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 .
Continued Examination Under 37 CFR 1.114
A request for continued examination under, including the fee set forth in 37 CFR1.17(e), was filed in this application after final rejection. Since this application is eligiblefor continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e)has been timely paid, the finality of the previous Office action has been withdrawnpursuant to 37 CFR 1.114. Applicant's submission filed on 06/18/2026 has been entered.
Status of the Claims
Claims 1-20 are pending.
Response to Applicant’s Arguments
In response to “In Jax, therefore, the surround sound signals 821 that are provided to prediction block 85 to generate predicted sound components 851 are lossy-decoded, not losslessly decoded. The reference signal in Jax is accordingly derived from lossy-decoded audio samples of the lower hierarchical layer-not from previously losslessly decoded audio samples of the previous unconstrained audio mix, as required, in part, by amended claim 1. Moreover, the lower hierarchical layer in Jax (surround sound 821) is not a separate, independently losslessly reconstructable deliverable mix. It is a lossy, near- lossless representation that exists only as a stepping stone to reconstructing the higher- order HOA signal 83q. Jax does not contemplate, disclose, or suggest the output of the lower layer as an independent lossless deliverable. The output of Jax (per Jax, par. [0074] and Fig. 8) is the reconstructed HOA signal 83q-a single end-product. The lower-layer surround signals are intermediate, lossy artifacts, not independently usable lossless mixes of the audio content” and “Specifically, Kim discloses a hierarchical bitstream comprising a lossy-coded base layer and a lossless- coded enhancement layer (Kim, par. [0036] and Fig. 4). As acknowledged by the Office Action (p. 7), Kim teaches restoring a lossless audio signal by "adding lossy signal restored by the lossless decoding unit 310 to the lossy signal restored by the lossy decoding unit 320 (i.e., previous or first layer in the hierarchy)." Thus, in Kim, the so- called "previous" or "first" layer in the hierarchy is itself a lossy base layer-restored by lossy decoding unit 320-and is not a losslessly reconstructed unconstrained mix. The decoded audio samples used to generate any reference in Kim are therefore lossy, not losslessly decoded, as required by amended claim 1”.
In view of such amendment to claims 1, 8, and 15, rejections under Kjoerling Jax and Kim are withdrawn. Upon further search and consideration, please see details of a new combination of references set forth below.
Claim Rejections - 35 USC § 103
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 103 that form the basis for the rejections under this section made 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 1-3, 7-10, 14-17, and 20 are rejected under 35 USC 103(a) as being unpatentable over Kjoerling et al. (US 2016/0055855 A1) in view of Melkote et al. (“Hierarchical and Lossless Coding of Audio Objects in Dolby TrueHD”) and Davis (US 2013/0317833 A1).
Regarding Claims 1, 8, and 15, Kjoerling discloses an apparatus (¶247) providing preservation of artistic intention for audio content (¶198, a corresponding decoder to facilitate lossless decoding) comprising:
a memory / non-transitory processor-readable medium storing instructions / program (¶247, non-transitory computer storage media storing computer readable instructions); and
at least one processor executes the instructions including a process (¶247, microprocessor) configured to:
provide a lossless audio reconstruction process including unconstrained audio mixes for an audio content (¶¶88-89, a decoder 100 for reconstruction of an encoded 5.1 surround sound; ¶198, decoder facilitates lossless decoding); and
reconstruct, using the lossless audio reconstruction process, each unconstrained audio mix in the audio content (¶198, at the corresponding decoder, use bit allocation to determine probabilities of quantization indices in order to facilitate lossless decoding).
Kjoerling does not teach the lossless audio reconstruction process is a hierarchical lossless audio reconstruction process.
