Ex parte TEH et al.

Board of Patent Appeals and Interferences

Jul 30, 2001

08161798 (B.P.A.I. Jul. 30, 2001)

1 Application for patent filed December 6, 1993, entitled "Bit
Allocation Method For Digital Audio Signals," which claims the
foreign filing priority benefit under 35 U.S.C. § 119 of Japanese
Application 5-030712, filed February 19, 1993.
- 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and is not binding precedent of the Board.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. 24
UNITED STATES PATENT AND TRADEMARK OFFICE

BEFORE THE BOARD OF PATENT APPEALS
AND INTERFERENCES

Ex parte DO HUI TEH,
PEK YEW TAN,
and SUA HONG NEO

Appeal No. 1999-1793
Application 08/161,7981

ON BRIEF

Before HAIRSTON, JERRY SMITH, and BARRETT, Administrative Patent
Judges.
BARRETT, Administrative Patent Judge.
DECISION ON APPEAL
Appeal No. 1999-1793
Application 08/161,798
- 2 -
This is a decision on appeal under 35 U.S.C. § 134 from the
final rejection of claims 4-7, 9-12, 14-17, and 19-23. Claims 8, 13,
18, and 24 are allowed. Claims 1-3 are canceled.
We reverse.
BACKGROUND
The disclosed invention relates to a dynamic bit allocation
method that is adapted to the human auditory system and yet has a low
level of complexity. The complexity is reduced by allocating an
initial bit number using approximations made in a mathematical model
before the remaining bit numbers are allocated by a precise iterative
process.
Claim 5 is reproduced below.
5. A method of determining a bit allocation for a
quantization of digital audio signals having spectral and
temporal structure wherein the digital audio signals are
obtained by buffering audio signals in frames and decomposing
the digital signals into spectral components, which comprises
the steps of:
a step for dividing at least one frequency interval into a
plurality of frequency-units;
a step for obtaining a representative of each
frequency-unit obtained;
a step for counting a number of available frequency-units
by using the representative, based on psychoacoustic criteria;
a first bit allocation step for determining an initial bit
allocation for each of said available frequency-units by using
Appeal No. 1999-1793
Application 08/161,798
2 The Examiner's statement of the rejection inadvertently omits
claim 4.
- 3 -
an approximate mathematical model using said counted number of
available frequency-units; and
a second bit allocation step for determining further bit
allocation for each of said available frequency-units based on
psychoacoustic criteria.
The Examiner relies on the following prior art:
Veldhuis et al. (Veldhuis) 5,105,463 April 14, 1992
Nishiguchi et al. (Nishiguchi) 5,151,941 September 29, 1992
Claims 42-7, 9-12, 14-17, and 19-23 stand rejected under
35 U.S.C. § 103(a) as being unpatentable over Veldhuis and
Nishiguchi.
We refer to the rejection (Paper No. 5), the final rejection
(Paper No. 8) (pages referred to as "FR__") and the examiner's answer
(Paper No. 15) (pages referred to as "EA__") for a statement of the
Examiner's position, and to the appeal brief (Paper No. 14) (pages
referred to as "Br__") and the reply brief (Paper No. 16) (pages
referred to as "RBr__") for Appellants' arguments thereagainst.
OPINION
The claims are grouped to stand or fall together with
independent claim 5 (Br5).
Appeal No. 1999-1793
Application 08/161,798
3 As noted by Appellants (RBr2 n.1), there have been numerous
examiners during prosecution.
- 4 -
Appellants argue that the Examiner has improperly read
limitations into claim 5 (e.g., Br7-8; RBr5-6). To the extent this
is so, we address only the actual limitations of claim 5.
The Examiner's3 discussion of Veldhuis (FR5) is very cursory and
does not correlate the actual limitations of claim 5 to the
disclosure of Veldhuis. Therefore, we begin by making findings
regarding the contents of Veldhuis and the differences between
Veldhuis and the subject matter of claim 5.
Claim 5 recites "buffering audio signals in frames and
decomposing the digital signals into spectral components" and "a step
for dividing at least one frequency interval into a plurality of
frequency-units." A "unit" is a frequency interval (specification,
p. 8). Veldhuis does not disclose that the audio signal is buffered
in frames; however, while this limitation is considered either
inherent or obvious to one of ordinary skill in the art, it is not
argued and will not be addressed. See 37 CFR § 1.192(c)(8)(iv)
(1997) (brief must specify the errors in the rejection). Veldhuis
discloses that the audio signal x(k) is applied to an analysis filter
