Abstract: An image rendering method for rendering a pixel of a virtual scene at a viewpoint includes: downloading a machine learning system corresponding to a current or anticipated state an application determining the virtual scene to be rendered from among a plurality of machine learning systems corresponding to a plurality of states of the application; providing a position and a direction based on the viewpoint to the machine learning system previously trained to predict a factor; combining the predicted factor from the machine learning system with a distribution function that characterises an interaction of light with a predetermined surface to generate the pixel value corresponding to an illuminated first element of the virtual scene at the position; and incorporating the pixel value into a rendered image for display.

Abstract: An audio personalisation method for a first user includes: testing a first user on a calibration test, the calibration test comprising requiring a user to match a test sound to a test location, either by controlling the position of the presented sound or controlling the position of the presented location, for a sequence of test matches, each test sound being presented at a position using a default head related transfer function ‘HRTF’, receiving an estimate of each matching location from the first user, and calculating a respective error for each estimate, to generate a sequence of location estimate errors for the first user; and comparing at least some of the location estimate errors for the first user with estimate errors of the same locations previously generated for at least a subset of a corpus of reference individuals; identifying a reference individual with the closest match of compared location estimation errors to those of the first user; and using an HRTF, previously obtained for the identified refe

Abstract: A data processing apparatus adapted to generate a recording for a content includes: input circuitry configured to obtain image data and corresponding audio data for the content, the image data including a plurality of image frames, a plurality of machine learning models including one or more first machine learning models and one or more second machine learning models, each first machine learning model configured to obtain at least a part of the image data and trained to output respective indicator data for the plurality of image frames, each second machine learning model configured to obtain at least a part of the audio data and trained to output respective indicator data for the plurality of image frames corresponding to the at least part of the audio data, where the respective indicator data associated with an image frame is indicative of an interest score for the image frame, authoring circuitry configured to select one or more groups of image frames from the plurality of image frames in dependence upon re

Abstract: An audio personalisation method for a user includes: testing the user with an audio test, the audio test comprising: moving a portable device, comprising a position tracking mechanism and a speaker, to a plurality of predetermined test positions relative to the user's head; when within a threshold distance of a predetermined test position, playing a test sound through the speaker of the portable device; and detecting the test sound using a microphone proximate to each of the user's ears, and associating resulting measurement data with the predetermined test position, where the resulting measurement data derived from the detected test sounds is characteristic of the user's head morphology; for a corpus of reference individuals for whom respective HRTFs have been generated, comparing the measurement data from the user's audio test or a head-related transfer function ‘HRTF’ derived from this measurement data with corresponding measurement data of some or all respective reference individuals in the corpus or HRTF

Abstract: An audio personalisation method for a user, to reproduce an area-based or volumetric sound source, includes the steps of, for a head related transfer function ‘HRTF’ associated with the user, smoothing HRTF coefficients relating to peaks and notches in the HRTF's spectral response, responsive to the size of the area or volume of the sound source; filtering the sound source using the smoothed HRTF for the notional position of the sound source; and outputting the filtered sound source signal for playback to the user.

Abstract: A method of video editing includes: receiving data identifying a videogame, and data representative of one or more in-game events, analysing one or more aspects of the in-game events to generate at least a first statistical model for identifying statistically significant in-game events, identifying whether a current in-game event is statistically significant based on a generated statistical model and if so, extracting video of the event from a video buffer, and evaluating whether to include the extracted video in a themed highlight reel and if so, including the extracted video in the themed highlight reel.

Abstract: A system for modifying interactions between a user and displayed content, the system comprising an input detection unit operable to detect one or more inputs from a peripheral device operated by a user, an input characterisation unit operable to characterise the detected one or more inputs, and a content modification unit operable to modify one or more aspects of the user’s interaction in dependence upon the characterisation of the detected inputs.

Abstract: A system for generating audio content in dependence upon an input audio track comprising audio corresponding to one or more sound sources, the system comprising an audio input unit operable to input the input audio track to one or more models, each representing one or more of the sound sources, and an audio generation unit operable to generate, using the one or more models, one or more audio tracks each comprising a representation of the audio contribution of the corresponding sound sources of the input audio track, wherein the generated audio tracks comprise one or more variations relative to the corresponding portion of the input audio track.

Abstract: A method of displaying streamed content includes receiving streamed content from a server, decoding the streamed content, outputting the decoded streamed content for display, and estimating a degree of attention of a user on the streamed content, where the decoding step comprises decoding the content to a quality that is dependent upon the estimated degree of attention of the use.

Abstract: A content generation system operable to generate one or more actions to be performed by an agent, the system comprising an input receiving unit operable to receive two or more actions for the agent, a model generation unit operable to input the actions to a machine learning model so as to generate a trained machine learning model, and an action generation unit operable to generate an action to be performed by the agent, wherein the generation comprises the selection of a latent space interpolation state associated with the trained machine learning model.

Abstract: A content generation system operable to generate one or more actions to be performed by an agent, the system comprising an input receiving unit operable to receive information defining an input action, the input action being an action associated with the agent, a constraint identifying unit operable to identify one or more constraints associated with the input action, and an action generation unit operable to generate, using a machine learning model, one or more actions in dependence upon the information defining the input action and the identified constraints wherein the one or more actions are variations of the defined action.

