Allvoice Developments U.S., LLC v. Microsoft Corp.

988 F. Supp. 2d 1248, 2013 WL 6825234, 2013 U.S. Dist. LEXIS 179946

• Court: District Court, W.D. Washington
• Decided: December 23, 2013
• Docket: Case No. C10-2102 RAJ
• Status: Published

Opinion

ORDER

RICHARD A. JONES, District Judge.

I. INTRODUCTION

This matter comes before the court on a motion for summary judgment for noninfringement by defendant Microsoft Corporation (“Microsoft”). Dkt. 188. Microsoft argues that the accused products do not contain “audio identifiers” or “link data” as construed by the court. Microsoft also argues that the accused products do not perform the “selectively disabling” step or store an audio message received from a speech recognition engine, as required by Claims 56 and 57. Finally, Microsoft argues that the accused products do not inflinge under the doctrine of equivalents and that Microsoft did not indirectly infringe the '273 patent. Plaintiff Allvoice Developments U.S., LLC (“Allvoice”) opposes the motion. Dkt. ## 191 & 195.¹

Having considered the memoranda, exhibits, oral argument and the record herein, the court GRANTS defendant’s motion for summary judgment.

II. BACKGROUND

Allvoice filed this patent infringement action alleging that Microsoft infringed several claims of the '273 Patent. The '273 Patent describes an invention that connects a speech-recognition engine with a user’s text-processing application, via an interface application program (“IAP”). The IAP allows a user to write by speaking into a microphone rather than typing on a keyboard, and the words are directly inserted into an end user’s preferred application.

In the prior art systems described in the '273 Patent, the speech-recognition engines would record the user’s speech and convert it into recognized words, which would be displayed to the user on the computer screen in a dictation window. If the user saw a misrecognized word in the dictation window, the user could play back the voice recording and make corrections in the dictation window, and the speech-recognition engine would then incorporate the corrections to optimize its accuracy over time. In order to transfer the text from the dictation window to a text-processing application, such as Word, the user would have to cut and paste the text from the dictation window into a Word document manually.

The '273 Patent sets out to improve what it identifies as a major disadvantage of the prior art systems: cutting and pasting the text from the dictation window eliminated the connection between the speech recording and the text, meaning that if the user made corrections in the Word document (rather than in the dictation window), the speech-recognition engine could not incorporate those corrections to improve its speech-recognition models. As a solution to this problem, the '273 Patent adds an IAP to “control the flow of text into the text processing application, to control the flow of updating information from the text processing application to the speech recognition application and for maintaining links between the text and the audio data.” '273 Patent at 5:7-14. The IAP creates and stores “link data,” which allows the user to play back the stored audio data corresponding to the words in the text processing application, and any corrections made in the text processing application can be incorporated by the speech-recognition engine to update and improve accuracy.

Allvoice contends that Microsoft has made, used and sold the Windows XP, Windows Vista, and Windows 7 operating systems, which include software referred to as SAPI Server and Text Services Framework (the “TSF”) that violate the '273 Patent.

III. ANALYSIS

1. Legal Standard

In patent cases, the court follows the law of the regional circuit for procedural questions not unique to patent law. See Koninklijke Philips Elecs. N.V. v. Cardiac Science Operating, 590 F.3d 1326, 1332 (Fed.Cir.2010) (applying Ninth Circuit law to vacate sua sponte grant of summary judgment).² Summary judgment is appropriate if there is no genuine dispute as to any material fact and the moving party is entitled to judgment as a matter of law. Fed.R.Civ.P. 56(a). The moving party bears the initial burden of demonstrating the absence of a genuine issue of material fact. Celotex Corp. v. Catrett, 477 U.S. 317, 323, 106 S.Ct. 2548, 91 L.Ed.2d 265 (1986). Where the moving party also bears the burden of persuasion at trial, the moving party must show that the evidence is so powerful that no reasonable jury would be free to disbelieve it. Shakur v. Schriro, 514 F.3d 878, 890 (9th Cir.2008).

On an issue where the nonmoving party will bear the burden of proof at trial, the moving party can prevail by either (1) producing evidence negating an essential element of the nonmoving party’s claim or defense or (2) showing that the nonmoving party does not have enough evidence of an essential element to carry its burden of persuasion at trial. Nissan Fire & Marine Ins. Co. Ltd. v. Fritz Cos., Inc., 210 F.3d 1099, 1102, 1106 (9th Cir.2000); see Celotex Corp., 477 U.S. at 325, 106 S.Ct. 2548 (“the burden on the moving party may be discharged by ‘showing’— that is, pointing out to the district court— that there is an absence of evidence to support the nonmoving party’s case.”). However, a “moving party may not require the nonmoving party to produce evidence supporting its claim or defense simply by saying that the nonmoving party has no such evidence.” Nissan Fire, 210 F.3d at 1105. “In a typical case, in order to carry its initial burden of production by pointing to the absence of evidence to support the nonmoving party’s claim or defense, the moving party will have made reasonable efforts, using the normal tools of discovery, to discover whether the nonmoving party has enough evidence to carry its burden of persuasion at trial.” Id. (citing Clark v. Coats & Clark, Inc., 929 F.2d 604, 608 (11th Cir.1991) for the proposition that “a moving party must ‘point to materials on file which demonstrate that a party will not be able to meet that burden.’ ”).

