An electronic musical instrument includes: a memory that stores a trained acoustic model obtained by performing machine learning on training musical score data and training singing voice data of a singer; and at least one processor, wherein the at least one processor: in accordance with a user operation on an operation element in a plurality of operation elements, inputs prescribed lyric data and pitch data corresponding to the user operation of the operation element to the trained acoustic model so as to cause the trained acoustic model to output the acoustic feature data in response to the inputted prescribed lyric data and the inputted pitch data, and digitally synthesizes and outputs inferred singing voice data that infers a singing voice of the singer on the basis of the acoustic feature data output by the trained acoustic model.

An electronic musical instrument includes: a memory that stores a machine-learning trained acoustic model mimicking voice of a singer and at least one processor. When a vocoder mode is on, prescribed lyric data and pitch data corresponding to a user operation of an operation element of the musical instrument are inputted to the trained acoustic model, and inferred singing voice data that infers a singing voice of the singer is synthesized on the basis of acoustic feature data output by the trained acoustic model and on the basis of instrument sound waveform data that are synthesized in accordance with the pitch data corresponding to the user operation of the operation element. When the vocoder mode is off, the inferred singing voice data is synthesized based on the acoustic feature data without using the sound waveform data.

An electronic musical instrument includes: a memory that stores a trained acoustic model obtained by performing machine learning on training musical score data and training singing voice data of a singer; and at least one processor, wherein the at least one processor: in accordance with a user operation on an operation element in a plurality of operation elements, inputs prescribed lyric data and pitch data corresponding to the user operation of the operation element to the trained acoustic model, and digitally synthesizes and outputs inferred singing voice data that infers a singing voice of the singer on the basis of at least a portion of acoustic feature data output by the trained acoustic model, and on the basis of instrument sound waveform data that are synthesized in accordance with the pitch data corresponding to the user operation of the operation element.

Digital signal processing and machine learning techniques can be employed in a vocal capture and performance social network to computationally generate vocal pitch tracks from a collection of vocal performances captured against a common temporal baseline such as a backing track or an original performance by a popularizing artist. In this way, crowd-sourced pitch tracks may be generated and distributed for use in subsequent karaoke-style vocal audio captures or other applications. Large numbers of performances of a song can be used to generate a pitch track. Computationally determined pitch trackings from individual audio signal encodings of the crowd-sourced vocal performance set are aggregated and processed as an observation sequence of a trained Hidden Markov Model (HMM) or other statistical model to produce an output pitch track.

Methods, systems and apparatus to characterize audio using transchromagrams are disclosed. An example apparatus includes a transchromagram generator to generate a data structure based on a set of transition matrices corresponding to a plurality of time frames of audio data, the data structure indicative of probabilities that first musical notes will transition to second musical notes, a database controller to prompt a database to store the data structure within the audio data, and a notification manager to generate, based on a comparison between query audio data and the stored data structure of the audio data, a notification identifying at least one characteristic of the query audio data.

Digital signal processing and machine learning techniques can be employed in a vocal capture and performance social network to computationally generate vocal pitch tracks from a collection of vocal performances captured against a common temporal baseline such as a backing track or an original performance by a popularizing artist. In this way, crowd-sourced pitch tracks may be generated and distributed for use in subsequent karaoke-style vocal audio captures or other applications. Large numbers of performances of a song can be used to generate a pitch track. Computationally determined pitch trackings from individual audio signal encodings of the crowd-sourced vocal performance set are aggregated and processed as an observation sequence of a trained Hidden Markov Model (HMM) or other statistical model to produce an output pitch track.

An electronic musical instrument in one aspect of the disclosure includes a keyboard, a processor and a memory that stores musical piece data that includes data of a vocal part, the vocal part including at least first and second notes together with associated first and second lyric parts that are to be successively played at the first and second timings, respectively, wherein if while a digitally synthesized first signing voice corresponding to the first note is being output, a user specifies, via keyboard, a third pitch that is different from the first and second notes prior to the arrival of the second timing, the at least one processor synthesizes a modified first singing voice having the third pitch in accordance with the data of the first lyric part, and causes the digitally synthesized modified first singing voice to be audibly output at the third timing.

An electronic musical instrument in one aspect of the disclosure includes; a plurality of operation elements to be performed by a user for respectively specifying different pitches; a memory that stores musical piece data that includes data of a vocal part, the vocal part including at least a first note with a first pitch and an associated first lyric part that are to be played at a first timing; and at least one processor, wherein if the user does not operate any of the plurality of operation elements in accordance with the first timing, the at least one processor digitally synthesizes a default first singing voice that includes the first lyric part and that has the first pitch in accordance with data of the first note stored in the memory, and causes the digitally synthesized default first singing voice to be audibly output at the first timing.

Embodiments of the present invention relate to automatically identifying structures of a music stream. A segment structure may be generated that visually indicates repeating segments of a music stream. To generate a segment structure, a feature that corresponds to a music attribute from a waveform corresponding to the music stream is extracted from a waveform, such as an input signal. Utilizing a signal segmentation algorithm, such as a Variable Markov Oracle (VMO) algorithm, a symbolized signal, such as a VMO structure, is generated. From the symbolized signal, a matrix is generated. The matrix may be, for instance, a VMO-SSM. A segment structure is then generated from the matrix. The segment structure illustrates a segmentation of the music stream and the segments that are repetitive.

The purpose of the present invention is to provide a system capable of analyzing intuitively-created improvisation performances without relying on music theories. There is provided an improvisation performance analysis system, comprising: a music information coding section 10 for analyzing and coding music data of an improvisation performer stored in a music storage medium; a tone sequence pattern extraction section 11 for extracting all of first- to n-th-order tone sequence patterns which are likely to occur as n-th Markov chains in order to perform a stochastic analysis with a Markov model using the coded music data; a pitch transition sequence extraction section 12 for obtaining a pitch transition sequence for each of the extracted tone sequence patterns; a transition probability/appearance probability calculation section 13 for using the Markov model to calculate a transition probability of each pitch transition sequence and an appearance probability of each transition sequence at each of the first- to n-th-order hierarchical levels; and an improvisation performance phrase structuring section 14 for rearranging the pitch transition probabilities at each hierarchical level based on the transition probabilities and the appearance probabilities, identifying pitch transition sequences which are statistically likely to occur and expressing the pitch transition sequences in all keys as music scores based on the twelve-tone equal temperament to thereby generate improvisation performance phrases.