Ex Parte Bodin et al

Patent Trial and Appeal Board

Nov 26, 2012

11266559 (P.T.A.B. Nov. 26, 2012)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________

BEFORE THE PATENT TRIAL AND APPEAL BOARD
____________

Ex parte WILLIAM K. BODIN, DAVID JARAMILLO,
JERRY W. REDMAN, and DERRAL C. THORSON
____________

Appeal 2010-007040
Application 11/266,559
Technology Center 2600
____________

Before JOHN A. JEFFERY, DENISE M. POTHIER, and
JENNIFER L. McKEOWN, Administrative Patent Judges.

JEFFERY, Administrative Patent Judge.

DECISION ON APPEAL
Appellants appeal under 35 U.S.C. § 134(a) from the Examiner’s
rejection of claims 1, 25, and 26. We have jurisdiction under 35 U.S.C.
§ 6(b). We affirm.

STATEMENT OF THE CASE
Appellants’ invention dynamically adjusts prosody for
voice-rendering synthesized data. See generally Abstract. Claim 1 is
illustrative:
Appeal 2010-007040
Application 11/266,559

2
1. A computer-implemented method for voice-rendering synthesized
data comprising:
retrieving synthesized data to be voice rendered;
identifying, for the synthesized data to be voice rendered, a particular
prosody setting including determining current voice characteristics of the
user and selecting the particular prosody setting in dependence upon the
current voice characteristics of the user;
determining, in dependence upon the synthesized data to be voice
rendered and the context information for the context in which the
synthesized data is to be voice rendered, a section of the synthesized data to
be rendered including determining the context information for the context in
which the synthesized data is to be voice rendered, identifying in
dependence upon the context information a section length, and selecting a
section of the synthesized data to be rendered in dependence upon the
identified section length;
wherein identifying in dependence upon the context information a
section length further comprises:
identifying in dependence upon the context information a rendering
time; and
determining a section length to be rendered in dependence upon the
prosody settings and the rendering time;
rendering the section of the synthesized data in dependence upon the
identified particular prosody setting.

THE REJECTIONS
1. The Examiner rejected claims 1, 25, and 261 under 35 U.S.C.
§ 102(b) as anticipated by Kibre (US 6,792,407 B2; Sept. 14, 2004). Ans.
3-5.2

1 Although the Examiner omits claims 25 and 26 from the statement of the
rejection, the Examiner nonetheless includes these claims in the rejection’s
discussion. Ans. 3. Accordingly, we presume that claims 25 and 26 are
intended to be rejected and present the correct claim listing here for clarity.
Accord App. Br. 6; Reply Br. 4.
2 Throughout this opinion, we refer to (1) the Appeal Brief filed September
25, 2009 (“App. Br.”); (2) the Examiner’s Answer mailed December 30,
Appeal 2010-007040
Application 11/266,559

3
2. The Examiner newly rejected claim 26 under 35 U.S.C. § 101 as
directed to non-statutory subject matter. Ans. 6.

THE § 101 REJECTION
The Examiner finds that claim 26 recites non-statutory subject matter
since it “attempts to claim a machine readable medium” that can be
interpreted as a magnetic carrier wave according to the Specification. Ans.
6. Appellants argue that the claim recites a computer-readable recording
medium—not a carrier wave. Reply Br. 10.

ISSUE
Under § 101, has the Examiner erred in rejecting claim 26 by finding
that it can include patent-ineligible transitory media?

PRINCIPLES OF LAW
Signals are unpatentable under § 101. In re Nuijten, 500 F.3d 1346,
1355 (Fed. Cir. 2007). According to U.S. Patent & Trademark Office
(USPTO) guidelines:
A claim that covers both statutory and non-statutory
embodiments . . . embraces subject matter that is not eligible for
patent protection and therefore is directed to non-statutory
subject matter. . . . For example, a claim to a computer readable
medium that can be a compact disc or a carrier wave covers a
non-statutory embodiment and therefore should be rejected
under § 101 as being directed to non-statutory subject matter.

2009 (“Ans.”); and (3) the Reply Brief filed February 12, 2010 (“Reply
Br.”).
Appeal 2010-007040
Application 11/266,559

4
U.S. Patent & Trademark Office, Interim Examination Instructions for
Evaluating Subject Matter Eligibility Under 35 U.S.C. § 101, at 2 (Aug.
2009), available at http://www.uspto.gov/web/offices/pac/dapp/opla/2009-
08-25_interim_101_instructions.pdf.
The USPTO also notes the following:
The broadest reasonable interpretation of a claim drawn to a
computer readable medium . . . typically covers forms of non-
transitory tangible media and transitory propagating signals per
se in view of the ordinary and customary meaning of computer
readable media, particularly when the specification is silent.
When the broadest reasonable interpretation of a claim covers a
signal per se, the claim must be rejected under 35 U.S.C. § 101
as covering non-statutory subject matter.
David J. Kappos, Subject Matter Eligibility of Computer Readable Media,
1351 Off. Gaz. Pat. Office 212 (Feb. 23, 2010) (citations omitted).