If the moving party meets the initial burden, the opposing party must set forth specific facts showing that there is a genuine issue of fact for trial in order to defeat the motion. Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 250, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986). The court must view the evidence in the light most favorable to the nonmoving party and draw all reasonable inferences in that party’s favor. Reeves v. Sanderson Plumbing Prods., 530 U.S. 133, 150-51, 120 S.Ct. 2097, 147 L.Ed.2d 105 (2000).

2. Microsoft’s Initial Burden of Production

Allvoice argues that Microsoft has not met its initial burden of production because it “offers no evidence” that the accused products lack the “audio identifier,” the “link data” or the “audio messages” elements of the claims. Dkt. ## 191 & 195 at 10, 13, 16. However, Microsoft is not required to produce evidence to negate an element of Allvoice’s claims. Instead, it has directed the court to Allvoice’s infringement contentions, this court’s claim construction, and the '273 Patent to demonstrate that its accused products do not meet the claim limitations. The court believes this is sufficient to meet the initial burden of production under Celotex, particularly where Allvoice has not requested additional time to complete discovery.

3. Claim Limitations

“Summary judgment on the issue of infringement is proper when no reasonable jury could find that every limitation recited in a properly construed claim either is or is not found in the accused device either literally or under the doctrine of equivalents.” U.S. Philips Corp. v. Iwasaki Elec. Co. Ltd., 505 F.3d 1371, 1374-75 (Fed.Cir.2007).

Microsoft argues that it is entitled to summary judgment because (1) the accused products do not have the claimed “audio identifiers” that all asserted claims require, (2) the accused products do not have the claimed “link data” that all asserted claims require, and (3) the accused products do not perform the claimed “selectively disabling” step or store audio messages received from a “speech recognition engine” that claims 56 and 57 require. Dkt. # 188 at 2.

i. “Audio Identifier”

The court previously construed “audio identifier” to mean “identifiers that indicate, for each recognized word, (1) the file containing the corresponding audio component and (2) the position of the corresponding audio component within that file.” Dkt. # 166 at 15.

Microsoft argues that Allvoice has not identified any “file” in the accused products, and that the Random Access Memory (“RAM”) referenced in Allvoice’s infringement contentions is a substantially different structure than a “file.” Dkt. # 188 at 7. Allvoice argues that nothing precludes “memory” from being a “file” and that its infringement contentions include storing audio data in “a paging file or other file.” Dkt. ## 191 & 195 at 11. The parties’ disagreement ultimately boils down to the scope and meaning of the term “file.”

The court requested additional briefing with respect to the parties’ construction of the term “file” and to identify parts of the specification that supported their construction. Dkt. #216. Allvoice relies on the IEEE Standard Dictionary of Electrical and Electronics Terms 504 (6th ed.1996) and proposes the following construction: “a collection of related records treated as a unit.”³ Dkt. # 224 at 1; Dkt. ## 191-2 & 195-1 (Sonnier Decl.) ¶ 21 n. 1. Microsoft argues that Webster’s New World Dictionary of Computer Terms (6th ed.1997) definition of “file” is the most consistent with the '273 patent: “a document or other collection of information stored on a disk and identified as a unit by a unique name.”⁴ Dkt. ##204 & 205 at 3; Dkt. # 206 at 12 (Byer Decl. iso Reply), Ex. C (emphasis omitted); Dkt. # 225 at 2-7. It appears that Microsoft does not dispute that the meaning of “file” includes a collection of related records treated as a unit. Rather, the disagreement is about whether the collection of records must be “stored on a disk” and whether it must have a “unique name.”