Melkote teaches providing a hierarchical lossless audio reconstruction process including a hierarchy of unconstrained audio mixes for an audio content (Abstract, “Dolby TrueHD is a lossless and hierarchical audio coding format that not only enables compact bit-exact representation of the source multichannel audio signal, but also facilitates low complexity reconstruction of downmixes thereof”), wherein each unconstrained audio mix in the hierarchy comprises a different deliverable mix of the audio content that is independently losslessly reconstructable (Fig. 1 reproduced below where at the decoder, where the lossless decoder independently reconstruct substream 0 of M channels and substream 1 of N-M channels):
PNG
media_image1.png
196
399
media_image1.png
Greyscale
and reconstructing, using the hierarchical lossless audio reconstruction process, each unconstrained audio mix in the hierarchy for the audio content, except a first unconstrained mix in the hierarchy, based on a previous unconstrained audio mix in the hierarchy (p. 384, “suffice to say that can be exactly inverted with finite precision arithmetic, and the lossless original can be recovered by decoding both substreams and applying the appropriate matrix inverses to the internal channels”; Fig. 1 shows lossless original corresponding to substream 1 N-M channels is reproduced based on substream 0 M channels), wherein reconstructing a current unconstrained audio mix comprises generating a reference signal using previously losslessly decoded audio samples of the previous unconstrained audio mix (Fig. 1 shows lossless original was reconstructed using inverse LPC and Huffman decoded substream 0 M channels as a reference for Inverse LPC and Huffman decoded substream 1 N-M channels), and outputting a lossless reconstructed multichannel audio signal of the current unconstrained audio mix (Fig. 1, lossless original).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to provide a hierarchical audio reconstruction process to reconstruct each unconstrained audio mix in a hierarchy for the audio content based on a previous unconstrained audio mix in the hierarchy except the first unconstrained mix in the hierarchy in order to enable compact bit-exact representation of a source multichannel audio signal and facilitate low complexity reconstruction of downmixes (e.g., 5.1 ch or stereo) thereof (Melkote, Abstract; compare Kjoerling, ¶42, reconstruct sampled sound wave in multi-channel format and ¶88, reconstruction of an encoded 5.1 surround sound).
Melkote further teaches TrueHD encoder using matrices to transform input channels into N internal channels, each of which is then individually encoded into the bitstream via linear prediction and Huffman coding and lossless original can be recovered by decoding both substreams and applying the appropriate matrix inverses to the internal channels (p. 384, 1. Introduction).
The combination does not disclose decoding residual information for the current unconstrained audio mix relative to the reference signal.
Davis teaches a prediction + Huffman coding based lossless Encoder (Fig. 1) and a lossless Decoder applying appropriate matrix inverses (Fig. 3):
PNG
media_image2.png
446
562
media_image2.png
Greyscale
PNG
media_image3.png
451
590
media_image3.png
Greyscale
In particular, Fig. 1 shows a lossless encoder applying matrices to transform input audio samples into coded samples Sx (¶33), applying a prediction operation on Sx to generate residuals (¶36), and perform Huffman Coding on the residual (¶44) to generate encoded bitstream via prediction and Huffman coding as required by Melkote.
Similarly, Fig. 3 shows a lossless decoder applying Huffman decoding to recover residual information (¶¶52-53), apply inverse prediction of the residual information to predict Sx (¶53), and output a lossless reconstructed audio sample by applying matrix inverses (¶63) as required by Melkote.
Davis demonstrated that encoding N internal channels into bitstreams via linear prediction and Huffman coding (Melkote, p.384) requires predicting corresponding residuals and performing Huffman coding thereon.
Davis further demonstrated recovering the lossless original by decoding both substreams (Melkote, p.384) requires Huffman decoding and inverse prediction to recover residual information for the current unconstrained audio mix relative to the reference signal (Melkote, Fig. 1, Inverse LPC and Huffman decoding of substream 1 N-M channels recovers encoded residuals, Inverse LPC and Huffman decoding of substream 0 M channels recovers corresponding residuals, where lossless original is recovered by applying appropriate matrix inverses to the residuals of substream 1 N-M channels and substream 0 M channels)
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to perform decoding residual information for the current unconstrained audio mix relative to the reference signal before applying appropriate matrix inverses in order to recover the lossless original (Melkote, p.384; Davis, ¶63).
Regarding Claims 2, 9, and 16, Kjoerling discloses store multiple unconstrained mixes of the audio content for transmission (i.e., as a bitstream) or storage / digital file container (¶42).
As modified by Melkote to implement a lossless and hierarchical audio coding format for low complexity reconstruction of downmixes thereof, Melkote further teaches wherein the multiple unconstrained mixes comprise at least two deliverable mixes of the audio content having different channel configurations (Fig. 1, substream 0 M-channels and substream 1 N-M channels; p. 384, “scalable, bitstream structure that facilitates low-complexity decoding of downmix presentations, such as 5.1 ch or stereo, of the source multichannel signal”); and
utilizing the hierarchical lossless audio reconstruction process to preserve artistic intention for the audio content (p. 384, Abstract, “Dolby TrueHD is a lossless and hierarchical audio coding format that not only enables compact bit-exact representation of the source multichannel audio signal”);
wherein the previously decoded audio samples are used as predictive input at corresponding temporal positions relative to the decoded audio samples of the current unconstrained audio mix (Fig. 1 showing Inverse LPC and Huffman decoded residuals from M channels are used with Inverse LPC and Huffman decoded residuals from N-M channels to recover the lossless original).
