadth of language explored, and
clarification imposed."). Accordingly, the claims, as presented,
do not define over the admitted prior art where coefficients of
an adaptive filter are always updated and where saturation may
occur. We next consider the examiner's rejection.
For the limitation of updating the filter coefficients only
when a prevailing value of the output signal is within a
Appeal No. 2002-1197
Application 09/131,167
- 8 -
predetermined range the examiner points to Koski, column 1,
lines 43-47; column 3, lines 20-36; column 5, lines 45-50; and
column 6, lines 1-10 (FR4). The examiner states that Koski
discloses a double talk detector "to freeze the adaptation of the
filter coefficients when the output signal of the microphone,
which is the superposition of the near-end speech signal and the
far end speech signals[,] said far end speech signal [being] the
cause of the echo, reaches a threshold that causes the detector
to freeze said coefficients" (FR2). The examiner states (FR3):
"Koski states that the filter coefficients of the adaptive filter
will only be aloud [sic, allowed] to update, i.e., they are
frozen, unless the signal from the microphone (the source) is
below a value. The signal from the microphone being the near end
users signal plus the echo signal coming from the speaker of the
communication device" (FR3).
Appellant argues: (a) Koski only freezes updating of the
coefficients upon detection of "double talk" and double talk is
not equivalent to an output signal of the microphone (Br9-10);
and (b) the filter coefficients in Koski are not updated only
when the value of the microphone output is within a predetermined
range (Br10-12). Appellant does not argue the other limitations
of the claims or the propriety of the combination with Haneda.
Arguments not raised are waived. See 37 CFR § 1.192(c)(8)(iv)
(1998) (brief must point out errors in the rejection); cf.
Appeal No. 2002-1197
Application 09/131,167
- 9 -
In re Baxter Travenol Labs., 952 F.2d 388, 391, 21 USPQ2d 1281,
1285 (Fed. Cir. 1991) ("It is not the function of this court to
examine the claims in greater detail than argued by an appellant,
looking for nonobvious distinctions over the prior art.").
We find that Koski teaches the limitation that "filter
coefficients of said adaptive filter are updated only when a
prevailing value of the output signal of said microphone is
within a predetermined range of possible output signal values"
(claim 1), as interpreted. Koski discloses that echo cancelers
are known and usually employ an adaptive digital filter where the
parameters (coefficients) are updated whenever far-end speech
occurs (col. 1, lines 21-42). Koski teaches that an improvement
is to not update the filter coefficients during double talk to
overcome the problem of distortion (col. 1, lines 43-63), which
indicates that the prior condition was to always update the
coefficients for every signal, which is considered to meet the
limitations of claims 1, 13, 17, and 25, as broadly interpreted.
Because these are the only limitations argued, we sustain the
rejection of claims 1-4, 8-28, and 32-37.
IF the claims recited updating the filter coefficients only
upon the condition that the output signal of the microphone is
within a predetermined range of possible output signal values,
where the predetermined range was less than the total range of
possible output signal values, the examiner's rejection could not
Appeal No. 2002-1197
Application 09/131,167
- 10 -
be sustained. We discuss this situation because of the dependent
claims and for the benefit of any further prosecution. Column 1,
lines 43-63, of Koski states:
Double-talking refers to the condition when the near
end subscriber (the user of the phone) and the far-end
subscriber talk simultaneously. When both parties talk
simultaneously, i.e. during double talk, the echo canceller
is no longer able to effectively block echo signals. This
is because the echo signals are included in the near-end
subscriber's signals to be transmitted, i.e. a desired
signal to be transmitted and an echo signal are
simultaneously applied to the send input. The super-
positioning of these signals causes distortion of the
adjustment of the echo canceller when it considers both the
echo signal and the desired signal to be transmitted. This
means that the replica produced by the echo canceller no
longer sufficiently cancels the current echo signal.
Accordingly, it is a current practice to provide a double
talk detector for preventing the disturbing influence of
double talk on echo canceller adjustment. This means that
the parameters of the adaptive filer performing the echo
cancellation are not updated during double talk. Echo and
double talk are problems especially in speaker phones and in
phones with hands-free equipment in which the far-end signal
from the speaker is captured by the microphone. [Emphasis
added.]
