From The Field

April 21, 2020

The USPTO’s Patent Trial Practice Guide (“Practice Guide”) states that, during an AIA trial proceeding, a petitioner may file a reply to a patent owner’s response.  (See Practice Guide, Section II(I).)  But the Practice Guide puts limits on what the petitioner may argue in its reply. 

Specifically, the reply should only respond to arguments raised in the patent owner’s response.  Responding to arguments does not mean that the petitioner in its reply can proceed “in a new direction with a new approach.”  (See id. at pg. 74.)  For example, a reply cannot raise a new issue that it had not addressed in its initial petition–which is often indicated by the petitioner’s submission of new evidence needed to make out a prima facie case for the unpatentability of a claim.  But the PTO does not prohibit all new evidence submitted in a reply. 

The recent Federal Circuit decision in Apple v. Andrea Electronics Corp., Slip Op. 2018-2382, -2383 (Fed. Cir. 2020), illustrates this.  In Apple v. Andrea Electronics, which is an appeal from an IPR proceeding, the Federal Circuit held that the petitioner Apple’s reply arguments and evidence were not “impermissible new matter,” Slip Op. at 17, and that the Board erred in refusing to consider them.  The court explained that the Board had “pars[ed] [Apple’s] arguments on reply with too fine a filter.”  Id. at 16 (citation omitted).  

The Federal Circuit’s decision offers important guidance for practitioners as to when arguments on reply are acceptable, and when they go too far “in a new direction.”  An understanding of the technology at issue in the IPR helps to illuminate the basis for the court’s holding. 

The Patent at Issue

Andrea’s U.S. Patent No. 6,363,345 relates to certain aspects of digital audio processing.  The ’345 patent describes a system that reduces or cancels noise in an audio signal by a process known as “spectral subtraction.”  The process converts an audio signal to the frequency domain.  The process then generates a histogram, which divides the signal into multiple “frequency bins,” with each “bin” corresponding to one of the frequencies present in the signal.  The process then sets separate “thresholds” for each frequency bin in order to distinguish voice in the signal from unwanted background noise, and employs “spectral subtraction” to remove ambient noise based on the thresholds.  

In order to determine the thresholds for each frequency bin, the system sets two “minimum values,” i.e., a “future minimum” and a “current minimum.”  A “future minimum” value is initiated, for example, every 5 seconds with the value of the current magnitude of the signal.  Over the next 5 seconds, the future minimum value is compared with the current magnitude value of the signal.  If the current magnitude is smaller than the future minimum, the future minimum is replaced with that smaller magnitude—which becomes the new future minimum. 

At the start of every 5-second time interval, the “current minimum” is set as the value of the future minimum that was determined over the previous 5 seconds.  The current minimum value then follows the minimum value of the signal over the next 5 seconds by comparing its value with the current magnitude value. 

The final “current minimum” is used to calculate the thresholds in the spectral subtraction process, while the “future minimum” values are used to initiate and refresh the current minimum values.  Using the calculated thresholds, estimated noise magnitude is subtracted from the current magnitude value of the bin, resulting in a cleaner signal.

Apple’s Initial Petition in The IPR Proceeding

Andrea sued Apple for infringement of the ’345 patent, and Apple responded by filing two IPRs—the ’626 IPR and the ’627 IPR.  The ’626 IPR focused on claims 1-25 of the ’345 patent, and the ’627 IPR focused on claims 38-47 of the patent.  The dispute over whether Apple raised an improper new argument on reply arose in the context of the ’626 IPR—specifically, in the context of whether prior art cited by Apple rendered claims 6-9 unpatentable. 

In the ’626 IPR petition, Apple argued that claims 6-9 of the ’345 patent would have been obvious over the Hirsch and Martin references.  Claims 6-9 recite various features related to the two “minimum values” discussed above.  For example, claim 7 recites “wherein said future minimum value is replaced with the current magnitude value when said future minimum value is greater than said current magnitude value,” and claim 8 recites “wherein said current minimum value is replaced with the current magnitude value when said current minimum value is greater than said current magnitude value.”   

Apple’s petition contended that Martin’s algorithm could be used to track the magnitude of the noise signal, in accordance with the process set out by claims 6-9.  According to Apple, Martin’s noise floor estimation process algorithm used a “current minimum” and a “future minimum” to track the minimum signal power over windows of L digital samples, with each window L corresponding to a predetermined time period and being dividable into a number of sub-windows W.  As one example of how Martin satisfied the claim elements, Apple’s expert witness derived certain estimated noise floor values (i.e., Martin’s “current minimum”), and minimum observed noise power values (i.e., Martin’s “future minimum”) by inputting certain variables with an assumption of no sub-windows (i.e., W=1). 

