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 .
This office action is sent in response to Applicant’s communication received on 04/03/2025 for the application number 19169733. The office hereby acknowledges receipt of the following placed of record in the file: Specification, Abstract, Oath/Declaration and claims.
Status of the claims
Claims 1-20 are presented for examination.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 04/03/2025 was filed before the
mailing date of the first office action. These submissions are in compliance with the provisions
of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the
examiner.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed
to an abstract idea without significantly more. The claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception because as
explained below.
Claim 1 recites a method for displaying information relating to an acoustic model that is established by being trained using a plurality of sound waveforms so as to generate acoustic features, the method comprising:
acquiring characteristic distribution of the plurality of sound waveforms used for training of the acoustic model, a characteristic of the characteristic distribution being one or more sound waveform characteristics;
and displaying information relating to the characteristic distribution.
Step (a) comprises a mental process. This step can be performed by a human as a person
can observe sound waveforms and determine characteristics of their distribution.
Step (b) comprises insignificant post-solutional activity because merely displaying the results of the foregoing analysis does not meaningfully alter the patentability.
Step 1: This part of the eligibility analysis evaluates whether the claim falls within any statutory category. See MPEP 2106.03. The claim recites at least method. Thus, the claim is a process, which is one of the statutory categories of invention. (Step 1: YES).
Step 2A, Prong One: This part of the eligibility analysis evaluates whether the claim recites a judicial exception. As explained in MPEP 2106.04, subsection II, a claim “recites” a judicial exception when the judicial exception is “set forth” or “described” in the claim. As discussed above, the broadest reasonable interpretation of step (a) recites a mental process.
Specifically, step (a) comprises a mental process. This step can be performed by a human as a person can observe sound waveforms and determine characteristics of their distribution.
Step 2A, Prong Two: This part of the eligibility analysis evaluates whether the claim as a whole integrates the recited judicial exception into a practical application of the exception or whether the claim is “directed to” the judicial exception. This evaluation is performed by (1) identifying whether there are any additional elements recited in the claim beyond the judicial exception, and (2) evaluating those additional elements individually and in combination to determine whether the claim as a whole integrates the exception into a practical application. See MPEP 2106.04(d).
The claim recites additional elements including displaying information relating to the characteristic distribution. However, they are recited at a high level of generality and perform a generic computer functions such as displaying data data. The use of a display merely automates the mental processes described above. Such processing does not impose any meaningful limit on the judicial exception. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea. (Step 2A, Prong Two: NO), and the claim is directed to the judicial exception. (Step 2A: YES).
Step 2B: This part of the eligibility analysis evaluates whether the claim as a whole
amounts to significantly more than the recited exception i.e., whether any additional element, or
combination of additional elements, adds an inventive concept to the claim. As explained with
respect to Step 2A, Prong Two, the displaying of characteristic data comprises an additional element that does not contribute to the patentability of the claim as a whole. It is at best mere instructions to “apply” the abstract ideas, which cannot provide an inventive concept. See MPEP 2106.05(f). At Step 2B, the evaluation of the insignificant extra-solutional activity consideration takes into account whether or not the extra-solutional activity is well understood, routine, and conventional in the field. See MPEP 2106.05(g). As known in the art these elements are well routine and conventional. Even when considered in combination, these additional elements represent mere instructions to implement an abstract idea or other exception on a computer and insignificant extra-solutional activity which do not provide an inventive concept. The claim is not patent eligible.
Claim 2 comprises a mental process, as a person can identify characteristics of the sound waveforms and determine their distribution.
Claim 3 comprises a mental process, as a person can determine whether training data lacks training of an acoustic model.
Claim 4 comprises a mental process, as a person can determine whether information was acquired during training.
Claim 5 comprises a mental process, as a person can acquire characteristics before training and compare them with characteristics acquired during training.
Claim 6 comprises a mental process, as a person can compare characteristics from previous training with characteristics from additional training.
Claim 7 comprises a mental process, as a person can select an acoustic model from among a plurality of trained acoustic models.