Judges may “rely on dictionary definitions when construing claims, ‘so long as the dictionary definition does not contradict any definition found in or ascertained by a reading of the patent documents.’ ” Meyer Intellectual Props. Ltd. v. Bodum, Inc., 690 F.3d 1354, 1368 (Fed.Cir.2012) (quoting Phillips v. AWH Corp., 415 F.3d 1303, 1322 (Fed.Cir.2005) (en banc)). Indeed, “heavy reliance on the dictionary divorced from the intrinsic evidence risks transforming the meaning of the claim term to the artisan into the meaning of the term in the abstract, out of its particular context, which is the specification.” Phillips, 415 F.3d at 1321. “The risk of systematic overbreadth is greatly reduced if the court instead focuses at the outset on how the patentee used the claim term in the claims, specification, and prosecution history, rather than starting with a broad definition and whittling it down.” Id. “Indeed, the specification ‘is the single best guide to the meaning of a disputed term’ and it ‘acts as a dictionary when it expressly defines terms used in the claims or when it defines terms by implication.’” Meyer Intellectual Props., 690 F.3d at 1368.

During oral argument, Allvoice again argued that there is nothing in the specification that requires “file” to mean stored on disk storage. However, every time the specification refers to a “file,” it explicitly provides that the file is stored on one of the directories within the disk storage:

• “The system is also provided with non-volatile storage in the form of disk storage 15. Within the disk storage 15 two directories are provided. A temporary directory used by the speech recognition engine 11 for the storage of run time files which contain the speech recognition output data. A user’s directory is also provided for the storage of document files by the text processor application 13 and associated link data formed by the speech recognition interface 12.” '273 Patent at 6:1-8 (Dkt. # 107-2); see also at 13:8-12.

• “In the specific embodiment the audio data is stored in one or more files in the temporary directory of the disk storage 15 since the storage audio data requires a great deal of storage capacity and it is impractical to hold audio data of any length in the volatile memory 20.” Id. at 6:29-34.

• “The output data is then passed to files which are opened in the temporary directory of the disk storage 15. The audio data for each period of dictation is stored in a single file.” Id. at 6:44-47.

• “Also in temporary directory on the disc storage 15, two files are stored by the speech recognition engine application 11 which includes information illustrated in tabular form in FIG. 3. For each period of dictation an audio data file and a pair of information files are generated containing the information illustrated in FIG. 3.” Id. at 6:48-53.

• “The link data 25 also includes information identifying where the audio data can be found in the files in the temporary directory of the disk storage 15. This information is provided in the tag field. The tag field will not only include the identified tag identifying the position of the audio component for a word within a file, it will also include an identification of which file contains the audio component.” Id. at 7:12-18.

• “In step S19, in order to save the link data and audio data, the document and link data, by default, is saved in the user’s directory and a copy of the speech recogniser run time created files is made in the user’s directory.” Id. at 8:40-44.

• “In step S30 the dictation is started and in step S31 the speech recognition engine application 11 outputs speech recognition data 24 and stores the data in run time files in a temporary directory of the disk storage 15. Also, the audio data is stored in parallel as a ran time file in the temporary directory in step S32.” Id. at 8:52-57.

• “The audio component is then retrieved from the speech recognition ran time created files in the temporary directory view the speech recognition engine application 11 in step S56 and in step S57 the audio component is played back via the speech recognition engine application 11.” Id. at 9:41-46.

• “In step S76 the audio component is retrieved from the speech recognition run time created files in the temporary directory via the speech recognition engine application 11 for playback of the audio component via the speech recognition engine application 11 in step S77.” Id. at 10:21-25.

• “Also, since not only audio data but also the link data is stored in nonvolatile storage such as the disk storage 15, the user is able to reopen a document and play back the corresponding audio data. This enables a user to dictate a document and store it without correction thereby allowing correction at a later date, i.e. delaying the correction.” Id. at 10:64-11:2.

• “If there are no audio messages associated with a document the process proceeds to step S9 where the document and link data is read and the speech recognition run time created files are copied from the user’s directory to the temporary directory and the system proceeds as described with regard to FIG 5.” Id. at 11:15-20.

Allvoice has not directed the court to a single citation in the specification that would suggest that a “file” could also mean volatile storage or memory. The court recognizes that the specification provides for the ability for the volatile memory to store speech recognition output data and audio data temporarily.⁵ '273 Patent at 5:44-62, 6:16-19, 6:41-44 (Dkt. #107-2). However, the term “file” is only used in conjunction with disk storage in the specification.

With respect to whether the collection of data must have a “unique name,” Microsoft only directs the court to several definitions and to one part of the specification, which provides: “The link data 25 also includes information identifying where the audio data can be found in the files in the temporary directory of the disk storage 15. This information is provided in the tag field. The tag field will not only include the identified tag identifying the position of the audio component for a word within a file, it will also include an identification of which file contains the audio component.” '273 Patent at 7:12-18 (Dkt. #107-2). However, nothing in this section requires a unique name.

Accordingly, the court construes the term “file” to mean “a collection of related records treated as a unit stored on a disk.”