bank 3 which divides the signal band of 0-22.05 kHz into P=26
Appeal No. 1999-1793
Application 08/161,798
- 5 -
subbands (col. 5, lines 58-64), the characteristics of which are
shown in the table of figure 3. The bandwidths W(p) of the subbands
approximately correspond to the critical bandwidths of the human
auditory system in the respective frequency ranges (col. 6,
lines 40-46). Thus, we find Veldhuis discloses "decomposing the
digital signals into spectral components" and "a step for dividing at
least one frequency interval into a plurality of frequency-units."
The claim limitation of "at least one frequency interval" reads on
the frequency interval corresponding to the upper subbands (13#p#26)
in Veldhuis. Thus, Appellants' arguments that the lower subbands do
not use adaptive bit allocation and that their invention allows for
the possibility of assigning and allocating bits to the whole
frequency spectrum (e.g., Br10-13) are not persuasive.
Claim 5 recites "a step for obtaining a representative of each
frequency-unit obtained." The term "representative" is not expressly
defined in the specification. However, from the statement "the
variance or a representative within a defined frequency interval as
an accurate representation of the signals in the interval"
(specification, p. 5), we interpret "representative" to mean a
"characteristic representative of the signal." Veldhuis discloses a
signal buffer 6(p) for each subband and a level detector 7(p)
Appeal No. 1999-1793
Application 08/161,798
- 6 -
connected to each signal buffer 6(p) to determine for each block
stored having block number m a characteristic parameter G(p;m)
representative of the signal level of the block (col. 6,
lines 17-23). The signal level can be represented by the average
value of the amplitude or the power of the signal samples of a block,
and also by the peak value of the amplitude of the signal samples in
a block (col. 6, lines 29-33). The set of "characteristic parameters
G(p;m)" in Veldhuis is the "representative of each frequency-unit
obtained." Thus, Veldhuis discloses "a step for obtaining a
representative of each frequency-unit obtained."
Claim 5 recites "a step for counting a number of available
frequency-units by using the representative, based on psychoacoustic
criteria." For the upper group of subbands (13#p#26) in Veldhuis,
the number of quantizing bits per signal sample B(p;m) is determined
by the set of all characteristic parameters G(p;m) (col. 10,
lines 8-18). As previously noted,
the "at least one frequency interval" in the first step of claim 5
reads on the frequency interval corresponding to the upper subbands
(13#p#26) in Veldhuis, so it is proper to focus on these subbands
(frequency-units). Parameter G(p;m) is compared to the threshold
T(p) for the subband of band number p and a binary comparator signal
Appeal No. 1999-1793
Application 08/161,798
- 7 -
C(p;m) is generated having a first value C(p;m)="1" for G(p;m)$T(p)
and a second value C(p;m)="0" for G(p;m)1, so as to evenly
distribute the numbers of quantization bits saved over the
blocks (p;m) having a value of C(p;m)="1" (col. 15, line 63 to
col. 16, line 1). The value D is not disclosed to be "based on
Appeal No. 1999-1793
Application 08/161,798
- 10 -
psychoacoustic criteria." Therefore, we find that Veldhuis does not
teach "a second bit allocation step for determining further bit
allocation for each of said available frequency-units based on
psychoacoustic criteria" (emphasis added).
Based on the above analysis, we summarize that the differences
between the subject matter of claim 5 and Veldhuis are that Veldhuis
does not teach: (1) "a step for counting a number of available
frequency-units by using the representative, based on psychoacoustic
criteria" (emphasis added); (2) "a first bit allocation step for
determining an initial bit allocation for each of said available
frequency-units by using an approximate mathematical model using said
counted number of available frequency-units" (emphasis added); and
(3) "a second bit allocation step for determining further bit
allocation for each of said available frequency-units based on
psychoacoustic criteria" (emphasis added).
The Examiner has further applied Nishiguchi. The Examiner
finds that "Nishiguchi et al. teach a method for determining a bit
allocation comprising a first and second bit allocation which
incorporates psychoacoustic criteria (their initial allocation
followed by a 'correcting step' col. 2 line 64- col. 3 line 14; and
