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 .
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
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.
Claim(s) 1-2, 8, 13, 15, 17, 23, 28 is/are rejected under 35 U.S.C. 103 as being unpatentable over Nam et al (12,542,138) in view of Ahmed et al (US2024/0127832).
Consider claims 1, 15. Nam teaches an apparatus comprising: a receiver circuit (fig. 9, 910), wherein the receiver circuit is arranged to receive an audio data signal (e.g., audio data) comprising: a downmix audio signal for a multichannel audio signal (the third audio signal includes one of a mono signal and a stereo signal (col. 25, lines 64-65), which represents a downscaled version of the original multichannel (ambisonics) signal); and upmix parametric data for upmixing the downmix audio signal (the frequency feature signal is used to reconstruct the fourth audio signal 145 that is equal/similar to the first audio signal 105" (col. 5, lines 57-60)); a subband generator circuit, wherein the subband generator circuit is arranged to generate a plurality of frequency subband signals for subbands of the downmix audio signal (e.g., the second audio signal 115 may be stereo-audio including a left (L) channel and a right (R) channel or mono-audio including 1 channel, col. 4, lines 31-50; The transformer 212 transforms the first audio signal 105 from a time domain into a frequency domain... according to various transformation methods including a short time Fourier transform (STFT)" (col. 7, lines 19-25). This frequency domain decomposition functions as the claimed subband generator circuit); an artificial neural network (a 'deep neural network (DNN)' is a representative example of an artificial neural network model simulating brain nerves" (col. 3, lines 54-56). NAM utilizes multiple DNNs (first, second, third, and fourth DNNs) to perform the decoding and processing tasks); and a generator circuit (the combiner 916 may obtain the fourth audio signal 145 by applying samples of the weight signal to samples of the intermediate audio signal" (col. 15, lines 29-32). The combiner in the decoding apparatus acts as the claimed generator circuit to reconstruct the final multichannel audio signal). NAM discloses a generator circuit, wherein the generator circuit is arranged to generate the multichannel audio signal from the subband samples of the subbands of the multichannel audio signal. NAM describes that "the combiner 916 may obtain the fourth audio signal 145 by applying samples of the weight signal to samples of the intermediate audio signal" (col. 15, lines 29-32). The combiner in the decoding apparatus acts as the claimed generator circuit to reconstruct the final multichannel audio signal.
NAM discloses multiple deep neural networks; but does not explicitly disclose that the artificial neural network comprises a plurality of subband artificial neural networks.
However, AHMED discloses a neural network configuration consisting of multiple cascaded or parallel learnable blocks, such as TADE Residual Blocks or multi-channel layers, that process subband-specific or band-limited frequency representations of the audio signal. AHMED describes that "the generator outputs simultaneously N different frequency bands samples at fs/N Hz" (para. [0450]), utilizing a plurality of learnable layers or blocks (para. [0357]). It would have been obvious to a person of ordinary skill in the art to modify NAM's full-band neural network upscaler to utilize a plurality of subband artificial neural networks as taught by AHMED to reduce computational complexity and processing delay.
NAM discloses generating multichannel audio signals; but does not explicitly show individual subband neural networks each dedicated to generating samples for a
corresponding frequency subband of the multichannel signal.
AHMED teaches generating samples for N frequency bands or channels simultaneously using learnable layers, describing that "Rather than synthesizing the speech signal in time domain at the output sample rate fs, the generator outputs simultaneously N different frequency bands samples at fs/N Hz" (para. [0450]). It would have been obvious to a person of ordinary skill in the art to combine these teachings to configure each subband artificial neural network to generate subband samples for a corresponding frequency subband of the multichannel audio signal to optimize parallel processing.
NAM discloses a third DNN that functionally matches a parameter generator by deriving weights from frequency features; but lacks an explicit disclosure of a dedicated parameter generator circuit generating subband-specific upmix parameter values.
AHMED teaches a conditioning learnable layer configured to process target data to obtain conditioning feature parameters across multiple bands, describing "at least one conditioning learnable layer configured, for the given frame, to process the target data to obtain conditioning feature parameters for the given frame" (Claim 1). It would have been obvious to a person of ordinary skill in the art to utilize AHMED's parameter generation structure to generate pluralities of upmix parameter values for subbands of the frequency subband representation from the upmix parametric data to condition the subband networks.