Sonnier opines that the “audio identifier” element is satisfied by the ISpRecoResult structures, which identify the file where the audio corresponding to a particular recognized word is stored and include an integer that provides the offset within that file where the audio for the particular word is stored. Dkt. ## 191-2 & 195-1 (Sonnier Deck) ¶ 21. Mr. Sonnier has testified that the accused audio data (ISpRecoResult) is stored in “allocated memory” which means “RAM memory.” Dkt. # 207 (Ex. A to Byer Deck iso Reply, Sonnier Depo. at 94:17-25). However, the court has rejected the definition of “file” offered and relied upon by Mr. Sonnier because it is divorced from the intrinsic evidence. Accordingly, the court finds that Allvoice’s infringement contentions based on audio data being stored in “memory” must be dismissed since audio data in memory or RAM is not stored in a “file.”

Nevertheless, Allvoice argues that its infringement contentions also allege that the computer stores the audio data in a paging file or other file. Dkt. ## 191 & 195 at 11; Dkt. # 191-10 (Attach. 2 to Ex. B to Shih Deck at 25, 36, 57). In its opening papers, Microsoft did not move for summary judgment with respect to All-voice’s infringement contentions alleging that the computer stores the audio data in a paging file or other file. Dkt. # 188. Since Microsoft did not provide any argument or analysis in its moving papers regarding the infringement contentions that allege “a paging file or other file,” the court finds that it has not met its initial burden of production. Accordingly, the court declines to address the arguments raised for the first time in reply and during oral argument regarding whether an audio identifier identifies a paging file or whether the accused “reference” is not an audio identifier.⁶

ii. “Link Data”

The court previously construed “Link Data” to mean “[l]ink data, which is stored in the interface application memory, includes the character positions of recognized words [or characters] in the text processing application and the corresponding audio identifiers for those words [or characters].” Dkt. # 166 at 16.

Microsoft argues that the court already rejected Allvoice’s motion for leave to amend its infringement theory to assert that the TSF, which Allvoice’s expert includes within the term SAPI Server, derives or obtains and stores the alleged character positions. Dkt. # 188 at 11. While the court denied Allvoice’s motion for leave to amend its infringement contentions based on lack of diligence, the operative infringement contentions provide:

For the purpose of the infringement analysis in Attachment 2, the foregoing versions of the SAPI are collectively referred to as part of the ‘SAPI Server,’ in part, because they all appear to operate generally in the same manner, at least with regard to the infringement analysis. SAPI Server as used herein includes the Text Services Framework, sapisvr.ee, related client APIs, and spsreng.dll.

Dkt. # 191-10 (Ex. B to Shih Decl. at 3). Accordingly, there is nothing new regarding Allvoice’s argument regarding TSF.⁷

During oral'argument, Microsoft argued that it challenged Mr. Sonnier’s outcome-driven and unscientific methodology as inadmissible, unreliable, and not competent. Specifically, Microsoft argued during oral argument that there are three reasons why Mr. Sonnier’s testimony is inadmissible and not competent: (1) Mr. Sonnier’s definition of SAPI Server contradicts the court’s claim construction order because it includes “spsreng.dll” — the speech recognition engine; (2) Mr. Sonnier’s outcome-driven definition of SAPI Server is inadmissible under Daubert and Rule 702; and (3) the definition of SAPI Server that All-voice relies on for Windows XP— CTFMON.EXE and the Language Bar— were not disclosed in the infringement contentions.

With respect to the first argument, Mr. Sonnier confirms that his definition of the term “SAPI Server” is the same as disclosed in the infringement contentions: SAPI Server refers to the TSF, SAPISVR.EXE,⁸ related client APIs, and spsreng.dll. Dkt. ## 191-2 & 195-1 (Sonnier Decl.) ¶ 14; Dkt. # 191-10 (Ex. B to Shih Decl. at 3). During oral argument plaintiff did not dispute that inclusion of the speech recognition engine, or spsreng. dll, within the meaning of “SAPI Server” contradicts the court’s Markman order. Rather, plaintiff argued that Mr. Sonnier was careful in his report and only referred to “SAPI Server” a few times, and, instead, referred to the individual components of what his definition of SAPI Server included, such as TSF and SAPISVR.EXE. The court agrees. See Dkt. ## 191-2 & 195-1 (Sonnier Decl.) ¶¶ 10, 15-17, 21. Nevertheless, to the extent that Mr. Sonnier uses the term “SAPI Server” to include spsreng.dll, the court strikes that part of the' definition.