psychoacoustics col. 4)" (FR6). The Examiner concludes that it would
Appeal No. 1999-1793
Application 08/161,798
- 11 -
have been obvious to use the two-step process of Nishiguchi in the
adaptive bit allocation taught by Veldhuis "because the number of
available bits can change dynamically which results in bits being in
surplus or deficit and leading to less efficient bit allocation,
thus, requiring more bandwidth (Nishiguchi et al., col. 1)" (FR6).
We analyze the three differences noted above.
Difference (1) - counting available frequency-units
Appellants argue that the Examiner has not shown where the step
of counting the number of available frequency-units is found in
Veldhuis or Nishiguchi (RBr6). We do not find any discussion of the
counting step by the Examiner.
We agree that neither Veldhuis nor Nishiguchi discloses
counting the number of available frequency-units. In Veldhuis, the
blocks (p;m) in the allocation window having C(p;m)="1" are
"available frequency-units." However, Veldhuis does not disclose
"counting" the number of available blocks (p;n) having C(p;m)="1".
We do not find any disclosure of counting available frequency-units
in Nishiguchi. Thus, the counting step is a difference which is not
accounted for in the rejection.
Appeal No. 1999-1793
Application 08/161,798
- 12 -
Difference (2) - mathematical model using counted
number of available frequency-units
The Examiner finds that Veldhuis teaches a "mathematical model"
and points to Veldhuis at column 2, lines 54-64, showing B(p),
G(p;m), and C(p;m) (EA6). The Examiner states that a mathematical
model or algorithm is a series of mathematical steps as shown in
Veldhuis at figures 8 and 9 (EA7).
Appellants respond that a "mathematical model" is not the same
thing as a "mathematical algorithm" (RBr8-10) and that Nishiguchi
does not teach or suggest a mathematical model using the counted
number of available frequency-units (RBr10).
We agree with Appellants that a "mathematical model" is not
necessarily a "mathematical algorithm." However, it is not necessary
to explore the nuances in any detail since the Examiner has not shown
a mathematical algorithm "using said counted number of available
frequency-units," as claimed. The initial bit allocations in
Veldhuis are all predetermined. Thus, the mathematical model using
the counted number of frequency-units is a difference which is not
accounted for in the rejection.
Difference (3) - second step of bit allocation
based on psychoacoustic criteria
Appeal No. 1999-1793
Application 08/161,798
- 13 -
Appellants argue that the second step of further bit allocation
in claim 5 is different than the second step of bit correction in
Nishiguchi (Br12; RBr11-12). Appellants argue that Nishiguchi
teaches only one bit allocation step that is amorphously connected
with a correction of the allocation and that Nishiguchi does not make
obvious two separate and distinct allocation steps (RBr7-8).
The Examiner disagrees (EA7), but we do not understand the
Examiner's rationale.
The arguments are directed to the existence of a two step
allocation process, an initial allocation followed by a further bit
allocation, rather than to the limitation of the further bit
allocation being based on psychoacoustic criteria. We have found
that Veldhuis discloses a two step process. Since the limitation
about further allocation "based on psychoacoustic criteria" is not
argued, it is not addressed. See 37 CFR § 1.192(c)(8)(iv).
Appeal No. 1999-1793
Application 08/161,798
- 14 -
CONCLUSION
Because the combination of Veldhuis and Nishiguchi does not
disclose at least the limitations of "a step for counting a number of
available frequency-units by using the representative, based on
psychoacoustic criteria" (emphasis added) and "a first bit allocation
step for determining an initial bit allocation for each of said
available frequency-units by using an approximate mathematical model
using said counted number of available frequency-units" (emphasis
added), the Examiner has failed to establish a prima facie case of
obviousness. The rejection of claims 4-7, 9-12, 14-17, and 19-23 is
reversed.
REVERSED
KENNETH W. HAIRSTON )
Administrative Patent Judge )
)
)
)
) BOARD OF PATENT
JERRY SMITH ) APPEALS
Administrative Patent Judge ) AND
) INTERFERENCES
)
)
Appeal No. 1999-1793
Application 08/161,798
- 15 -
)
LEE E. BARRETT )
Administrative Patent Judge )
Appeal No. 1999-1793
Application 08/161,798
- 16 -
BEVERIDGE, DeGRANDI, WEILACHER & YOUNG, L.L.P.
1850 M Street, N.W.
Suite 800
Washington, DC 20036