Claim 8 comprises a mental process, as a person can identify pitch, intensity, phoneme, duration, or style characteristics of sound waveforms.
Claim 9 comprises a mental process, as a person can identify first and second characteristics and determine their joint distribution.
Claim 10 comprises a mental process, as a person can determine whether a characteristic falls within a prescribed range.
Claim 11 comprises a mental process, as a person can identify a range in which the characteristic distribution is below a threshold.
Claim 12 comprises insignificant post solutional activity as displaying the limits of the identified range merely presents the results of the analysis.
Claim 13 comprises a mental process, as a person can identify musical pieces having a characteristic value within a lacking range and present the identified pieces to a user.
Claim 14 comprises a mental process, as a person can acquire training history and evaluate characteristics based on that history.
Claim 15 comprises a mental process, as a person can evaluate proficiency for a musical piece based on its score and acquired characteristics.
Claim 16 comprises a mental process, as a person can evaluate proficiency based on a musical score and acquired characteristics.
Claim 17 comprises a mental process, as a person can separately evaluate proficiency for sections of a musical piece.
Claim 18 comprises a mental process, as a person can evaluate multiple musical pieces and recommend a piece based on the evaluations.
Claim 19 comprises a mental process, as a person can compare acoustic models with a musical piece and recommend a matching model.
Claim 20 comprises a mental process, as a person can evaluate proficiency for portions of a musical piece as those portions are performed.
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1, 2, 3, 4, 5, 8 are rejected under 35 U.S.C. 103 as being unpatentable over Mase et al. (“HMM-based singing voice synthesis system using pitch-shifted pseudo training data”, INTERSPEECH 2010, pg. 845-848, 2010) in view of Gupta (US 20220277763 A1).
Regarding claim 1, Mase teaches a method for displaying information relating to an acoustic model that is established by being trained using a plurality of sound waveforms so as to generate acoustic features (Fig. 1 shows the system as a whole), the method comprising: acquiring characteristic distribution of the plurality of sound waveforms used for training of the acoustic model (Pg. 847, “Figure 2 and 3 show the distributions of pitch in training data (10 and 60 songs)” and Pg. 845, the training data including the pitch is extracted “from the waveform of a song in the singing voice database” where the “pitch is represented by the log F0 parameter” which is extracted from the waveform), a characteristic of the characteristic distribution being one or more sound waveform characteristics (Pg. 845, “Training data are mel-Cepstral coefficients, log fundamental frequencies (F0), and vibrato parameters” and Pg. 846, “Since pitch is represented by the log F0 parameter…”). Mase does not teach and displaying information relating to the characteristic distribution.
However, Gupta teaches displaying information relating to the characteristic distribution (Para 0020, teaching that an absolute measure of pitch may be based on “overall pitch distribution” and that the processor may assess pitch by “producing a pitch histogram”. See also para 0068, “The various measures calculated by module 202 and 204 … may be displayed on display 208” teaching display of information calculated from the pitch distribution).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify Mase in such a way as to incorporate the teachings of Gupta in order to allow the characteristics of the training data to be more readily presented to the user (Para 0068).
Regarding claim 2, Mase teaches wherein the characteristic distribution is acquired by analyzing the plurality of sound waveforms (Pg. 845, “we first extract various parameters to be used as training data from the waveform of a song in the singing voice database” and pg. 847, “Figure 2 and 3 show the distributions of pitch in training data (10 and 60 songs)”).
Regarding claim 3, Mase teaches wherein the information relating to the characteristic distribution indicates a lack of the training of the acoustic model (Pg. 845, “HMMs corresponding to contextual factors that rarely appear in training data cannot be well-trained” and pg. 847, “Before the pitch-shift is applied, the distributions are sparse”, wherein these teach the training data having a sparse area that indicates a lack of training).
