Abstract: Disclosed embodiments include systems and methods for distributed control of media-playback components of a webcast. In an example, a webcast presenter's device can include a webcast compositing engine for creating a webcast from a variety of data sources, a media-playback engine for playing media content items, and a message processing engine for processing messages sent from audience members to the presenter. The message processing engine can obtain the messages and parse the messages for tokens indicative of a requested media content item. The message processing engine, having identified the requested media content item, can then cause the media-playback engine to play the requested content or add it to a queue.

Abstract: Audio content episodes are received. Using machine learning, one or more audio segments of interest are identified in each of the audio content episodes based at least in part on an analysis of content included in a corresponding audio content episode. Each of the identified audio segments is associated with one or more automatically determined tags. Using machine learning, a recommended audio segment is selected for a specific user from the identified audio segments based at least in part on attributes of the specific user and the automatically determined tags of the identified audio segments. The recommended audio segment are automatically provided in an audio segment feed.

Abstract: Audio translation system includes a feature extractor and a style transfer machine learning model. The feature extractor generates for each of a plurality of source voice files one or more source voice parameters encoded as a collection of source feature vectors, and generates for each of a plurality of target voice files one or more target voice parameters encoded as a collection of target feature vectors. The style transfer machine learning model trained on the collection of source feature vectors for the plurality of source voice files and the collection of target feature vectors for the plurality of target voice files to generate a style transformed feature vector.

Abstract: An audio cancellation system includes a voice enabled computing system that is connected to an audio output device using a wired or wireless communication network. The voice enabled computing device can provide media content to a user and receive a voice command from the user. The connection between the voice enabled computing system and the audio output device introduces a time delay between the media content being generated at the voice enabled computing device and the media content being reproduced at the audio output device. The system operates to determine a calibration value adapted for the voice enabled computing system and the audio output device. The system uses the calibration value to filter the user's voice command from a recording of ambient sound including the media content, without requiring significant use of memory and computing resources.

Abstract: A system, method and computer product for combining audio tracks. In one example embodiment herein, the method comprises determining at least one music track that is musically compatible with a base music track, aligning those tracks in time, and combining the tracks. In one example embodiment herein, the tracks may be music tracks of different songs, the base music track can be an instrumental accompaniment track, and the at least one music track can be a vocal track. Also in one example embodiment herein, the determining is based on musical characteristics associated with at least one of the tracks, such as an acoustic feature vector distance between tracks, a likelihood of at least one track including a vocal component, a tempo, or musical key. Also, determining of musical compatibility can include determining at least one of a vertical musical compatibility or a horizontal musical compatibility among tracks.

Abstract: A method comprises the following steps: providing a Gaussian process variational autoencoder (GP-VAE) including a Gaussian process (GP) encoder and a neural network decoder; selecting a plurality of inducing points in a data space; generating a mapping of the plurality of inducing points in a latent space; and training the GP-VAE using a training dataset.

Abstract: A system is provided for streaming media content in a vehicle. The system includes a personal media streaming appliance system configured to connect to a media delivery system and receive media content from the media delivery system at least via a cellular network. The personal media streaming appliance system includes one or more preset buttons for playing media content associated with the preset buttons. Data about the preset buttons and the media content associated with the preset buttons can be stored in the media delivery system.

Abstract: Apparatus, methods and computer-readable medium are provided for processing wind noise. Audio input is processed by receiving an audio input. A wind noise level representative of a wind noise at the microphone array is measured using the audio input and a determination is made, based on the wind noise level, whether to perform either (i) a wind noise suppression process on the audio input on-device, or (ii) the wind noise suppression process on the audio input on-device and an audio reconstruction process incloud.

Abstract: A method includes displaying a user interface of a digital audio workstation, which includes a first region for generating a composition. The first region includes a first compositional segment that has been added to the composition by a user. Based on the first compositional segment, one or more recommended predefined compositional segments are identified and displayed in a second region. The method includes receiving the selection of a second compositional segment. The method includes adding the compositional segment to the composition.

