DETAILED ACTION
This communication is in response to the Amendments and Arguments filed on 1/21/2026.
Claims 1 ,4-10, 13-16, 18, and 21-27 are pending and have been examined.
All previous objections / rejections not mentioned in this Office Action have been withdrawn by the examiner.
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Amendments
Applicant has amended independent claim 1, 10, and 18 by adding “is less than” limitation and has added dependent claim 27. The added limitations raises new grounds for rejection. Since Applicant’s arguments are directed towards the new amendment, the arguments are moot in view of new grounds for rejection.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1, 4, 5, 10, 13, 14, 18, 22, and 26 are rejected under 35 U.S.C. 103 as being unpatentable over Talwar et al. (U.S. PG Pub No. 20160118042), hereinafter Talwar, in view of Garcia Martinez et al. (U.S. PG Pub No. 20130054236), hereinafter Garcia Martinez.
Regarding claim 1, 10 and 18 Talwar teaches:
(Claim 1) A method comprising: (P0005, Method of front-end processing an audio signal.)
(Claim 10) An apparatus comprising: one or more network processor units to communicate with devices in a network; and a processor coupled to the one or more network processor units and configured to perform: (P0019, Processor can be any type of device capable of processing electronic instructions. … Processor executes various types of digitally-stored instructions, such as software or firmware programs stored in memory, which enable the telematics unit to provide a wide variety of services.; P0020, Telematics unit can be used to provide a diverse range of vehicle services that involve wireless communication.)
(Claim 18) A non-transitory computer readable medium encoded with instructions that, when executed by a processor, cause the processor to perform: (P0019, Processor can be any type of device capable of processing electronic instructions. … Processor executes various types of digitally-stored instructions, such as software or firmware programs stored in memory, which enable the telematics unit to provide a wide variety of services.)
by a microphone, capturing audio including analogue audio frames; (P0034, The voice-activity detector may receive an incoming audio signal comprised of the user's utterance and noise via the microphone.)
digitizing the analog audio frames into audio frames of audio samples; (P0034, The voice-activity detector … may determine the speech frames or speech-dominant frames associated with the user's utterance. )
performing frame-based voice activity detection on the audio samples; (P0035, The voiced-unvoiced classifier may determine or classify each of the speech frames as either voiced or unvoiced speech frames.)
(Claim 1) detecting audio to produce audio frames; (P0040, FIG. 2 further illustrates a flow diagram beginning with receiving the user's utterance into the microphone. The utterance may be processed by the voice-activity detector, then the voiced-unvoiced classifier, and then the SNR evaluator. As will be explained below (FIG. 3), when the SNR evaluator determines that the SNR of a speech frame exceeds a predetermined threshold, that speech frame may bypass the noise suppressor.)
(Claim 10 + 18) receiving audio frames; (P0040, FIG. 2 further illustrates a flow diagram beginning with receiving the user's utterance into the microphone. The utterance may be processed by the voice-activity detector, then the voiced-unvoiced classifier, and then the SNR evaluator. As will be explained below (FIG. 3), when the SNR evaluator determines that the SNR of a speech frame exceeds a predetermined threshold, that speech frame may bypass the noise suppressor.)
detecting that voice is continuously present or is not continuously present across consecutive audio frames based on the frame-based voice activity detection; (P0042, In step 320, the ASR engine—or more specifically the voiced-unvoiced classifier—determines for each of the speech frames whether they are voiced or unvoiced. The voiced/unvoiced determination may include pitch and/or formant analysis, or any other method known to skilled artisans.)
computing a full-spectrum energy of an audio frame of the audio frames; (P0036, The SNR evaluator may determine the relative signal strength of the classified speech frames.; P0038, Nonlimiting examples of feature extraction used in sub-stage include various tools such as: Mel frequency cepstral coefficients (MFCC).)
computing a signal-to-noise ratio (SNR) of the audio frame; (P0043, A signal-to-noise (SNR) value is determined.)
bypassing background noise removal (BNR) on the audio frame when the voice is continuously present across consecutive audio frames and the SNR is greater than an SNR threshold; (P0035, The voiced-unvoiced classifier may determine or classify each of the speech frames as either voiced or unvoiced speech frames.; P0043, SNR value is compared to (or against) a predetermined threshold (TV1) that may be stored in memory. If the SNR value is greater than threshold TV1, the voiced-frame bypasses the noise suppressor.)