Regarding claim 4, Mase teaches wherein the information relating to the characteristic distribution indicates an ability acquired by the acoustic model through the training (Mase, Pg. 845, “HMMs corresponding to contextual factors that rarely appear in training data cannot be well-trained” and “pitch should be properly covered since it has a great impact on the synthesized singing voice”, see also “pitches that rarely appear in training data cannot be properly generated because the system cannot model their fundamental frequency (F0) contours” teaching that the distribution of pitch in the training data indicates the pitches that the acoustic model is capable of properly modeling and wherein being able to generate a specific pitch constitutes an ability).
Regarding claim 5, Mase teaches wherein the acquiring is performed by acquiring, before the training of the acoustic model that is expected to be established through the training, the characteristic distribution of the plurality of sound waveforms used for the training of the acoustic model (Mase, Pg. 845, “first we increase the training data by adding pitch-shifted pseudo data… Then HMMs are trained with the training data” and pg. 847, “Figure 2 and 3 show the distributions of pitch in the training data (10 and 60 songs)” teaching that the characteristic distribution pertains to the training data that is prepared before the HMMs are trained).
Mase does not teach displaying information relating to the characteristic distribution of the plurality of sound waveforms used for the training of the acoustic model is displayed. However, Gupta teaches displaying information relating to the characteristic distribution of the plurality of sound waveforms used for the training of the acoustic model is displayed (Para 0020, teaching that an absolute measure of pitch may be based on “overall pitch distribution” and that the processor may assess pitch by “producing a pitch histogram”. See also para 0068, “The various measures calculated by module 202 and 204 … may be displayed on display 208” teaching display of information calculated from the pitch distribution).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify Mase in such a way as to incorporate the teachings of Gupta in order to allow the distribution of acoustic characteristics to be more readily presented to the user (Para 0068).
Regarding claim 8, Mase teaches wherein the one or more sound waveform characteristics include one or more of pitch, intensity, phoneme, duration, or style (Pg. 846, “Since pitch is represented by the log F0 parameter…”).
Claim 6 is rejected under 35 U.S.C. 103 as being unpatentable over Mase (“HMM-based singing voice synthesis system using pitch-shifted pseudo training data”, INTERSPEECH 2010, pg. 845-848, 2010) in view of Gupta (US 20220277763 A1) as above and further in view of Daido et al. (US 20210256959 A1).
Regarding claim 6, Mase teaches where the characteristic distribution is a characteristic distribution obtained by analyzing a plurality of sound waveforms (Pg. 847, “Figure 2 and 3 show the distributions of pitch in training data (10 and 60 songs)”, teaching that a characteristic distribution is obtained by analyzing training waveform data).
Mase does not teach wherein the training is additional training, and the plurality of sound waveforms have been used for previous training of the acoustic model before the additional training and a sound waveform is additionally used for the additional training.
However, Daido teaches wherein the training is additional training, and the plurality of sound waveforms have been used for previous training of the acoustic model before the additional training and a sound waveform is additionally used for the additional training (Para 0034, “the machine learning carried out by the learning processor 26 is classified into pre-training and additional training” and “the additional training is carried out after the pre-training” and para 0038, teaching pre-training data including audio signals used to train the synthesis model. See also para 0062-0064, teaching an additional audio signal V1 from an additional singer, generation of feature data from V1, addition of that data to the existing training data, and additional training using the resulting training data).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify Mase in such a way as to incorporate the teachings of Daido in order to adapt an already-trained synthesis model using newly obtained waveform data while retaining the benefit of the model’s previous training, thereby avoiding the need to retrain the model entirely from scratch (Para 0034).
Claim 7 is rejected under 35 U.S.C. 103 as being unpatentable over Mase (“HMM-based singing voice synthesis system using pitch-shifted pseudo training data”, INTERSPEECH 2010, pg. 845-848, 2010) in view of Gupta (US 20220277763 A1) as above and further in view of Forand et al. (US 20030130840 A1).
Regarding claim 7, Mase modified by Daido does not teach wherein the acoustic model before the additional training is selected by a user from among a plurality of trained acoustic models.