With respect to the second argument, Microsoft argues that Allvoice cannot “cobble together” separate programs with different purposes and functions to satisfy the interface application memory limitation, and that Mr. Sonnier’s “litigation-outcome-driven methodology is biased and unsound.” Dkt. ## 204 & 205 at 7-8. During oral argument, Microsoft argued that Mr. Sonnier came to a conclusion first, and then did the research after the fact to support it, citing Claar v. Burlington, 29 F.3d 499 (9th Cir.1994). The court disagrees.

Under Rule 702, an expert may testify in the form of an opinion if (1) the expert’s scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue, (2) the testimony is based on sufficient facts or data, (3) the testimony is the product of reliable principles and methods, and (4) the expert has reliably applied the principles and methods to the facts of the case. Fed. Rule Evid. 702. The court must scrutinize the reasoning and methodology underlying the expert’s affidavit. Claar, 29 F.3d at 501. Before admitting expert affidavits, a district court is affirmatively required to find that the experts’ conclusions were based on scientific knowledge, which means that the experts arrived at their conclusions using scientific ■ methods and procedures, not subjective beliefs or unsupported speculation. Id. at 502.

Mr. Sonnier has significant technical and specialized knowledge that would assist a trier of fact to understand the evidence. Mr. Sonnier has undergraduate degrees in Physics and Computer Science, and he has taken graduate courses in computer sciences. Dkt. ##191-2 & 195-1 (Sonnier Decl.) ¶ 4. Mr. Sonnier has been an information technology consultant for computers, software and networks for over 25 years. Id. ¶ 3. He has studied and obtained extensive knowledge and understanding of numerous speech recognition products and the technical documents regarding application programmer interfaces (“API”) and other aspects of speech recognition products that were available in 1996 and earlier. Id. Through his work as an expert in patent cases involving speech recognition technology, he has analyzed speech recognition products and APIs that were promoted during that period by Dragon Systems, Inc., IBM Corp., Microsoft Corp., NCC, Inc., and Kurzweil Artificial Intelligence, Inc. Id. Mr. Sonnier is the president and founder of a company that provides information technology services to clients, including software development, hardware integration, and network design for the past twelve years. Id. ¶ 4. He has over twenty years of experience writing and analyzing software in various programming languages, and over eighteen years of experience writing and analyzing source code for Microsoft Windows based applications. Id. Mr. Sonnier has developed software and related systems for various applications, and has developed extensive knowledge regarding not only software but also the interrelation between software applications, hardware, related data structures, and the applicable operating system. Id.

Mr. Sonnier’s testimony is also based on sufficient facts and data. Mr. Sonnier’s opinions are based on his education, experience, and background, as well as his “review of relevant pleadings, orders, and discovery responses in this ease, and document productions and source code for the Accused Instrumentalities .... ” Id. ¶ 8. These documents included the '273 patent, its prosecution history, the interference proceeding record, the referenced prior art, Microsoft’s MSDN website and CDROMs, Microsoft’s SDK and DDK for the various Windows releases, and background information from various sources. Id. Mr. Sonnier also reviewed the produced source code on multiple occasions. Id.

Finally, Mr. Sonnier’s testimony is the product of reliable principles and methods and he has applied those principles to the facts of this case. Mr. Sonnier has testified that his definition of SAPI Server is the '273 Patent’s interface application program, that the SAPI Server makes its appearance in Windows products in slightly different ways based on the developers’ choice to reorganize some of the functionality, and that he found the functionality in the Microsoft products that implement those recited functions. Dkt. # 207 (Ex. A to Byer Decl. iso Reply, Sonnier Depo. at 74:21-76:6). Mr. Sonnier explained that he used the original engineering documents that he received during discovery, which talk about various code names and other references to SAPI Server, to develop the notion that what was generally referred to as SAPI Server was a primary element of the functionality that he could see in the products that were implementing the functions recited in the '273 Patent. Id. (at 76:11-19). Once he received the code, he went through and detailed out “exactly what all went into this what was just referred to in engineering documents as SAPI Server” and he included those pieces as described in the source code that performed the recited function. Id. (at 76:20-25).

Accordingly, the court finds that the requirements of Rule 702 and Daubert are met. Microsoft’s arguments go to the weight of Mr. Sonnier’s evidence and credibility determinations, not admissibility.

With respect to the third argument, Mr. Sonnier confirms that he uses the term SAPI Server to refer to the TSF, SAPISVR.EXE, related client APIs, and spsreng.dll, and for Windows XP, the term also includes CTFMON.exe and the Language Bar. Dkt. # 191-2 & 195-1 (Sonnier Decl.) ¶ 14. Since the infringement contentions did not disclose the programs CTFMON.exe and the Language Bar, All-voice is precluded from relying on this aspect of the definition at this late stage. However, Mr. Sonnier has testified that SAPISVR.EXE and the corresponding client APIs are also programs in Windows XP that create the link data. Dkt. # 207 (Ex. A to Byer Decl. iso Reply, Sonnier Depo. at 206:14-207:10). Accordingly, the court declines Microsoft’s invitation to dismiss all infringement claims against Windows XP based on the preclusion of CTFMON.exe and the Language Bar.