The Patent Owner Response and Apple’s Reply

Andrea, in its Patent Owner Response, argued that Martin does not disclose a “future minimum” because there is no future minimum in both the monotonically increasing and the non-monotonically increasing applications of the algorithm.  According to Andrea, to determine whether the signal power is monotonically increasing over a window L, at least two sub-window minimum values must be available for the comparison, and therefore at least two sub-windows must be used (i.e., W must be at least 2).  Andrea supported its arguments by giving an example where Martin’s algorithm is applied in the context of multiple sub-windows because a determination of whether a signal’s power is monotonically increasing over window length L requires at least two sub-windows.

Apple rebutted Andrea’s arguments in its petitioner’s reply brief.  Apple argued first that Martin teaches the use of any number of sub-windows, and that where Martin is configured to use one sub-window (i.e., W=1), it is unnecessary to determine whether the sound is monotonically increasing.  Thus, according to Apple, when there is only one sub-window, Martin discloses a future minimum because the current minimum is set to be the same as the future minimum.  

Apple further contended that Andrea’s argument for situations where W > 1 relied on a non-existent claim requirement, i.e., that a future minimum be a minimum across the entire window L containing multiple sub-windows.  This is a scenario Apple had not addressed in its initial papers.  For the application where a signal is monotonically increasing, Apple contended that the claims do not specify the period over which the future minimum must be calculated, and thus that Martin satisfies the claims.  And for the application where a signal is non-monotonically increasing, Apple argued that the claims do not prohibit calculating the future minimum over a particular data window or using the minimum of a previous sub-window as the future minimum. 

Apple supported its arguments with a reply expert declaration, and Andrea did not file a sur-reply or move to strike any portion of Apple’s reply submission.  However, the Board refused to consider Apple’s reply arguments related to applying Martin’s algorithm in the context of multiple sub-windows (W > 1).  The Board held that these arguments raised a new theory of unpatentability for the first time on reply.

The Federal Circuit Vacates the Board’s Decision

The Federal Circuit disagreed with the Board and found that Apple’s legal ground did not change in its reply.  The court held that Apple’s reply relied on the same algorithm from the same prior art reference to support the same legal argument as it had asserted in its initial petition. 

The Federal Circuit distinguished this case from cases where the petitioner had asserted a new theory of unpatentability.  The first such case was Ariosa Diagnostics v. Verinata Health Inc., 805 F.3d 1359 (Fed. Cir. 2015), where the petitioner’s reply raised a new argument because the petitioner relied on an embodiment of the prior art that was not discussed in the petition.  The second case was Intelligent Bio-Sys., Inc. v. Illumina Cambridge Ltd., 821 F.3d 1359 (Fed. Cir. 2016), where the reply brief and the accompanying declaration exceeded the scope of the petition.  The reply there cited several new references and argued that the claim was obvious for reasons other than those described in the originally relied-upon prior art.

By contrast, in this case, the Federal Circuit stated that Apple’s reply brief merely demonstrated another example of the same algorithm to further explain why Martin discloses the claim limitations.  “The Board ‘pars[es] [the petitioner’s] arguments on reply with too fine a filter.’ …  It is unreasonable to hold petitioners to such a high standard that, if they choose to rely on one example of an algorithm, they must either discuss all potential permutations of the variables or risk waiving the opportunity to further discuss other relevant examples in their reply.”  Apple v. Andrea Electronics Corp. at 16 (citation omitted).

The Federal Circuit further pointed out that “the petitioner in an inter partes review proceeding may introduce new evidence after the petition stage if the evidence is a legitimate reply to evidence introduced by the patent owner.”  Id. at 16 (citation omitted).  That is precisely what Apple did here.  Apple’s reply brief “squarely responds to Andrea’s Patent Owner Response,” by (1) “rebutting Andrea’s arguments addressing the [no sub-windows] scenario;” (2) “argu[ing] that Andrea’s multiple sub-window discussion ‘reads non-existent limitations into the claims;’” and (3) explaining “why the algorithm” discloses the required claim elements “even in the context of multiple sub-windows[.]”  Id. at 16-17. 

The Apple v. Andrea Electronics Corp. helps clarify where the line should be drawn between improperly raising a new issue in a petitioner’s reply, and properly presenting arguments and new evidence on reply that respond to the patent owner’s response.  While the petitioner cannot advance a new theory of unpatentability by relying on new references or previously-undiscussed embodiments of an existing reference, the petitioner should be able to introduce evidence that rebuts or responds to the evidence introduced by the patent owner—as long as the petitioner relies on an unpatentability theory already asserted in its petition.

Author: Ray Huang