However, Forand teaches wherein the acoustic model before the additional training is selected by a user from among a plurality of trained acoustic models (Para 0003, “the user may select a suitable one among a few acoustic models” see also claim 1, teaching making a plurality of speech samples associated with respective acoustic models available to a user, “receiving the selection” and identifying the acoustic model associated with the selected sample).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify Mase in such a way as to incorporate the teachings of Forand in order to allow a user to select a model already suited to the desired characteristics, reducing the amount of training needed to obtain their desired model (Para 0003).
Claim 9 is rejected under 35 U.S.C. 103 as being unpatentable over Mase (“HMM-based singing voice synthesis system using pitch-shifted pseudo training data”, INTERSPEECH 2010, pg. 845-848, 2010) in view of Gupta (US 20220277763 A1) as above and further in view of Parsa et al. (WO 2008110016 A1)
Regarding claim 9, Mase does not teach wherein the one or more sound waveform characteristics include a first characteristic and a second characteristic, and in the displaying, a graph indicating a two-dimensional distribution of the first characteristic and the second characteristic is displayed.
However, Parsa teaches wherein the one or more sound waveform characteristics include a first characteristic and a second characteristic (Pg. 13, “fundamental frequency estimation to determine the pitch and intensity of the talker”, teaching first and second characteristics being pitch and intensity), and in the displaying, a graph indicating a two-dimensional distribution of the first characteristic and the second characteristic is displayed (Pg. 14, “a phonetogram is a two dimensional display of the pitch and intensity parameters” and “This display is therefore useful to quantify the dynamic range of the pitch and intensity values).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify Mase in such a way as to incorporate the teachings of Parsa in order to simultaneously visualize the relationship between multiple characteristics in training data (Pg. 14).
Claim 10 is rejected under 35 U.S.C. 103 as being unpatentable over Mase (“HMM-based singing voice synthesis system using pitch-shifted pseudo training data”, INTERSPEECH 2010, pg. 845-848, 2010) in view of Gupta (US 20220277763 A1) as above and further in view of Bobrow et al. (US 7629986 B2).
Regarding claim 10, Mase teaches wherein the acquiring of the characteristic distribution is performed by analyzing the plurality of sound waveforms and acquiring, as the one or more sound waveform characteristics, a first characteristic, and a second characteristic (Pg. 845, “we first extract various parameters to be used as training data from the waveform of a song in the singing voice database” and “Training data are mel-Cepstral coefficients, log fundamental frequencies (F0), and vibrato parameters”)
Mase does not teach in the displaying, the characteristic distribution of the second characteristic in a state in which the first characteristic is within a prescribed range is displayed.
However, Bobrow in the displaying, the characteristic distribution of the second characteristic in a state in which the first characteristic is within a prescribed range is displayed. (Col. 5, Ln. 46-51 , teaching an income histogram having bins representing respective income ranges and an age histogram of the same population, “the user has selected one of the income bins, and … the display system illustrated the user-selected income group’s distribution among the various age groups” thereby teaching displaying the distribution of a second characteristic for data having a first characteristic with a selected range).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify Mase in such a way as to incorporate the teachings of Bobrow in order to improve interpretation of multidimensional training data by allowing relationships to be identified more readily (Col. 5).
Claim 11 is rejected under 35 U.S.C. 103 as being unpatentable over Mase (“HMM-based singing voice synthesis system using pitch-shifted pseudo training data”, INTERSPEECH 2010, pg. 845-848, 2010) in view of Gupta (US 20220277763 A1) as above and further in view of Nadav et al. (US 11599579 B1).
Regarding claim 11, Mase does not teach wherein the displaying includes detecting, in the characteristic distribution that has been acquired, a lacking range of the plurality of sound waveforms related to the characteristic distribution, which is a range in which the characteristic distribution is smaller than a threshold value.