Regarding Claims 3, 10, and 17, Kjoerling as modified by Melkote disclose wherein the process is further configured to: add different coded mix bitstreams to the digital file container (Kjoerling, ¶42, audio data are quantized and coded in a format suitable for transmission or storage; Melkote, Fig. 1, p. 384, “scalable, bitstream structure” comprising substream 0 of M channels and substream 1 of N-M channels; i.e., storing the scalable bitstream structure in a storage).
Regarding Claims 7, 14, and 20, Kjoerling as modified by Melkote disclose wherein each unconstrained audio mix in the hierarchy for the audio content comprise different versions of a same mix that are correlated (Melkote, Abstract, “TrueHD was primarily employed for lossless carriage of speaker feeds (typically, 7.1 ch), in which case downmixes of interest (5.1 ch or stereo) could be obtained via time-invariant linear transformation”; i.e., Fig. 1 shows decoding substream 0 M channels into Downmix of 5.1 ch or stereo while decoding substream 0 N-M channels into lossless original of 7.1 ch; Kjoerling, ¶88, reconstruction of encoded 5.1 or 7.1 surround sound), and the decoded residual information is combined with the reference signal to output the reconstructed multichannel audio signal as a lossless reconstruction of the current unconstrained audio mix (Melkote, Fig. 1, Lossless Original).
Claims 4-5, 11-12, and 18 are rejected under 35 USC 103(a) as being unpatentable over Kjoerling et al. (US 2016/0055855 A1) in view of Melkote et al. (“Hierarchical and Lossless Coding of Audio Objects in Dolby TrueHD”) and Davis (US 2013/0317833 A1) as applied to claims 1, 8, and 15, in view of McGrath et al. (US 2021/0166708 A1).
Regarding Claims 4-5, 11-12, and 18, Kjoerling as modified by Melkote discloses the process is further configured to: provide metadata that indicates specific signal processing operations that are utilized to construct one or more downmixes at one or more temporal points (Melkote, p. 384, Dolby Atmos format enables an immersive audio experience via audio objects: audio signals associated with time-varying spatial metadata representing sound sources with arbitrary location / motion in the room; p. 385, “An object audio rendering algorithm converts the position metadata into panning gains so that the object signal may be distributed into speaker signals…the panning gains define an MxN downmix matrix at each time instant the metadata is specified”; therefore, Fig. 1 shows Input Primitive Matrices being defined at each time instant the metadata is specified), the metadata is utilized in a hierarchical lossless coding predictor of the hierarchical lossless audio reconstruction process to emulate a mixing operation when reconstructing the current unconstrained audio mix from the previous unconstrained audio mix (Mekote, p. 384, “the encoder receives an M x N downmix matrix A (M < N) that defines an M channel downmix, Ax, of the source…N x N input primitive matrices, that are sequentially applied to transform the input channels into N internal channels, each of which is then individually encoded into the bitstream via linear prediction and Huffman coding”; p. 385, “define an MxN downmix matrix at each time instant the metadata is specified”; therefore, Fig. 1 shows recovering the lossless original comprising, at each time instant the metadata is specified, define a downmix matrix for the encoder that defines an M channel downmix, and apply the downmix matrix to transform input channels into N internal channels for linear prediction and Huffman coding).
The combination does not teach wherein the process is further configured to: provide metadata during creation of the audio content.