The superposition of the near-end and echo (far-end signals
coupled from the loudspeaker into the microphone) signals must
refer to superposition of signals to become the output signal of
the microphone, as noted by the examiner (e.g., FR3). The
superposition of the near-end and echo signals during double talk
creates an output signal that causes a distortion in the echo
canceller, and the need to prevent distortion suggests that the
resultant superposed signal be kept below a predetermined value
that causes distortion, i.e., that the output of the microphone
Appeal No. 2002-1197
Application 09/131,167
- 11 -
should be kept within a predetermined range that does not cause
distortion. However, Koski teaches not updating the filter
coefficients based on the condition of double talk detection, not
only updating the coefficients based upon the condition that the
output signal of the microphone is within a predetermined range
of possible output signal values, as claimed. Not updating the
filter coefficients based on double talk detection will
indirectly keep the output signal of the microphone within a
predetermined range of output signal values, but this is not the
same as updating based on the output signal value. We agree with
appellant's argument "that when double-talk does not exist, the
double-talk detector of Koski will not prevent the parameters of
the adaptive filter from being updated, regardless of the
prevailing value of the output signal of the microphone because
Koski discloses freezing the parameters of the adaptive filter
only when double-talk exists" (RBr3). Thus, the condition under
which Koski freezes the coefficients of the adaptive filter are
not based on whether the output signal is within a predetermined
range. The examiner has not presented any reasoning that since
distortion is caused when the magnitude of the superposed near-
end and echo signals is too great, due to double talk, it would
have been obvious to one skilled in the art to freeze the
coefficients based on the magnitude of the output signal rather
than based on the presence of double talk.
Appeal No. 2002-1197
Application 09/131,167
- 12 -
The examiner's findings that Koski acts "to freeze the
adaptation of the filter coefficients when the output signal of
the microphone . . . reaches a threshold that causes the detector
to freeze said coefficients" (FR2) and that "Koski states that
the filter coefficients of the adaptive filter will only be aloud
[sic, allowed] to update, i.e., they are frozen, unless the
signal from the microphone (the source) is below a value" (FR3)
are in error. As noted by appellant (Br11), Koski only freezes
adaptation of the filter coefficients when double talk exists,
not when the output signal of the microphone reaches a threshold.
The thresholds in Koski have to do with detection of double talk
and are not used for comparison with the output signal.
Claims 5, 7, 29, and 31
The examiner finds that Koski discloses updating filter
coefficients only when the output signal is at a predetermined
threshold value, referring to column 5, line 20, to column 6,
line 15 (FR6). The examiner states that "Koski discloses that
the threshold values are based in part on the distortion through
A to D conversion (see col. 4 lines 25-62)" (FR6).
Appellant argues that the combination of Koski and Haneda
does not disclose or suggest that the first predetermined
threshold value is less than or equal to an upper full-scale
value of the analog-to-digital converter (ADC) and that the
second predetermined threshold value is greater than or equal to
Appeal No. 2002-1197
Application 09/131,167
- 13 -
a lower full-scale value of the ADC, as recited in claims 5
and 29 (Br13). It is argued that column 4 of Koski does not
disclose or suggest threshold values, much less that the
threshold values are based in part on distortion through the ADC
or on an upper and lower full-scale value of an ADC (Br13-14).
The examiner states that Koski states (col. 1, lines 50-55)
that the reason for setting the threshold is to stop distortion
and therefore Koski is setting two thresholds to stop distortion
in the device (EA10).
We will not sustain the rejection. The thresholds in Koski,
columns 5 and 6, are for determining whether there is double talk
based on whether the power ratio (PR) of the far-end signal to
the near-end signal in the sub-channels are above or below the
threshold, taken in conjunction with the voice activity detector
(VAD) signal. Koski does not determine whether the output signal
is above or below a threshold, much less that the threshold
values relate to full-scale values of the ADC, as recited in
claims 5 and 29, or that at least one of the predetermined
threshold values is adjusted to account for ringing of the ADC
converter, as recited in claims 7 and 31. The fact that Koski
seeks to prevent distortion does not mean that it prevents
distortion in the way that is claimed. The rejection of
claims 5, 7, 29, and 31 is reversed.
Appeal No. 2002-1197
Application 09/131,167
- 14 -
Claim 6
Appellant argues that the combination of Koski and Haneda
does not disclose or suggest that at least one of the first and
second predetermined threshold values is adjusted to account for
analog distortion (Br14).
As discussed in connection with claim 5, the thresholds in
Koski relate to the detection of double talk and Koski does not
determine whether the output signal from the microphone is below
a first threshold or above a second threshold. Because Koski
does not have predetermined thresholds, as claimed, it does not
teach or suggest that one of the predetermined threshold values
is adjusted to account for analog distortion. Although Koski may
detect double talk to prevent distortion, this does not suggest
keeping the output signal below or above a threshold value to
account for analog distortion. The rejection of claim 6 is
reversed.
Claim 30
Appellant argues that the combination of Koski and Haneda
does not disclose or suggest that at least one of the first and
second predetermined threshold values is adjusted to account for
analog compression (Br14).
As discussed in connection with claim 5, the thresholds in
Koski relate to the detection of double talk and Koski does not
determine whether the output signal from the microphone is below
Appeal No. 2002-1197
Application 09/131,167
- 15 -
a first threshold or above a second threshold. Because Koski
does not have predetermined thresholds, as claimed, it does not
teach or suggest that one of the predetermined threshold values
is adjusted to account for analog compression. Although Koski
may detect double talk to prevent distortion, this does not
suggest keeping the output signal below or above a threshold
value to account for analog compression. The rejection of
claim 30 is reversed.