ANALYSIS
We will not sustain the Examiner’s rejection of claim 26 under § 101.
As Appellants explain (Reply Br. 10), claim 26 unambiguously recites a
computer-readable recording medium—media that is expressly
distinguished from transmission media. Spec. 60:17-27. Although
exemplary recording media broadly includes magnetic disks, compact disks,
magnetic tape, and “others as will occur to those of skill in the art” (Spec.
60:22-24 (emphasis added))—a non-limiting and open-ended list of
examples—we nonetheless find that when read in context, the recited
recording medium is limited to non-transitory recordable media consistent
with the listed examples. As such, the recited recording medium does not
encompass ineligible transitory media such as signals.
Appeal 2010-007040
Application 11/266,559

5
We are therefore persuaded that the Examiner erred in rejecting claim
26 under § 101.

THE ANTICIPATION REJECTION
The Examiner finds that Kibre’s method for voice-rendering
synthesized data has every recited feature of independent claim 1 including
retrieving synthesized data to be voice-rendered3—a feature which is said to
correspond to data retrieved from “disparate data sources and types,” and is
used to provide a “final rendering” to be synthesized from text to speech.
Ans. 3, 6-7. According to the Examiner, Kibre also (1) determines a section
of the synthesized data to be rendered depending on the synthesized data and
context information, and (2) selects a section of the synthesized data
depending on an identified section length as claimed. Ans. 4-5, 7-9.
Appellants argue that Kibre does not retrieve synthesized data to be
voice-rendered, let alone determine a section of that data depending on the
synthesized data to be rendered and context information as claimed. App.
Br. 7-10; Reply Br. 6-9. Appellants add that Kibre does not identify a
section length depending upon the context information, much less select a
section of the synthesized data depending upon the identified length. App.
Br. 10-11; Reply Br. 9-10.

ISSUE
Under § 102, has the Examiner erred in rejecting claim 1 by finding
that Kibre (1) retrieves synthesized data to be voice-rendered; (2) determines

3 Although Appellants do not hyphenate this term, we nonetheless hyphenate
it for clarity and proper grammatical form.
Appeal 2010-007040
Application 11/266,559

6
a section of the synthesized data depending on the synthesized data to be
rendered and context information; (3) identifies a section length depending
on the context information; and (4) selects a section of the synthesized data
to be rendered depending on the identified length?

ANALYSIS
As noted above, the first disputed issue is whether Kibre retrieves
synthesized data to be voice-rendered. As Appellants indicate (App. Br. 8),
the Specification defines “synthesized data” as “aggregated data which has
been synthesized into data of a uniform data type.” Spec. 9:1-2. The
Specification also defines “aggregated data” as “the accumulation, in a
single location, of data of disparate types.” Spec. 8:28-31 (emphasis added).
Notably, “[d]isparate data types are data of different kind and form” which
can include differences in data structure, file format, transmission protocol,
“and other distinctions as will occur to those of skill in the art.” Spec. 6:5-8
(emphasis added). Exemplary disparate data types include MP3 files, XML
documents, email documents, “and so on as will occur to those of skill in the
art.” Spec. 6:9-14 (emphasis added). We emphasize this permissive and
open-ended language, for it underscores that disparate data types that are
synthesized into a uniform type to form synthesized data are hardly limited
to the examples in the Specification.
Turning to Kibre, concatenative synthesizer 24 produces synthesized
speech from text using recorded “snippets” (i.e., recorded sound units stored
as allophones, diphones, and/or triphones)4 for particular speakers that are

4 See also Kibre, col. 7, l. 49 – col. 8, l. 2 (noting that snippets may be stored
as samples of digitized recorded speech or parameterized).
Appeal 2010-007040
Application 11/266,559

7
stored in a database 18. Kibre, Abstract; col. 1, ll. 9-13; col. 2, ll. 23-27;
col. 4, ll. 61-65; Fig. 1. The snippet database is based on text 20 that is (1)
acquired from a text selection technique shown in Figure 2, and (2) read by a
speaker to provide the recorded snippets. Kibre, col. 5, ll. 35-42; Figs. 1-2.
As shown in Figure 2, Kibre selects text from a variety of sources, including
databases 31, digitized literature 33, electronic dictionaries 34, technical
reports 32, etc. Kibre, col. 5, l. 63 – col. 6, l. 9; Fig. 2. This text is parsed
into words and phrases that are analyzed to produce phonemes which are, in
turn, analyzed by sound analysis module 46 to produce sound units that are
stored in a data structure. Kibre, col. 6, ll. 10-64; Fig. 2. Optimal set
selection module 52 then (1) uses a “greedy” selection algorithm 54 to
identify the smallest text subset containing all unit types needed to represent
the database, and (2) provides an optimal set of words and phrases that is
displayed to a speaker. Kibre, col. 7, ll. 9-23. The speaker then reads these
optimal words and phrases while the speech is captured and digitized to
develop the database. Kibre, col. 7, ll. 25-29.
Although Kibre retrieves text data from different sources to develop
the snippet database as Appellants contend (App. Br. 8; Reply Br. 7), this
text data is nonetheless derived from disparate data types even assuming,
without deciding, that the textual sources (databases 31, digitized literature
33, electronic dictionaries 34, technical reports 32, etc.) are the same data
type (e.g., the same type of text file, such as ASCII). Notably, the Examiner
reasons that since Kibre uses not only text, but also voiceto provide the final
rendering for text-to-speech synthesis, Kibre retrieves synthesized data to be
voice-rendered as claimed. Ans. 7. This use of voice and text to personalize
text-to-speech synthesis by updating stored snippets for new speakers is a
Appeal 2010-007040
Application 11/266,559