bypassing the BNR when the voice is not continuously present across the consecutive audio frames and the full-spectrum energy is less than a full-spectrum energy threshold. (P0047, Returning to the frames in step determined to be unvoiced, these frames may be evaluated by SNR evaluator. … For each unvoiced frame, a signal-to-noise (SNR) value is determined, and the SNR value is compared to (or against) a predetermined threshold (TU1) that may be stored in memory. If the SNR value is greater than threshold TU1, the unvoiced-frame bypasses the noise suppressor. [A person of ordinary skill in the art, with simple math, would understand that SNR comparison to threshold value is equivalent to comparing signal strength to a value derived from noise floor value and a threshold value.])
upon bypassing the BNR first encoding the audio frame to produce a first encoded audio frame; (P0043, SNR value is compared to (or against) a predetermined threshold (TV1) that may be stored in memory. If the SNR value is greater than threshold TV1, the voiced-frame bypasses the noise suppressor.)
upon not bypassing the BNR, performing the BNR on the audio frame to produce a reduced-noise audio frame, and second encoding the reduced-noise audio frame to produce a second encoded audio frame; and (P0043, If the SNR value is less than or equal to the threshold TV1, the voiced-frame is provided to the noise suppressing step before it is downstream processed (i.e., provided first to the noise suppressor).)
transmitting the first encoded audio frame or the second encoded audio frame to a media server over a network. (P0040, Speech frames with an SNR less than (or equal to) the threshold may undergo noise suppression using one or more tools of the suppressor. In some instances, the nonspeech frames may also under noise suppression via suppressor. According to the diagram, the preprocessing stage then concludes, and thereafter, the speech frames (and nonspeech frames) proceed to the downstream processing stage and may pass through sub-stages in order to complete the interpretation of the utterance. Upon exiting the decoding sub-stage, the downstream processing stage ends, and the output may or may not be provided to the additional processing.; P0015, Communications system generally includes a vehicle, one or more wireless carrier systems, a land communications network.; P0017, The telematics unit preferably uses radio transmissions to establish a communications channel (a voice channel and/or a data channel) with wireless carrier system so that voice and/or data transmissions can be sent and received over the channel.)
Talwar does not specifically teach:
bypassing the BNR when the voice is not continuously present across the consecutive audio frames and the full-spectrum energy is less than a full-spectrum energy threshold.
Garcia Martinez, however, teaches:
bypassing the BNR when the voice is not continuously present across the consecutive audio frames and the full-spectrum energy is less than a full-spectrum energy threshold. (P0015, Making the decision of classifying the segments of the input signal as speech or as noise. Specifically, a first criterion relating to the energy of the signal based on the comparison with a threshold is used.; P0016, The proposed method for the detection of speech segments is performed in three stages. In the first stage the signal frames the energy of which does not exceed a certain energy threshold.)
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, compare energy with a threshold when deciding to bypass BNR. It would have been obvious to combine the references because the detection of speech and noise segments through energy is fundamental because without eliminating the periods in which the voice of the user is not present, would involve triggering processing loads. (Garcia Martinez P0003)
Regarding claim 4 and 13 Talwar in view of Garcia Martinez teach claim 1 and 10.
Talwar further teaches:
not bypassing the BNR includes not bypassing the BNR when the voice is continuously present across the consecutive audio frames and the SNR is less than the SNR threshold. (P0035, The voiced-unvoiced classifier may determine or classify each of the speech frames as either voiced or unvoiced speech frames.; P0043, Step 330 occurs at the SNR evaluator. For each voiced frame, a signal-to-noise (SNR) value is determined, and the SNR value is compared to (or against) a predetermined threshold (TV1) that may be stored in memory. … However, if the SNR value is less than or equal to the threshold TV1, the voiced-frame is provided to the noise suppressing step 350 before it is downstream processed (i.e., provided first to the noise suppressor).)
Regarding claim 5 and 14 Talwar in view of Garcia Martinez teach claim 1 and 10.