NAM discloses multiplying intermediate subband samples by weights, which mathematically mirrors node-level operations, but lacks the specific internal node structure of parallel subband neural networks configured to receive these inputs.
AHMED teaches neural network structures where input signals are processed in layers that are conditioned by parameter values, describing a styling element where "an element-by-element multiplication may be performed between a column of the styling conditions 74, 75 (conditioning feature parameters) and the first data 15" (para. [0343]). It would have been obvious to a person of ordinary skill in the art to configure each subband artificial neural network with a plurality of nodes arranged to receive both the upmix parameter values and the frequency subband signal samples to perform the upmixing calculations.
NAM discloses a general correlation and mapping between frequency feature signals and target multichannel outputs; but does not explicitly state that the input subband signal must correspond to the exact subband generated by the network.
AHMED teaches spectral consistency and time-frequency alignment between target data coefficients and the generated multi-band waveforms, describing that "different coefficients will take place in different entries of each column in association with coefficients associated with the frequency bands" (para. [0343]). It would have been obvious to a person of ordinary skill in the art to align the inputs and outputs such that the input frequency subband signal is for the subband for which the subband artificial neural network generates subband samples of the multichannel audio signal to maintain spectral consistency.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teachings of Ahmed into the teachings of Nam, since one of ordinary skill in the art, recognizing the efficiency problems identified in NAM, would have been motivated to apply AHMED's subband-specific neural network processing to NAM's system to allow the neural network upmixing to be performed in parallel across multiple frequency subbands, thereby reducing computational complexity and taking advantage of parallel processing architectures.
Consider claims 2, 17. Nam further teaches at least a first subband artificial neural network of the plurality of subband artificial neural networks comprises nodes for receiving parameter values of pluralities of upmix parameter values for other subbands than the subband of the subband artificial neural network (fig. 2, legacy downscaler 250, col. 8, lines 62-67).
Consider claims 8, 23. Nam further teaches at least one first subband of the plurality of subband artificial neural networks comprise at least two subband artificial neural networks, wherein the at least two subband artificial neural networks are arranged to generate samples for different components of a subband signal for the first subband (col. 4, lines 6-12).
Consider claims 13, 28. Ahmed further teaches the upmix parametric data provides parametric data for sequential time intervals, wherein at least a first subband artificial neural network of the plurality of subband artificial neural networks comprises first nodes, wherein the first nodes are arranged to receive parameter values of a plurality of upmix parameter values for another time interval of the sequential time interval than a time interval for which subband samples of the multichannel audio signal are generated (e.g., different times for different frames; fig. 2b, para 346).
Claim(s) 3-4, 12, 18-19, 27 is/are rejected under 35 U.S.C. 103 as being unpatentable over Nam et al (12,542,138) in view of Ahmed et al (US2024/0127832) as applied to claims 1, 15 above, and further in view of Sheaffer et al (11,997,463).
Consider claims 3, 18. Nam in view of Ahmed does not teach wherein at least a portion of parameters of the pluralities of upmix parameter values for different subband artificial neural networks are the same.
Sheaffer teaches a portion of parameters of the pluralities of upmix parameter values for different subband artificial neural networks are the same (e.g., the different or same spatial parameters by ML model 4, col. 6, lines 43-48).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teachings of Sheaffer into the teachings of Nam in view of Ahmed in order to provide a method performed by an audio system for generating realistic audio effects and ambiences using machine learning (ML) models (col. 1, lines 49-52)
Consider claim 4, 19. Sheaffer further teaches at least a portion of parameters of the pluralities of upmix parameter values for different subband artificial neural networks are different (e.g., the different or same spatial parameters by ML model 4, col. 6, lines 43-48).
Consider claims 12, 27. Sheaffer further teaches the pluralities of upmix parameter values have a different number of parameters for at least two subbands (e.g., the different or same spatial parameters by ML model 4, col. 6, lines 43-48).
Claim(s) 5, 14, 16, 20, 29 is/are rejected under 35 U.S.C. 103 as being unpatentable over Nam et al (12,542,138) in view of Ahmed et al (US2024/0127832) as applied to claims 1, 15 above, and further in view of Yu et al (10,510,360).