8
key aspect of Kibre’s system. See Kibre, Abstract; col. 4, l. 66 – col. 5, l.
23; col. 7, l. 30 – col. 8, l. 23; Figs. 1-3. This update is based on iteratively
comparing the new speaker’s snippets with those in the database—snippets
that are derived from text as noted above. Kibre, col. 2, ll. 23-27; col. 5, ll.
6-23; Figs. 1, 3. These snippets, then, fully meet “synthesized data” under
Appellants’ definition since they result from aggregating disparate-type data
from different documents and speakers that is synthesized into data of a
uniform type, namely snippet data structures for storage and comparison.
And as the Examiner indicates (Ans. 7), these snippets are to be voice-
rendered, for they are retrieved and used by the concatenative synthesizer 24
to produce synthesized speech. See Kibre, col. 5, ll. 40-45; col. 7, ll. 25-32;
Fig. 3.
Nor are we persuaded of error in the Examiner’s finding that Kibre
determines a section of the synthesized data depending on the synthesized
data to be rendered and context information for the context in which the data
is to be voice-rendered as claimed. Ans. 4, 7-9. As the Examiner indicates
(Ans. 8), Appellants define context information quite broadly as “data
describing the context in which synthesized data is to be voice rendered” and
provide various non-limiting examples including “other context information
. . . as will occur to those of skill in the art.” Ans. 8 (quoting Spec. 47:19-23
(emphasis added)). Despite Appellants’ arguments to the contrary (App. Br.
8-10; Reply Br. 7-9), nothing in the claim nor this open-ended description
precludes Kibre’s using “context information” as the Examiner indicates
(Ans. 7-9), for Kibre’s system not only determines the context of word and
phrases, but also sound units which are ultimately used for voice rendering
as noted above. See Kibre, col. 6, ll. 35-64; Fig. 2 (numerals 44, 48).
Appeal 2010-007040
Application 11/266,559

9
Appellants’ contention that Kibre merely refers to the internal organization
of the data itself and not the external voice-rendering environment (App. Br.
9-10; Reply Br. 8-9) is unavailing and not commensurate with the scope of
the claim.
Lastly, we are unpersuaded of error in the Examiner’s finding that
Kibre (1) identifies a section length depending on the context information,
and (2) selects a section of the synthesized data to be rendered depending on
the identified length as claimed. Ans. 4, 10. As the Examiner indicates,
Appellants define “section length” quite broadly as “typically implemented
as a quantity of the synthesized content” and provide several non-limiting
examples including (1) a particular number of (a) synthesized data bytes; (b)
lines or paragraphs of text; or (c) chapters of content, or (2) “any other
quantity of the synthesized content . . . as will occur to those of skill in the
art.” Ans. 10 (quoting Spec. 58:3-10 (emphasis added)). The Examiner
refers to various quantities in Kibre that correspond to this open-ended
description of “section length” including syllables, demi-syllables, and pairs
of half-syllables that can constitute sound units depending on the nature of
the synthesizer. Ans. 10 (citing Kibre, col. 6, ll. 51-54). Since these units
represent a certain quantity of synthesized content, they constitute “section
lengths” under Appellants’ definition. And since sections of synthesized
data are selected for rendering depending at least partly on this length, we
find no error in the Examiner’s reliance on Kibre in this regard. Appellants’
arguments regarding Kibre’s recording context of words and phrases to
generate a snippets database (App. Br. 10-11; Reply Br. 9-10) are unavailing
and not commensurate with the scope of the claim.
Appeal 2010-007040
Application 11/266,559

10
We are therefore not persuaded that the Examiner erred in rejecting
representative claim 1, and claims 25 and 26 not separately argued with
particularity.

CONCLUSION
The Examiner erred in rejecting claim 26 under § 101, but did not err
in rejecting claims 1, 25, and 26 under § 102.

ORDER
The Examiner’s decision rejecting claims 1, 25, and 26 is affirmed.
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a)(1)(iv).

AFFIRMED

babc