Microsoft also argues that Allvoice raised the TSF property store for the first time in its motion for leave to amend, which the court denied, and that the operative infringement contentions do not identify the TSF property store.⁹ Dkt. ## 204 & 205 at 8; see also Dkt. # 188 at 11. During oral argument Allvoice conceded that the sole element Allvoice accuses of being “link data” is data in the property store of the TSF. In response to the court’s question regarding where Allvoiee identifies the property store in its infringement contentions, Allvoice directed the court to one passage in the infringement contentions and documents cited in the chart that were not part of the record.¹⁰

According to Allvoice, the relevant portion of the infringement contention is the following:

SAPI Server receives recognition data and stores it in memory. That data includes the recognized words and corresponding references to audio stream offsets. SAPI Server derives or obtains the character offset where the recognized words are inserted into an end user application. It forms “link data” by associating a reference to the character positions of the recognized words with a reference to the audio stream offset for their corresponding audio data (i.e., the sound of the word spoken). SAPI Server Stores the link data in its memory.

See MS-ALLVOICE00174110-24; MSALLVOICE00068057-76; MS-ALLVOICE000107348-50; MS-ALLVOICE00103940-41.

Attach. 3 at 15 (emphasis in original); see Dkt. # 191-10 (Attach. 2 to Ex. B to Shih Decl. at 26). Nothing in this passage of the infringement contentions, or in any other, refers to the property store of the TSF. Nevertheless, Allvoice argued that the documents cited were very clear regarding what Allvoice meant, and that the documents cited make clear that the information was stored as a property.¹¹

One of the documents cited during oral argument provides:

Properties

Cicero[¹²] provides properties that associate metadata with a range of text. The Speech TIP uses a property to associate (an abstract pointer to) the recognition data with each dictated word. When the user wishes to correct a range of text, the Speech TIP retrieves the properties in order to recover the recognition data.

Attach. 3 at 16 (emphasis in original); Dkt. #234-5 (Ex. J to Supp. Shih Decl. at 68061). Allvoice also argued that an insertion dataflow diagram purported to show that after text is inserted, the arrow following that refers to the property store, which is used to set certain information. Attach. 3 at 17; see Dkt. # 234-5 (Ex. J to Supp. Shih Decl. at 68070). Another document discusses implementing corrections by “retaining the recognition results of all dictated text, and storing an index to that recognition result as a property to the inserted text.” Attach. 3 at 18; see Dkt. # 234-6 (Ex. K to Supp. Shih Decl. at 174123).

Contrary to Allvoice’s argument, it is not clear from the documents cited during oral argument, or any other document cited in the infringement contentions, that Allvoice referred to the property store in the TSF. Rather, a careful review of those documents cited indicate that they refer to the Speech TIP, which appears to be a different program. Dkt. ## 234-5 & 234-6 (Exs. J & K to Supp. Shih Decl. at 68061-62, 70, 174123). The court finds that Allvoice did not disclose the property store of the TSF in its infringement contentions, and therefore cannot now rely upon it to avoid summary judgment.

Since Allvoice cannot demonstrate that the accused products contain “link data,” the court GRANTS summary judgment as to all claims on these grounds.

iii. “Audio Messages”

Microsoft argues that Claims 56 and 57 should be dismissed because accused products do not perform the “selectively disabling” step or store an audio message received from a speech recognition engine, as required by Claims 56 and 57. Dkt. # 188 at 12-13. With respect to the latter, Allvoice argues that Microsoft’s construction ignores the structure of claim 56 that combines two separate functionalities, the first of which requires audio data to come from the speech recognition engine, but the second of which is silent about the source of the audio data. Dkt. ## 191 & 195 at 17. Accordingly, the court must construe Claim 56 to determine whether the audio data for audio messages must be received from the speech recognition engine.

Claim 56 provides in relevant part:

A data processing method comprising:

Receiving recognition data and corresponding audio data from a speech recognition engine, said recognition data including recognised characters and audio identifiers identifying audio components corresponding to text components in the recognized text;

Selectively disabling one of the importation of recognised characters into the processor and the recognition of speech by said speech recognition engine for a period of time;

Storing the audio data for the period of time as an audio message associated with the file; ...

'273 Patent at 27:50-55, 28:9-14 (Dkt. # 107-2) (emphasis added).