Claims 38-41
The issue is whether the combination of Koski and Haneda
discloses or suggests "means for updating filter coefficients of
said adaptive filter only when an absolute value of the output
signal of the microphone is less than a saturation constant"
(claim 38). Unlike claims 1, 13, 17, and 25, claim 38 positively
limits the updating of coefficients when the output is less than
the full range of the output signal and, as discussed in
connection with the rejection of claim 1, this is not taught by
Koski. The rejection of claims 38-41 is reversed.
Requirement under 37 CFR § 1,196(d)
The specification describes that "practical experience" has
demonstrated that adaptive filtering techniques do not work well
when the source signal (e.g., the near-end microphone signal)
becomes saturated. Specification, p. 3, first full paragraph.
Appeal No. 2002-1197
Application 09/131,167
- 16 -
It is not clear whether appellant is describing a known problem
in the art or a problem which he discovered. This fact may have
a bearing on the obviousness analysis; e.g., if the saturation
problem was known, the issue would be whether it would have been
obvious to freeze the coefficients when the source signal was of
a magnitude that would cause saturation. In any response to this
opinion, appellant is required to state whether the problems
described in this paragraph on page 3 were known to those in the
art of echo cancelation. It will be sufficient if appellant
admits, denies, or states that he is without knowledge of whether
the described problems were known in the art.
CONCLUSION
The rejection of claims 1-4, 8-28, and 32-37 is sustained.
The rejection of claims 5-7, 29-31, and 38-41 is reversed.
A new ground of rejection has been entered as to claims 1-24
pursuant to 37 CFR § 1.196(b).
In addition to affirming the Examiner’s rejection of one or
more claims, this decision contains a new ground of rejection
pursuant to 37 CFR § 1.196(b)(amended effective Dec. 1, 1997, by
final rule notice, 62 Fed. Reg. 53,131, 53,197 (Oct. 10, 1997),
1203 Off. Gaz. Pat. & Trademark Office 63, 122 (Oct. 21, 1997)).
37 CFR § 1.196(b) provides, "A new ground of rejection shall not
be considered final for purposes of judicial review."
Appeal No. 2002-1197
Application 09/131,167
- 17 -
Regarding any affirmed rejection, 37 CFR § 1.197(b)
provides:
(b) Appellant may file a single request for rehearing
within two months from the date of the original
decision . . . .
37 CFR § 1.196(b) also provides that the Appellants, WITHIN
TWO MONTHS FROM THE DATE OF THE DECISION, must exercise one of
the following two options with respect to the new ground of
rejection to avoid termination of proceedings (37 CFR § 1.197(c))
as to the rejected claims:
(1) Submit an appropriate amendment of the claims
so rejected or a showing of facts relating to the
claims so rejected, or both, and have the matter
reconsidered by the examiner, in which event the
application will be remanded to the examiner. . . .
(2) Request that the application be reheard under
§ 1.197(b) by the Board of Patent Appeals and
Interferences upon the same record. . . .
Should appellants elect to prosecute further before the
primary examiner pursuant to 37 CFR § 1.196(b)(1), in order to
preserve the right to seek review under 35 U.S.C. §§ 141 or 145
with respect to the affirmed rejection, the effective date of the
affirmance is deferred until conclusion of the prosecution before
the examiner unless, as a mere incident to the limited
prosecution, the affirmed rejection is overcome.
If appellants elect prosecution before the examiner and this
does not result in allowance of the application, abandonment or a
second appeal, this case should be returned to the Board of
Appeal No. 2002-1197
Application 09/131,167
- 18 -
Patent Appeals and Interferences for final action on the affirmed
rejection, including any timely request for rehearing thereof.
A requirement has also been entered pursuant to 37 CFR
§ 1.196(d). Any response to this opinion, whether to the Board
or to the examiner, either in response to the new ground of
rejection or by way of further prosecution, must respond to this
requirement.
Appeal No. 2002-1197
Application 09/131,167
- 19 -
No time period for taking any subsequent action in
connection with this appeal may be extended under 37 CFR
§ 1.136(a).
AFFIRMED-IN-PART - 37 CFR §§ 1.196(b) & 1.196(d)
LEE E. BARRETT )
Administrative Patent Judge )
)
)
)
) BOARD OF PATENT
JOSEPH L. DIXON ) APPEALS
Administrative Patent Judge ) AND
) INTERFERENCES
)
)
)
ANITA PELLMAN GROSS )
Administrative Patent Judge )
Appeal No. 2002-1197
Application 09/131,167
- 20 -
COATS & BENNETT, PLLC
P.O. BOX 5
RALEIGH, NC 27602