However, Nadav teaches wherein the displaying includes detecting, in the characteristic distribution that has been acquired, a lacking range of the plurality of sound waveforms related to the characteristic distribution, which is a range in which the characteristic distribution is smaller than a threshold value (Col. 9, Ln. 15-42, “the model system 104 bins event feature values to find frequency” and “the model system 104 may generate bind or ranges of event feature values, assign a given event feature to a bin based on its value, and find the frequency of the event feature value by considering the frequency of event feature values in the same bin”, see also col. 4, Ln. 29-36, “identifies event features that have values with a frequency less than a threshold frequency” teaching detection of value ranges whose distribution is lower than a threshold value).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify Mase in such a way as to incorporate the teachings of Nadav in order to identify underrepresented ranges of a distribution using a threshold (Col. 4).
Mase modified by Nadav does not teach displaying the lacking range.
However, Gupta teaches displaying the lacking range (Para 0020, teaching that an absolute measure of pitch may be based on “overall pitch distribution” and that the processor may assess pitch by “producing a pitch histogram”. See also para 0068, “The various measures calculated by module 202 and 204 … may be displayed on display 208” teaching display of information calculated from the pitch distribution, wherein the displaying of this pitch distribution could be applied to displaying the lacking range).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify Mase in such a way as to incorporate the teachings of Gupta in order to visually present the identified lacking range to the user (Para 0068).
Claim 12 is rejected under 35 U.S.C. 103 as being unpatentable over Mase (“HMM-based singing voice synthesis system using pitch-shifted pseudo training data”, INTERSPEECH 2010, pg. 845-848, 2010) in view of Gupta (US 20220277763 A1) and Nadav (US 11599579 B1) as above and further in view of Parsa et al. (WO 2008110016 A1)
Regarding claim 12, Mase does not teach wherein in the displaying, a characteristic value of at least one of an upper limit and a lower limit of the lacking range is displayed.
However, Parsa teaches wherein in the displaying, a characteristic value of at least one of an upper limit and a lower limit of the lacking range is displayed (Pg. 14, “in the pitch display, the clinician may set a minimum and maximum values for the pitch” and “the client can adjust his/her voice production to keep their pitch within this range”).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify Mase in such a way as to incorporate the teachings of Parsa in order to visually identify the boundaries of the detected lacking range to the user (Pg. 14).
Claim 13 is rejected under 35 U.S.C. 103 as being unpatentable over Mase (“HMM-based singing voice synthesis system using pitch-shifted pseudo training data”, INTERSPEECH 2010, pg. 845-848, 2010) in view of Gupta (US 20220277763 A1) and Nadav (US 11599579 B1) as above and further in view of Ref. A (JP 4286583 B2).
Mase does not teach wherein the displaying includes detecting, from among a plurality of musical pieces, a plurality of candidate musical pieces containing a note with a characteristic value within the lacking range, and presenting the plurality of candidate musical pieces to a user.
However, Ref. A teaches wherein the displaying includes detecting, from among a plurality of musical pieces, a plurality of candidate musical pieces containing a note with a characteristic value within the lacking range (Para 0023, “The supplemental corpus creation unit 13 searches the corpus storage unit 16 for a corpus including the deficient phoneme sequence obtained by the deficient phoneme sequence search unit 12” teaching searching source material for content containing a deficient characteristic; see also 0057, teaching that a deficient sequence may be determines base don pitch), and presenting the plurality of candidate musical pieces to a user (Para 0023, “The additional corpus may be output in any format such as print output, display on a display, or output to a file” teaching presentation of the selected candidate material to a user).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify Mase in such a way as to incorporate the teachings of Ref. A in order to identify and underrepresented characteristics to allow for more complete training (Para 0023).
Claim 14 is rejected under 35 U.S.C. 103 as being unpatentable over Mase (“HMM-based singing voice synthesis system using pitch-shifted pseudo training data”, INTERSPEECH 2010, pg. 845-848, 2010) in view of Gupta (US 20220277763 A1) as above and further in view of McCusker et al. (US 20220058511 A1).