Talwar further teaches:
computing the full-spectrum energy includes peaks of the audio frame; and (P0038, Feature extraction used in sub-stage include various tools such as: linear predictive codes (LPC), perceptual linear prediction (PLP), Mel frequency cepstral coefficients (MFCC). [A person of ordinary skill in the art would understand that MFCC includes values of energy levels of different frequency bins for each frame segment of audio.])
the method further comprises computing an energy of a noise floor of the audio frame that excludes the peaks, wherein computing the SNR includes computing the SNR as a ratio of the full-spectrum energy of the audio frame to the energy of the noise floor of the audio frame. (P0036, The SNR evaluator may determine the relative signal strength of the classified speech frames. For example, a voiced or unvoiced speech frame may be compared against one or more predetermined threshold values to determine whether the speech frame is greater than a first predetermined threshold.; P0041, The method begins with step 310 where the voice-activity detector detects a noise floor and multiple speech frames (voiced, unvoiced) associated with a user's utterance from among other audio frames (e.g., nonspeech frames).; P0043, Signal-to-noise (SNR) value is determined.)
Regarding claim 22 and 26 Talwar in view of Garcia Martinez teach claim 1 and 18.
Talwar further teaches:
establishing a teleconference collaboration session over the network; and during the teleconference collaboration session, performing receiving, digitizing, the frame-based voice activity detection, detecting, computing the full-spectrum energy, computing the SNR, bypassing the BNR when the voice is continuously present, bypassing the BNR when the voice is not continuously present, first encoding, second encoding, and transmitting. (P0017, Telematics unit can be an OEM-installed (embedded) or aftermarket device that is installed in the vehicle and that enables wireless voice and/or data communication over wireless carrier system and via wireless networking. This enables the vehicle to communicate with call center, other telematics-enabled vehicles, or some other entity or device.; P0040, FIG. 2 further illustrates a flow diagram beginning with receiving the user's utterance into the microphone. The utterance may be processed by the voice-activity detector, then the voiced-unvoiced classifier, and then the SNR evaluator. As will be explained below (FIG. 3), when the SNR evaluator determines that the SNR of a speech frame exceeds a predetermined threshold, that speech frame may bypass the noise suppressor.; P0043, A signal-to-noise (SNR) value is determined.; P0043, SNR value is compared to (or against) a predetermined threshold (TV1) that may be stored in memory. If the SNR value is greater than threshold TV1, the voiced-frame bypasses the noise suppressor. … SNR value is compared to (or against) a predetermined threshold (TV1) that may be stored in memory. If the SNR value is greater than threshold TV1, the voiced-frame bypasses the noise suppressor.)
Claim 6 is rejected under 35 U.S.C. 103 as being unpatentable over Talwar in view of Garcia Martinez and further view of Mohammad et al. (U.S. PG Pub No. 20120250882), hereinafter Mohammad.
Talwar in view Garcia Martinez does not specifically teach:
echo-canceling the audio frame to produce an echo-canceled audio frame prior to performing the BNR, performing first encoding, and performing second encoding.
Mohammad, however, teaches:
echo-canceling the audio frame to produce an echo-canceled audio frame prior to performing the BNR, performing first encoding, and performing second encoding. (P0037, Tools may be any suitable device or algorithm for minimizing, quashing, or suppressing noise in the speech frames. Nonlimiting examples of tools include: filtering devices, noise reducers (NR), dynamic noise reducers (DNR), echo cancellers.)
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to echo cancel the audio frame prior to performing the BNR. It would have been obvious to combine the references because linear echo cancellers may be at the front of the processing chain, i.e., to avoid any non-linear processes and also to avoid re-modeling/re-learning the rapid variations in the processing path due to the spatial processor. (Mohammad P0031)
Claims 7-9, 15-17, 21, 24, and 25 are rejected under 35 U.S.C. 103 as being unpatentable over Talwar in view of Garcia Martinez and further view of Gass et al. (U.S. PG Pub No. 20020188442), hereinafter Gass.
Regarding claim 7 and 15 Talwar in view of Garcia Martinez teaches claim 1 and 10.