Abstract: A method of obtaining a head-related transfer function for a user is provided. The method comprises generating an audio signal for output by a handheld device and outputting the generated audio signal at a plurality of locations by moving the handheld device to those locations. The audio output by the handheld device is detected at left-ear and right-ear microphones. A pose of the handheld device relative to the user's head is determined for at least some of the locations. One or more personalised HRTF features are then determined based on the detected audio and corresponding determined poses of the handheld device. The one or more personalised HRTF features are then mapped to a higher-quality HRTF for the user, wherein the higher-quality HRTF corresponds to an HRTF measured in an anechoic environment. This mapping may be learned using machine learning, for example. A corresponding system is also provided.

Abstract: A system for determining the impulse response of an environment, the system comprising an audio emitting unit operable to emit a predetermined sound in the environment, an audio detection unit operable to record the sound output by the audio emitting unit, and an impulse response generation unit operable to identify an impulse response of the environment in dependence upon a frequency response of the audio emitting unit and/or the audio detection unit, and a difference between the predetermined sound and the recorded sound.

Abstract: An image rendering method for an entertainment device for rendering a pixel at a viewpoint includes: for a first element of a virtual scene, having a predetermined surface at a position within that scene, obtaining a machine learning system previously trained to predict a factor that, when combined with a distribution function that characterises an interaction of light with the predetermined surface, generates a pixel value corresponding to the first element of the virtual scene as illuminated at the position; providing the position and a direction based on the viewpoint to the machine learning system; combining the predicted factor from the machine learning system with the distribution function to generate the pixel value corresponding to the illuminated first element of the virtual scene at the position; and incorporating the pixel value into a rendered image for display; where the obtaining step comprises: identifying a current or anticipated state of an application determining the virtual scene to be rend

Abstract: An image rendering method for rendering a pixel at a viewpoint includes, for a first element of a virtual scene, having a predetermined surface at a position within that scene; providing the position and a direction based on the viewpoint to a machine learning system previously trained to predict a factor that, when combined with a distribution function that characterises an interaction of light with the predetermined surface, generates a pixel value corresponding to the first element of the virtual scene as illuminated at the position, combining the predicted factor from the machine learning system with the distribution function to generate the pixel value corresponding to the illuminated first element of the virtual scene at the position, and incorporating the pixel value into a rendered image for display, where the machine learning system was previously trained with a training set based on images comprising multiple lighting conditions.

Abstract: A system for determining prioritisation values for two or more sounds within an audio clip includes: a feature extraction unit operable to extract characteristic features from the two or more sounds, a feature combination unit operable to generate a combined mix comprising extracted features from the two or more sounds, an audio assessment unit operable to identify the contribution of one or more of the features to the combined mix, a feature classification unit operable to assign a saliency score to each of the features in the combined mix, and an audio prioritisation unit operable to determine relative priority values for the two or more sounds in dependence upon the assigned saliency scores for each of one or more features of the sounds.

Abstract: Audio communication apparatus comprises a set of two or more audio communication nodes; each audio communication node comprising: an audio encoder controlled by encoding parameters to generate encoded audio data to represent a vocal input generated by a user of that audio communication node, the encoded data being agnostic to which user who generated the vocal input; and an audio decoder controlled by decoding parameters to generate a decoded audio signal as a reproduction of a vocal signal generated by a user of another of the audio communication nodes, the decoding parameters being specific to the user of that other of the audio communication nodes.

Abstract: A system for training a model to select actions to be taken by an agent within an environment, the system including: a state determination unit operable to determine a state of the environment, a latency determination unit operable to determine a latency associated with interactions between the agent and the environment, an action determination unit operable to determine one or more actions to be performed by the agent in dependence upon the state, wherein actions are determined for each of one or more latencies determined by the latency determination unit, an action evaluation unit operable to evaluate the success of each of the actions, and a generation unit operable to generate the model in dependence upon identifying correlations between the success of each of the actions and the determined latency associated with those actions, so as to identify an action to be taken by the agent in dependence upon both a determined state and a latency.

Abstract: An audio personalisation method for a user, to reproduce an area-based or volumetric sound source, includes the steps of, for a head related transfer function ‘HRTF’ associated with the user, smoothing HRTF coefficients relating to peaks and notches in the HRTF's spectral response, responsive to the size of the area or volume of the sound source; filtering the sound source using the smoothed HRTF for the notional position of the sound source; and outputting the filtered sound source signal for playback to the user.

Abstract: An audio personalisation method for a user includes: testing the user with an audio test, the audio test comprising: moving a portable device, comprising a position tracking mechanism and a speaker, to a plurality of predetermined test positions relative to the user's head; when within a threshold distance of a predetermined test position, playing a test sound through the speaker of the portable device; and detecting the test sound using a microphone proximate to each of the user's ears, and associating resulting measurement data with the predetermined test position, where the resulting measurement data derived from the detected test sounds is characteristic of the user's head morphology; for a corpus of reference individuals for whom respective HRTFs have been generated, comparing the measurement data from the user's audio test or a head-related transfer function ‘HRTF’ derived from this measurement data with corresponding measurement data of some or all respective reference individuals in the corpus or HRTF