Regarding claim 14, Mase modified by Gupta teaches wherein the acoustic model is a model that has acquired an ability to generate second acoustic features in accordance with musical score features of a second musical piece (Mase, Pg. 846, “a given musical score including lyrics to be synthesized is converted into a context dependent label sequence” and “the speech parameter generation algorithm [4] generates spectral, excitation, and vibrato parameters” teaching generation of acoustic features according to musical score features) by being trained using training data that contain musical score features of at least a part of the plurality of sound waveforms of a first musical piece and contain first acoustic features of the plurality of sound waveforms, (Mase, Pg. 847, “We used a singing voice database including 70 Japanese children’s songs… This database also includes the musical scores of these songs represented by MusicXML” and “Feature vectors consisted of spectral, F0, and vibrato feature vectors” teaching corresponding musical-score information and acoustic features for the waveforms used as training data) and acquiring information related to the characteristic distribution of the plurality of sound waveforms used for training the acoustic model (Mase, Pg. 847, “Figure 2 and 3 show the distributions of pitch in training data (10 and 60 songs). Before the pitch-shift is applied, the distributions are sparse. As the amount of data increases, the distribution has been smooth after pitch-shift”, see also Figs 2 and 3) where the information that relates to the characteristic distribution and has been acquired is displayed (Gupta, Para 0020, teaching that an absolute measure of pitch may be based on “overall pitch distribution” and that the processor may assess pitch by “producing a pitch histogram”. See also para 0068, “The various measures calculated by module 202 and 204 … may be displayed on display 208” teaching display of information calculated from the pitch distribution).
Mase does not teach where the acquiring includes acquiring history data indicating a history of the plurality of sound waveforms used for the training of the acoustic model.
However, McCusker teaches where the acquiring includes acquiring history data indicating a history of the plurality of sound waveforms used for the training of the acoustic model (Para 0031, “The AI engine may track and log data lineage associated with changes to training data serialized with a model. Data lineage may include an original training data used to train a model and subsequent changes applied to the training data” see also para 0046, teaching a genealogical tree associating the target model with “various instances of training data that have been used to train the target model” thereby teaching stored history identifying training data).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify Mase in such a way as to incorporate the teachings of McCusker in order to allow for identification of the training data to provide more information to the user (Para 0031).
Claims 15, 16, and 17 are rejected under 35 U.S.C. 103 as being unpatentable over Mase (“HMM-based singing voice synthesis system using pitch-shifted pseudo training data”, INTERSPEECH 2010, pg. 845-848, 2010) in view of Gupta (US 20220277763 A1) and McCusker (US 20220058511 A1) as above and further in view of Daido et al. (US 20210256959 A1).
Regarding claim 15, Mase modified by Gupta and McCusker teaches wherein, in the displaying, a degree is displayed based on the characteristic distribution that has been acquired (Gupta, Para 0020, teaching that an absolute measure based on “overall pitch distribution” and a pitch histogram, and Para 0068, teaching that the calculated measures “may be displayed on display 208”).
However, Mase modified by Gupta and McCusker does not teach wherein the degree is a degree of proficiency of the acoustic model with respect to musical features of the second musical piece.
However, Daido teaches wherein the degree is a degree of proficiency of the acoustic model with respect to musical features of the second musical piece (Para 0077, teaching calculation of an evaluation function representing an error between feature data generated by the synthesis model and ground truth data. Paras 0081-0084, teaching determination of whether the feature data generated by the model reached a “predetermined quality” based on the calculated error, thus teaching a degree of proficiency of the acoustic model with respect to a musical score which could be applied to a second musical score).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify Mase in such a way as to incorporate the teachings of Daido in order to allow the quality of data generated by the model to be better evaluated (Para 0081).
Regarding claim 16, Mase as modified by Gupta and McCusker teaches wherein the displaying includes evaluating a degree based on the characteristic distribution that has been acquired, and displaying the degree (Gupta, Para 0020, teaching that an absolute measure based on “overall pitch distribution” and a pitch histogram, and Para 0068, teaching that the calculated measures “may be displayed on display 208”).