Talwar further teaches:
performing the frame-based voice activity detection on the audio frames to produce decisions that indicate that the voice is present or that the voice is not present for the audio frames; and detecting that the voice is continuously present includes detecting that the voice is continuously present when the decisions include a first number of consecutive decisions, which all indicate that the voice is present. (P0042, Voiced-unvoiced classifier—determines for each of the speech frames whether they are voiced or unvoiced. The voiced/unvoiced determination may include pitch and/or formant analysis, or any other method known to skilled artisans.)
Talwar in view of Garcia Martinez does not specifically teach:
performing the frame based voice activity detection on the audio frames to produce decisions that indicate that the voice is present or that the voice is not present for the audio frames; and detecting that the voice is continuously present includes detecting that the voice is continuously present when the decisions include a first number of consecutive decisions, which all indicate that the voice is present.
Gass, however, teaches:
performing the frame based voice activity detection on the audio frames to produce decisions that indicate that the voice is present or that the voice is not present for the audio frames; and detecting that the voice is continuously present includes detecting that the voice is continuously present when the decisions include a first number of consecutive decisions, which all indicate that the voice is present. (P0013, Method of detecting voice activity in a signal divided into frames, the method including a step of smoothing a “voice” or “noise” initial decision made for each frame, the smoothing step including a step that makes a “voice” final decision for a frame n.; P0080-P0081, The second step 132 to 135 (analogous to the step 32 to 35) consists in making the “voice” decision if: The decision for the preceding two frames was “voice”.)
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to detect presence of voice in a number of consecutive decisions. It would have been obvious to combine the references because the method avoids the phenomenon of loss of speech segments because the smoothing function has an inertia corresponding to the duration of i frames for the return to a “noise” decision. (Gass P0020)
Regarding claim 8 and 16 Talwar in view of Garcia Martinez teaches claim 7 and 15.
Talwar in view of Garcia Martinez does not specifically teach:
detecting that the voice is not continuously present includes detecting that the voice is not continuously present when the decisions include a second number of consecutive decisions, which all indicate that the voice is not present.
Gass, however, teaches:
detecting that the voice is not continuously present includes detecting that the voice is not continuously present when the decisions include a second number of consecutive decisions, which all indicate that the voice is not present. (P0084-P0085, The third step 136 to 139, 143 (differing little from the step 36 to 39) makes the “noise” final decision 142 if: a “noise” decision was made for the last ten frames.)
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to detect presence of voice in a number of consecutive decisions. It would have been obvious to combine the references because the method avoids the phenomenon of loss of speech segments because the smoothing function has an inertia corresponding to the duration of i frames for the return to a “noise” decision. (Gass P0020)
Regarding claim 9 Talwar in view of Garcia Martinez teach claim 8.
Talwar in view of Garcia Martinez does not specifically teach:
wherein the first number of consecutive decisions is less than the second number of consecutive decisions.
Gass, however, teaches:
wherein the first number of consecutive decisions is less than the second number of consecutive decisions. (P0080-P0081, The second step 132 to 135 (analogous to the step 32 to 35) consists in making the “voice” decision if: The decision for the preceding two frames was “voice”.; P0084-P0085, The third step 136 to 139, 143 (differing little from the step 36 to 39) makes the “noise” final decision 142 if: a “noise” decision was made for the last ten frames.)
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to detect presence of voice in a number of consecutive decisions. It would have been obvious to combine the references because the method avoids the phenomenon of loss of speech segments because the smoothing function has an inertia corresponding to the duration of i frames for the return to a “noise” decision. (Gass P0020)
Regarding claim 21, 24, and 25 Tarwar in view of Garcia Martinez teach claim 1, 10, and 18.
Talwar in view of Garcia Martinez does not specifically teach:
detecting that the voice is continuously present or is not continuously present includes implementing a state machine that processes voice is present and voice is not present decisions from the frame-based voice activity detection on a frame-by-frame to produce a first state that indicates the voice is stable and a second state that indicates the voice is not stable across the consecutive audio frames, respectively.