A plain reading of the claim indicates that “a speech recognition engine” in the first paragraph is the antecedent basis¹³ of the term “said speech recognition engine” in the second paragraph above. Similarly, a plain reading of the claim also indicates that “corresponding audio data” in the first paragraph is the antecedent basis for the term “the audio data” in the third paragraph above. Accordingly, “the audio data” in the third paragraph above also comes “from a speech recognition engine.” Thus, to survive dismissal, there must be evidence that the accused products store audio data received from the speech recognition engine. However, Allvoice has not directed the court to any evidence that the accused products store audio data received from the speech recognition engine. Rather, Mr. Sonnier testified that in the accused products, the audio data for audio messages is never received from the speech recognition engine. Dkt. # 207 (Ex. A to Byer Decl. iso Reply, Sonnier Depo. at 39:18-22, 43:12-23, 51:2-6).

Accordingly, the court finds that the accused products lack the limitation that the audio data be received from the speech recognition engine, as required by Claims 56 and 57.

4. Doctrine of Equivalents

Microsoft argues that the accused products do not infringe under the doctrine of equivalents because the operative infringement contentions are devoid of any opinion regarding infringement by Microsoft’s products under the doctrine of equivalents. Dkt. # 188 at 14. Allvoice argues that arguments about the doctrine of equivalents are premature, and that it should be allowed to amend its infringement contentions. Dkt. ## 191 & 195 at 19-21. However, on December 27, 2012, and September 27, 2013, the court denied Allvoice’s motions for leave to amend its infringement contentions to, among other things, include the doctrine of equivalents for “link data” and “audio identifiers” because it was not diligent in seeking amendment. Dkt. ## 183, 212.

A finding of infringement under the doctrine of equivalents requires a showing that the difference between the claimed invention and the accused product or method was insubstantial or that the accused product or method performs the substantially same function in substantially the same way with substantially the same result as each claim limitation of the patented product or method. AquaTex Indus., Inc. v. Techniche Solutions, 479 F.3d 1320, 1326 (Fed.Cir.2007); see also Local Patent Rule 120(e) (requiring that infringement contentions contain a statement regarding whether each element of each asserted claim is claimed to be literally present and/or present under the doctrine of equivalents in the accused device).

The court finds that Allvoice has not made the requisite showing to create an issue of material fact with respect to infringement under the doctrine of equivalents.

5. Indirect Infringement

Microsoft argues that Allvoice’s claim that Microsoft indirectly infringed¹⁴ the claims fails because there is no direct infringement. Dkt. # 188 at 15; Akamai Techs., Inc. v. Limelight Networks, Inc., 692 F.3d 1301, 1308 (Fed.Cir.2012) (en banc) (“there can be no indirect infringement without direct infringement.”). The court has already found that the accused products lack the requisite “link data” for all claims and the limitation that the audio data be received from the speech recognition engine, as required by Claims 56 and 57. Accordingly, the court finds that there is no indirect infringement.

Microsoft also argues that both inducement and contributory infringement require proof of the indirect infringer’s culpable purpose, which Microsoft lacked. Dkt. # 188 at 15. Allvoice argues that it has presented sufficient evidence with respect to Microsoft’s culpable mental state. Dkt. ## 191 & 195 at 22.

“A finding of inducement requires both knowledge of the existence of the patent and ‘knowledge that the induced acts constitute patent infringement.’ ” Commit USA LLC v. Cisco Sys., Inc., 720 F.3d 1361, 1367 (Fed.Cir.2013) (quoting Global-Tech Appliances, Inc. v. SEB S.A., — U.S. —, 131 S.Ct. 2060, 2068, 179 L.Ed.2d 1167 (2011)). The knowledge requirement “may be satisfied by showing actual knowledge or willful blindness.” Id. at 1366. “[A] willfully blind defendant is one who takes deliberate actions to avoid confirming a high probability of wrongdoing and who can almost be said to have actually known the critical facts.” Globalr-Tech, 131 S.Ct. at 2070-71. “By contrast, a reckless defendant is one who merely knows of a substantial and unjustified risk of such wrongdoing, and a negligent defendant is one who should have known of a similar risk but, in fact, did not[.]” Id. at 2070 (citations omitted).

In 1997, John Mitchell, one of the named inventors of the '273 Patent, met with Xuedong Huang, Microsoft’s manager responsible for the development of its speech recognition software. Dkt. # 191-7 (Mitchell Decl.) ¶¶ 1, 4. At that time, Microsoft had not yet introduced speech recognition software as part of its Windows operating system. Id. ¶4. During that meeting, Mr. Mitchell explained that All-voice was interested in entering into a strategic relationship with Microsoft, and that it had patent applications pending. Id. Mr. Huang asked to be notified when Allvoice received the patent grants. Id. Mr. Mitchell informed Mr. Huang when Allvoice was awarded the patents, and subsequently, in 2001, Mr. Mitchell sent a copy of a patent infringement complaint that Allvoice filed against Dragon Systems. Id. ¶¶ 5-6.