McGrath teaches an audio coding system comprising an encoding unit 110 and a decoding unit 120 where the encoding unit generates a bitstream based on an immersive audio signal for transmission to the decoding unit for reconstruction of the immersive audio signal (¶28), the encoding unit configured to provide metadata indicating specific signal processing operations that are utilized to construct one or more downmixes at one or more temporal points during creation of the audio content (¶30, encoding unit comprises downmix module configured to downmix the immersive audio signal and a joint coding module configured to determine joint coding metadata 205 (spatial audio resolution reconstruction metadata or SPAR) configured to reconstruct the immersive audio signal from the downmix channel signals; ¶32, insert joint coding metadata 205 into bitstream 101 for transmission to corresponding decoding unit; ¶36, decoding unit comprises a reconstruction module configured to derive a reconstructed signal from the joint coding metadata, the joint coding metadata conveys the time elements of an upmix matrix that allows reconstructing the immersive audio signal), wherein the metadata is utilized in a coding predictor (¶45, the encoding unit having a joint coding unit modified to determine the joint coding metadata which allows the plurality of downmix channel signals to be reconstructed and hence the metadata is indicative of the predictor which has been used to generate compacted channel signals from the plurality of downmix channel signals).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to provide metadata indicating specific signal processing operations that are utilized to construct one or more downmixes at one or more temporal points and utilize the metadata in a hierarchical lossless coding predictor when performing the hierarchical lossless audio reconstruction process of Melkote to emulate a mixing operation when reconstructing the current unconstrained audio mix from the previous unconstrained audio mix (Mekote, Fig. 1, metadata being applicable to linear prediction LPC and Huffman coding of residual: when encoding substream 0 M channels, using the metadata to reconstruct downmix from Substream 0 and to reconstruct Lossless Original from Substream 1 to enable “an immersive audio experience via audio objects: audio signals associated with time-varying spatial metadata representing sound sources with arbitrary location/motion in the room”; p. 384; see also p. 385, “define an MxN downmix matrix at each time instant the metadata is specified”) in order to reconstruct multi-channel input signal from a plurality of downmix channel signals (McGrath, ¶31; compare Kjoerling, ¶53).
Claims 6, 13, and 19 are rejected under 35 USC 103(a) as being unpatentable over Kjoerling et al. (US 2016/0055855 A1) in view of Melkote et al. (“Hierarchical and Lossless Coding of Audio Objects in Dolby TrueHD”) and Davis (US 2013/0317833 A1) as applied to claims 5, 12, and 18, in further view of Vilkamo et al. (US 2024/0284134 A1).
Regarding Claims 6, 13, and 20, Kjoerling and Melkote as modified by McGrath disclose wherein the metadata is utilized in a hierarchical lossless coding predictor (Melkote, p. 385, “An object audio rendering algorithm converts the position metadata into panning gains so that the object signal may be distributed into speaker signals…the panning gains define an MxN downmix matrix at each time instant the metadata is specified”; Fig. 1, shows Input Primitive Matrices or downmix matrix at each time instant the metadata is specified for linear prediction and Huffman coding; compare McGrath, ¶45, the encoding unit having a joint coding unit modified to determine the joint coding metadata which allows the plurality of downmix channel signals to be reconstructed and hence the metadata is indicative of the predictor which has been used to generate compacted channel signals from the plurality of downmix channel signals), and the metadata is created during a mixing stage (McGrath, ¶30, encoding unit comprises a downmix module configured to downmix the multi-channel input signal / immersive audio signal to a plurality of downmix channel signals; ¶¶31-32, transform the downmix channel signals into subband domain to determine the joint coding metadata) and parsed during a coding stage (Melkote, p. 385, “the panning gains define an MxN downmix matrix at each time instant the metadata is specified”; i.e., use metadata to specify matrix for linear prediction and Huffman coding at the encoder side of Fig. 1; compare McGrath, ¶32, determine the joint coding metadata on a per subband basis and insert the joint coding metadata for the different subbands into the bitstream).
The combination does not teach the metadata is created using machine learning.
Vilkamo teaches using machine learning to create metadata (¶¶62-63, trained machine learning model 109 outputs spatial metadata 115 for rendering spatial audio signals, the spatial metadata 115 comprises information relating to one or more spatial properties of spatial sound environments corresponding to microphone signals; see Fig. 4, determine spatial metadata at 409 for audio processing 411) during a coding stage (Fig. 6, create spatial metadata for encoding at 605).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to use machine learning to create metadata in order to obtain high quality spatial metadata even from sub-optimal microphone arrays (Vilkamo, ¶46) comprising information relating to one or more spatial properties of spatial sound environments corresponding to microphone signals (Vilkamo, ¶62; compare McGrath, ¶31, joint coding module 230 for determining joint coding metadata is a spatial audio resolution reconstruction “SPAR” module).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to examiner Richard Z. Zhu whose telephone number is 571-270-1587 or examiner’s supervisor Hai Phan whose telephone number is 571-272-6338. Examiner Richard Zhu can normally be reached on M-Th, 0730:1700.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/RICHARD Z ZHU/Primary Examiner, Art Unit 2654 06/27/2026
Prosecution Insights