Gass, however, teaches:
detecting that the voice is continuously present or is not continuously present includes implementing a state machine that processes voice is present and voice is not present decisions from the frame-based voice activity detection on a frame-by-frame to produce a first state that indicates the voice is stable and a second state that indicates the voice is not stable across the consecutive audio frames, respectively. (P0074, The smoothing comprises four steps, which follow on from the “voice” or “noise” initial decision 21 based on a plurality of criteria. Of these four steps, three (tests 131, 132, 136) are analogous to three steps described above (tests 31, 32, 36), the fourth step 40 previously described is eliminated, and a preliminary step is added before the first step 31 described above. Inertia counting is added to obtain an inertia with a duration equal to five times the duration of a frame, for example, before changing from the “voice” decision to the “noise” decision when the energy of the frame has become weak. This duration is therefore equal to 50 ms in this example.)
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to detect presence of voice in a number of consecutive decisions. It would have been obvious to combine the references because the method avoids the phenomenon of loss of speech segments because the smoothing function has an inertia corresponding to the duration of i frames for the return to a “noise” decision. (Gass P0020)
Claim 23 is rejected under 35 U.S.C. 103 as being unpatentable over Talwar in view of Garcia Martinez and further view of Moradi et al. (U.S. PG Pub No. 20230402043), hereinafter Moradi.
Regarding claim 23 Talwar in view of Garcia Martinez teach claim 5.
Talwar in view of Garcia Martinez does not specifically teach:
computing the full-spectrum energy includes computing a discrete Fourier transform (DFT) of the audio frame, and summing first energy across all frequency bins of the DFT; and computing the energy of the noise floor includes finding the peaks and particular frequency bins corresponding to the peaks, and summing second energy across all of the frequency bins except the particular frequency bins.
Moradi, however, teaches:
computing the full-spectrum energy includes computing a discrete Fourier transform (DFT) of the audio frame, and summing first energy across all frequency bins of the DFT; and computing the energy of the noise floor includes finding the peaks and particular frequency bins corresponding to the peaks, and summing second energy across all of the frequency bins except the particular frequency bins. (P0048, The frequency domain analysis may be a discrete Fourier transform in accordance with [equation].; P0050, Each peak may be associated with a number of frequency bins representing the peak. [Equation] G (i) is the group of bins representing the peak at frequency fi.; P0013, [Equation - noise disregarding peaks]; P0052, Ex is the energy of the complete spectrum, Exnoise is the energy of the noise spectrum.; P0063, Exnoise [equation - sum of energy disregarding G(i) energy])
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to calculate the noise floor by disregarding the detected peaks. It would have been obvious to combine the references because disregarding the peaks obtains an accurate isolated noise spectrum that is later used for noise signal ratio calculation and noise attenuation. (Moradi P0013, P0014)
Claim 27 is rejected under 35 U.S.C. 103 as being unpatentable over Talwar in view of Garcia Martinez and further view of Jin et al. (U.S. PG Pub No. 20070262819), hereinafter Jin.
Regarding claim 27 Talwar in view of Garcia Martinez teach claim 1.
Talwar in view of Garcia Martinez does not specifically teach:
prior to digitizing, performing analog gain control and echo-canceling on the audio frames; and after digitizing, and prior to the BNR, performing digital gain control on the audio samples.
Jin, however, teaches:
prior to digitizing, performing analog gain control and echo-canceling on the audio frames; and after digitizing, and prior to the BNR, performing digital gain control on the audio samples. (P0017, The microphone gain control is separated into two parts: the analog gain ADC gain and the digital gain Gd. Commercially available codecs, such as the MT8960 sold by Zarlink Semiconductor Inc. allow for both analog and digital gain control. The analog gain control acts on the input signal prior to analog to digital conversion, and the digital gain control acts on the digital samples of the input signal.; P0011, An echo canceller and noise reduction circuitry for canceling echo and reducing noise in the signal output by the first automatic gain control unit.)
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to perform analog and digital gain control and echo cancellation. It would have been obvious to combine the references because utilizing both analog and digital gain maintains speech quality and echo cancellation performance and maximizes signal dynamic range. (Jin P0008)
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to DANIEL WONSUK CHUNG whose telephone number is (571)272-1345. The examiner can normally be reached Monday - Friday (7am-4pm)[PT].
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, PIERRE-LOUIS DESIR can be reached at (571)272-7799. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/DANIEL W CHUNG/Examiner, Art Unit 2659
/PIERRE LOUIS DESIR/Supervisory Patent Examiner, Art Unit 2659