Mr. Huang testified that he never reviewed the patents identified by John Mitchell, had no idea what technology was covered, and was not interested in those types of patents. Dkt. # 191-9 (Ex. A to Shih Deck, Huang Depo. at 83:16-84:16, 85:8-86:4, 87:17-23). Additionally, Microsoft’s practice was that Microsoft would not review anyone’s patent, and instead, focused on creating innovative technologies. Id. (Huang Depo. at 104:18-21, 105:22-107:5,108:6-14,108:24-109:7).

While this evidence is sufficient to demonstrate knowledge of the patent, the court finds that it does not create a genuine issue of material fact regarding whether Microsoft had knowledge (actual or willful blindness) that the induced acts constituted patent infringement.

Accordingly, the court dismisses All-voice’s induced infringement claim.

IV. CONCLUSION

For all the foregoing reasons, the court GRANTS defendant’s motion for summary judgment of non-infringement. Dkt. # 188. The court also GRANTS plaintiffs motion to seal exhibits J and K to the Supplemental Shih Declaration. Dkt. #235. The Clerk is directed to enter judgment in favor of Microsoft and against Allvoice. The Clerk is also directed to attach the following documents as attachments to this order: (1) Microsoft’s Presentation for Summary Judgment Hearing, (2) Microsoft’s Claim Construction Presentation for “File” and (3) Allvoice’s Presentation for Summary Judgment Hearing. The court found these visual aids helpful in understanding the arguments made by counsel during the hearing, and believes they should be available in the record.

1 . A number of filings have been filed under seal. In accordance with this District’s Local Rules, the parties have also filed redacted versions of all documents filed under seal. This court has attempted to cite to both the sealed and redacted documents where applicable.

2 . The summary judgment standard is a procedural issue not unique to patent law. Accordingly, the court applies Ninth Circuit law to the summary judgment standard.

3 . In its opening supplemental brief, Allvoice did not provide any citations to the specification that supported its construction. Dkt. # 224 at 1-4.

4 . Microsoft also argued that the IEEE Standard Dictionary provides other definitions of “file” that could be used: (a) "One named collection of data” and (b) “A set of related records usually treated as a named unit of storage.” Dkt. ## 204 & 205 at 3; Dkt. • 206 at 8 (Byer Decl. iso Reply), Ex. B (emphasis omitted).

5 . Indeed, it is for this reason that the court previously rejected Microsoft’s argument that the specification required that audio data be stored permanently in non-volatile disk storage. Dkt. # 166 at 23.

6 . The court notes that during oral argument, Allvoice conceded that the hibernation file is not at issue in this case.

7 . The same is not true for Mr. Sonnier’s attempt to include within the meaning of SAPI Server CTFMON.exe and the Language Bar for Windows XP, which the court addresses below.

8 . SAPISVR.EXE is a shared recognition service. Dkt. ## 191-2 & 195-1 (Sonnier Decl.) ¶ 13.

9 . This argument is one aspect of Microsoft's argument in its opening brief that Allvoice’s infringement contentions do not identify a data structure stored in the interface application memory that forms, stores, or updates the required character positions, much less store the character positions together with the audio identifiers in a data structure in an interface application memory as required by the construction of link data. Dkt. #188 at 11.

10 . The court has attached the parties’ visual presentations to this order as Attachments 1, 2 and 3. The court has referenced Allvoice's presentation as Attachment 3 in this order. The court has referred to the presentation solely for the purpose of clarification of All-voice’s arguments. Allvoice has also included the documents cited in its infringement contentions in a supplemental filing after the court advised that it would not accept additional evidence. Nevertheless, for purposes of this order, the court has considered these documents.

11 . Allvoice referred the court to documents bate-stamped MS-ALLVOICE00068057, 061, 070, and 174123 that were cited on page 26 of its infringement contentions. Attach. 3 at 15-18; Dkt. ## 234-5 & 234-6 (Exs. J & K to Supp. Shih Decl.).

12 . Cicero is also known as TSF. Dkt. # 234-5 (Ex. J to Supp. Shih Decl. at 68057).

13 . See Process Control Corp. v. HydReclaim Corp., 190 F.3d 1350, 1356 (Fed.Cir.1999).

14 . All voice alleges that Microsoft "indirectly infringe[s] through induced or contributory infringement of the Asserted Claims of the '273 patent through the import, offer for sale, sale and use of the Accused Instrumentalities by third parties.” Dkt. # 191-10 (Ex. B to Shih Deck at 2).