Abstract: A training set for use during content generation is generated by applying a first machine learning process P1 to a first finite sequence s wherein s has a length Ls, to generate a first statistical model M(s). The first statistical model M(s) is sampled using a first sampling process G to generate a second finite sequence t wherein t has a length Lt. A second machine learning process P2 is applied to the second finite sequence t to generate a second statistical model M(t), wherein no substring of the second finite sequence t of length d is identical to a substring of the first finite sequence s, wherein d is a predetermined number of elements in a sequence.

Abstract: A system and method for generating dynamic playlists utilizing device co-presence proximity. In accordance with an embodiment, the method can include identifying music preferences by analyzing the listening history on a device of a user; and recognizing what devices are co-present within a set proximity for the purposes of matching such devices. In accordance with an embodiment, the system can be determine what music each individual in a group has been listening to historically, and determine when those individuals are within a certain proximity to one another, including for example analyzing the media content played on a smartphone and recognising when two smartphones are in a same location.

Abstract: Systems, apparatuses, components, methods, and techniques for location-based tagging and retrieving of media content are provided. Some embodiments include a media-playback device. The media-playback device includes a location-determining device that determines a geographical location of the media-playback device, a media-output device that plays media content items, at least one computer readable data storage device. The media-playback device operates to receive a selection of a location-based media playlist, the location-based media playlist including playlist data that identifies a set of media content items and defines a route. A geographic location of the media-playback device is determined using the location-determining device. A media content item of the location-based playlist is identified based on the determined geographic location and the route. The media-playback device plays the media content item.

Abstract: Term masking is performed by generating a time-alignment value for a plurality of units of sound in vocal audio content contained in a mixed audio track, force-aligning each of the plurality of units of sound to the vocal audio content based on the time-alignment value, thereby generating a plurality of force-aligned identifiable units of sound, identifying from the plurality of force-aligned units of sound a force-aligned unit of sound to be altered, and altering the identified force-aligned unit of sound.

Abstract: An electronic device stores a shared playback queue for the shared playback session, the shared playback queue comprising one or more media content items, including a first media content item requested by a first user. While the shared playback queue includes the first media content item, the device receives a second request, from a second user, to add a second media content item to the shared playback queue. The device determines an order of the shared playback queue based at least in part on the musical preferences indicated in a profile of a third user of the plurality of users participating in the shared playback session. The device provides, for playback in the shared playback session, the first media content item and the second media content item based on the order of the shared playback queue.

Abstract: Methods, systems and computer program products are provided for determining acoustic feature vectors of query and target items in a first vector space, and mapping the acoustic feature vectors to a second vector space having a lower dimension. The distribution of vectors in the second vector space can then be used to identify items from the same songs, and/or items that are complementary. A mapping function is trained using a machine learning algorithm, such that complementary audio items are closer in the second vector space than the first, according to a given distance metric.

Abstract: A system, method and computer product for training a neural network system. The method comprises inputting an audio signal to the system to generate plural outputs f(X, ?). The audio signal includes one or more of vocal content and/or musical instrument content, and each output f(X, ?) corresponds to a respective one of the different content types. The method also comprises comparing individual outputs f(X, ?) of the neural network system to corresponding target signals. For each compared output f(X, ?), at least one parameter of the system is adjusted to reduce a result of the comparing performed for the output f(X, ?), to train the system to estimate the different content types. In one example embodiment, the system comprises a U-Net architecture. After training, the system can estimate various different types of vocal and/or instrument components of an audio signal, depending on which type of component(s) the system is trained to estimate.

Abstract: Systems, devices, apparatuses, components, methods, and techniques for predicting user and media-playback device states are provided. Systems, devices, apparatuses, components, methods, and techniques for representing cached, user-selected, and streaming content are also provided.

Abstract: A first client device is associated with a first user hosting a shared playback session. While a first media content item from the shared playback session is being presented on a set of presentation devices, the first client device communicates with a set of observer devices for the shared playback session. The first client device receives a request to modify playback of the shared playback session from a second client device, the second client device being one observer device of the set of observer devices. In response to the request to modify playback of the shared playback session, the first client device determines an action to take with respect to the shared playback session. In response to determining the action to take with respect to the shared playback session, the first client device sends a command for the action to each of the set of presentation devices.