Consider claims 5, 20. Nam in view Ahmed does not clearly teaches the plurality of subband artificial neural networks for at least one subband comprises separate artificial neural networks for different channels of the multichannel audio signal.
Yu teaches plurality of subband artificial neural networks for at least one subband comprises separate artificial neural networks for different channels of the multichannel audio signal (see fig 1A, DNN1-DNNn).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teachings of Yu into the teachings of Nam in view Ahmed in order to provide embodiments that can combine certain weight groups of sub-bands that are more important for specific tasks and less important for other tasks (col. 2, lines 40-43).
Consider claims 14, 29. Yu further teaches the plurality of subband artificial neural networks are arranged to receive only subband samples of the downmix audio signal and parameter values generated from the upmix parametric data (see fig 1A, DNN1-DNNn). It is noted that the signal received by apparatus 100A (fig. 1A of Yu) must include subband samples of the downmix audio signal and parameter values generated from the upmix parametric data.
Consider claim 16. Nam further teaches a computer program stored on a non-transitory medium, wherein the computer program when executed on a processor performs the method of claim 14 (col. 25, lines 52-67).
Claim(s) 9-10, 24-25 is/are rejected under 35 U.S.C. 103 as being unpatentable over Nam et al (12,542,138) in view of Ahmed et al (US2024/0127832) as applied to claims 1, 15 above, and further in view of Arteaga et al (12,574,696).
Consider claims 9, 24. Nam in view of Ahmed does not teach the plurality of subband artificial neural networks are trained by training data, wherein the training data comprises training input audio signals, wherein the training input audio signals comprise samples of input multichannel audio signals, wherein the plurality of subband artificial neural networks are trained by using a cost function. Arteaga teaches the plurality of subband artificial neural networks are trained by training data, wherein the training data comprises training input audio signals, wherein the training input audio signals comprise samples of input multichannel audio signals, wherein the plurality of subband artificial neural networks are trained by using a cost function (col. 3, line 57 to col. 4, line 3). Arteaga further teaches the cost function comprises a component indicative of a difference between the training input audio signals and multichannel audio signals generated by the plurality of subband artificial neural networks (col. 2, lines 32-38).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teachings of Arteaga into the teachings of Nam in view of Ahmed in order to provide a method that includes improving the machine learning algorithm based on a difference between the first input multichannel audio signal and the first output multichannel audio signal (col. 2, lines 7-10).
Consider claims 10, 25. Nam in view of Ahmed does not teach the plurality of subband artificial neural networks are trained by training data, wherein the training data comprises training input audio signals, wherein the training input audio signals comprises samples of input multichannel audio signals, wherein the plurality of subband artificial neural networks are trained by using a cost function wherein the cost function comprises a component indicative of a difference between upmix parameter values for the input audio signals and upmix parameter values for the multichannel audio signals generated by the plurality of subband artificial neural networks. Arteaga teaches the plurality of subband artificial neural networks are trained by training data, wherein the training data comprises training input audio signals, wherein the training input audio signals comprise samples of input multichannel audio signals, wherein the plurality of subband artificial neural networks are trained by using a cost function (col. 3, line 57 to col. 4, line 3). Arteaga further teaches the cost function comprises a component indicative of a difference between the upmix parameter values for the input audio signals and upmix parameter values for the multichannel audio signals generated by the plurality of subband artificial neural networks (col. 2, lines 32-38). It is noted that the training input audio signals and multichannel audio signals generated by the plurality of subband artificial neural networks must include the upmix parameter values for the input audio signals and upmix parameter values for the multichannel audio signals generated by the plurality of subband artificial neural networks.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teachings of Arteaga into the teachings of Nam in view of Ahmed in order to provide a method that includes improving the machine learning algorithm based on a difference between the first input multichannel audio signal and the first output multichannel audio signal (col. 2, lines 7-10).
Allowable Subject Matter
Claims 6-7, 11, 21-22, 26 are objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to DUC NGUYEN whose telephone number is (571)272-7503. The examiner can normally be reached 6:30AM-3:45PM.
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, Duc Nguyen can be reached at 571-272-7503. 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.
DUC NGUYEN
Supervisory Patent Examiner
Art Unit 2691
/DUC NGUYEN/Supervisory Patent Examiner, Art Unit 2691