Mase as modified by Gupta and McCusker does not teach wherein the displaying includes evaluating a degree of proficiency of the acoustic model with respect to the second musical piece based on a musical score of the second musical piece.
However, Daido teaches wherein the displaying includes evaluating a degree of proficiency of the acoustic model with respect to the second musical piece based on a musical score of the second musical piece (Para 0035, teaching that “the score data of the tune” may be used as data supplied to the model. Para 0077, teaching calculation of an evaluation function generated by the synthesis model and ground truth data, paras 0081-0084, teaching determination of whether the output of the model has reached a “predetermined quality” based pm the calculated error, thus teaching evaluation of a degree of proficiency of the model with respect to the musical score).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify Mase in such a way as to incorporate the teachings of Daido in order to allow the quality of data generated by the model to be better evaluated (Para 0081).
Regarding claim 17, Mase as modified by Gupta and McCusker teaches wherein the displaying includes evaluating, based on the characteristic distribution that has been acquired, a degree with respect to each section of the second musical piece, and displaying the degree for each section (Gupta, Para 0020, teaching assessment based on overall pitch distribution and a pitch histogram. See Para 0170, teaching that “Four different snippets were evaluated from each song” and referring to “individual snippet ranks” thereby teaching evaluation of a degree for individuals sections of a musical piece based on pitch distribution. See also Para 0068, teaching display of calculated measures of display 208).
Mase as modified by Gupta and McCusker does not teach wherein the displaying includes evaluating, based on a musical score of each section of the second musical piece, a degree of proficiency of the acoustic model with respect to each section of the second musical piece.
However, Daido teaches wherein the displaying includes evaluating, based on a musical score of each section of the second musical piece, a degree of proficiency of the acoustic model with respect to each section of the second musical piece (Paras 0034-0035, teaching condition data specifying pitch, phonetic data, and sound period for each note and that “the score data of the tune” may be used as the condition data supplied to the model. See Para 0126-0127, teaching evaluation of the quality of the audio generated by the model and a representation of the sound quality by an index value. See also para 0125 and 0130, teaching that a note corresponds to a time section and that audio may be generated for a “time section of a tune”, thereby teaching evaluation of the proficiency of the model with respect to individual score defined sections).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify Mase in such a way as to incorporate the teachings of Daido in order to allow the quality of the output to be evaluated by section, allowing more precise review of the output (Para 0125).
Claim 18 is rejected under 35 U.S.C. 103 as being unpatentable over Mase (“HMM-based singing voice synthesis system using pitch-shifted pseudo training data”, INTERSPEECH 2010, pg. 845-848, 2010) in view of Gupta (US 20220277763 A1) and McCusker (US 20220058511 A1) above and further in view of Takayoshi et al (JP 2015125268 A).
Regarding claim 18, Mase modified by Gupta and McCusker teaches wherein the displaying includes evaluating a degree based on a musical score of a musical piece and the characteristic distribution that has been acquired (Gupta, Para 0085, teaching evaluation of singing based on comparison of a pitch distribution with a MIDI pitch histogram representing the notes of the original piece).
Mase as modified by Gupta and McCusker does not teach wherein the displaying includes evaluating a degree of proficiency of the acoustic model with respect to each of a plurality of second musical pieces based on a plurality of musical scores, and displaying a recommendation of at least one of second musical pieces based on a degree of proficiency.
However, Takayoshi teaches wherein the displaying includes evaluating a degree of proficiency of the acoustic model with respect to each of a plurality of second musical pieces based on a plurality of musical scores (Pg. 12, teaching determination, from among a plurality of playable musical pieces, of musical pieces containing insufficient voice-synthesis data, thus teaching evaluation of each of a plurality of musical pieces based on whether the pieces correspond to insufficient data), and displaying a recommendation of at least one of second musical pieces based on a degree of proficiency (Pg. 12, teaching display of the filtering result on a music recommendation screen as recommended music, and further teaching display of the music in descending order of update rate, thus teaching recommendation of pieces based on the evaluation).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify Mase in such a way as to incorporate the teachings of Takayoshi in increase user-friendliness by recommended specific pieces (Pg. 12).
Claim 19 is rejected under 35 U.S.C. 103 as being unpatentable over Mase (“HMM-based singing voice synthesis system using pitch-shifted pseudo training data”, INTERSPEECH 2010, pg. 845-848, 2010) in view of Gupta (US 20220277763 A1) and McCusker (US 20220058511 A1) as above and further in view of Fructuoso et al. (US 20160343366 A1).
Regarding claim 19, Mase as modified by Gupta and McCusker teaches wherein the acquiring is performed for an acoustic model to acquire a characteristic distribution of the acoustic model (Mase, P. 847, “Figure 2 and 3 show the distributions of pitch in training data (10 and 60 songs)”, teaching acquisition of a characteristic distribution associated with the acoustic model).
Mase as modified by Gupta and McCusker does not teach wherein the system contains multiple models, and in the displaying, a recommendation of one or more models matching the input is displayed based on the input and information associated with the plurality of models.
However, Fructuoso teaches wherein the system contains multiple models (Para 0060, “the second neural network 220 may map each set of input acoustic features to multiple models”), and in the displaying, a recommendation of one or more models matching the input is displayed based on the input and information associated with the plurality of models (Para 0061, teaching determination of “a confidence score for each one of the multiple models” wherein the confidence score “indicates a degree of confidence in each model being the most suitable model for the given portion of textual input”. See also paras 0062 and 0063 teaching selection of one or more of the models based on the confidence scores).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify Mase in such a way as to incorporate the teachings of Fructuoso in order to improve performance by allowing for the best performing models to be selected (Para 0062).
Claim 20 is rejected under 35 U.S.C. 103 as being unpatentable over Mase (“HMM-based singing voice synthesis system using pitch-shifted pseudo training data”, INTERSPEECH 2010, pg. 845-848, 2010) in view of Gupta (US 20220277763 A1), McCusker (US 20220058511 A1) as above and further in view of Feng et al. (US 20220223127 A1).
Regarding claim 20, Mase modified by Gupta and McCusker as above in claim 14 teaches displaying the degree of proficiency evaluated in real time (Gupta, Para 0020, teaching that an absolute measure of pitch may be based on “overall pitch distribution” and that the processor may assess pitch by “producing a pitch histogram”. See also para 0068, “The various measures calculated by module 202 and 204 … may be displayed on display 208” teaching display of information calculated from the pitch distribution).
Mase modified by Gupta and McCusker does not teach wherein during execution of a generation method, in which portions of a musical score of a second musical piece are sequentially received, features of each of the portions of the musical score is subjected to real-time processing using the acoustic model, and a portion of the second acoustic features which corresponds to each of the portions of the musical score is generated in real time.
However, Feng teaches wherein during execution of a generation method, in which portions of a musical score of a second musical piece are sequentially received, features of each of the portions of the musical score is subjected to real-time processing using the acoustic model, and a portion of the second acoustic features which corresponds to each of the portions of the musical score is generated in real time (Para 0063, teaching that signal conversion block 354 extracts “frame-by-frame tonic and chords pitches from the playing music in real time” including extraction of frame-by-frame information from a MIDI stream) and a portion of the second acoustic features which corresponds to each of the portions of the musical score is generated in real time (Para 0065, teaching generation of the singing signal “frame by frame”. See also Para s 0083-0084, teaching calculation of tonic and chord pitches in real time).
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify Mase in such a way as to incorporate the teachings of Feng in order to allow for more seamless output by allowing for generation of singing output in real time (Para 0065).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MICHAEL ALAN FOSTER JR. whose telephone number is (571)272-8874. The examiner can normally be reached M - F 8:00am - 5:00pm, Alternate Fridays Off.
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, Hai Phan can be reached at (571) 272-6338. 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.
/MICHAEL A FOSTER JR/Examiner, Art Unit 2654 /Richa Sonifrank/Primary Examiner, Art Unit 2654
