Assistive device for audio processing

A predictive model in audio processing systems simplifies live audio production by generating audio control data based on context, reducing operational complexity and improving sound quality for nonprofessional operators.

GB2706196APending Publication Date: 2026-11-18AIASK LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
GB2025005830
Authority / Receiving Office
GB · GB
Patent Type
Applications
Current Assignee / Owner
Filing Date
2025-04-17
Publication Date
2026-11-18

AI Technical Summary

Technical Problem

Live audio production is difficult and inefficient, particularly for nonprofessional operators, due to repetitive operations and the need for deep knowledge of audio processing, leading to suboptimal sound quality and increased operational complexity.

Method used

A predictive model, such as a machine learning model or set of predefined rules, is used to generate audio control data based on audio processing context data, simplifying system setup and operation by providing baseline presets and adapting to changing contexts, reducing the need for operator expertise.

Benefits of technology

This approach accelerates system setup, simplifies operations, reduces cognitive load, and improves sound quality by providing consistent and efficient audio processing, even for less experienced operators.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 00000000_0000_ABST
    Figure 00000000_0000_ABST
Patent Text Reader

Abstract

An audio processing device 106a (eg. mixer, controller deck, keyboard, mouse, laptop, phone, interface or sensor) for live audio production obtains audio processing context data (eg. audio device stat
Need to check novelty before this filing date? Find Prior Art

Description

FIELD OF THE INVENTION The invention generally relates to audio processing systems, and particularly to the use of predictive models therewith. BACKGROUND TO THE INVENTION Live audio production refers to the process of applying audio effects to sounds coming from multiple instruments, vocal microphones or other audio devices (e.g. audio recorders), with two goals: (a) so that the mix of the processed sounds forms an audio track which sounds cohesive and satisfying to its audience, and (b) so that the mixed track can be rendered as physical sound through other audio devices such as loudspeakers and power amplifier systems (PA systems), in rooms with various acoustic characteristics, while sounding equally pleasing across such varying playback conditions. Executing live audio production requires audio processing devices that take audio as an input and return audio as output which is derived from the input audio by applying a desired audio effect. The effect itself, for example equalisation, compression or other sound effects, can be realised either by analogue electronic devices, or by digital computing devices able to realise the desired effect by running related software. The audio effects can be modulated by parameters which can be adjusted by an operator, for example equalisation filters’ frequencies, compressor’s ratio or reverberation time. In analogue electronics, the parameters are usually controlled via manually turning physical dials or pushing electric switches. But as live audio production becomes increasingly digital, the parameters of computer-based digital effects can be controlled remotely by data coming from other devices called controllers. As an example, MIDI controllers, currently the most widespread type of controllers, are physical devices, e.g. piano-type keyboards, beat pads or touch screens, which transform human action (e.g. pressing a key, beating a pad, touch-turning a virtual dial etc.) into corresponding digital control messages according to the MIDI standard, and the MIDI messages can then be forwarded to the effects computer in order to control the parameters of the audio processing. SUMMARY OF THE INVENTION The inventors have recognised that live audio production can be difficult and inefficient to operate. In particular, operators may lose time in repetitive operations, and nonprofessional operators may struggle to produce satisfying audio results. According to one aspect of the present disclosure there is provided a method of controlling an audio processing system, the audio processing system comprising an audio processing device, and an audio device, wherein the audio processing device is configured to process an input audio signal to generate an output audio signal and provide the output audio signal to the audio device, the method performed by a processor of a computing device and comprising: obtaining audio processing context data; processing the audio processing context data using a predictive model to generate audio control data, the audio control data defining a configuration of the audio processing device for processing the input audio signal to produce the output audio signal; and outputting the audio control data for configuring the audio processing device. The use of a predictive model for controlling an audio processing system enables accelerated system setup, and simplifies the task of operators by supplying baseline audio presets based on context and reacting to changing context during operation. Furthermore, the use of a predictive model may alleviate the need for operators to acquire deep knowledge of audio processing, thus potentially saving operators time and money, and may alleviate the operators’ cognitive load by reducing the operational complexity of the audio processing system, thereby increasing simplicity and ease of use. Finally, audio rendering quality may be improved compared to manual control, particularly in situations where an operator is less experienced, thereby providing improved and more consistent sound quality. Systems described herein therefore provide quicker, more efficient operation, increased ease of use of audio processing systems, less onerous knowledge requirements, and better sounding results. The predictive model may comprise a machine learning model. The machine learning model may comprise a deep neural network, a decision tree, a random forest, a Bayesian inference model or a Gaussian model. The predictive model may comprise a set of predefined rules. Outputting the audio control data may comprise communicating with the audio processing device to adjust a setting of the audio processing device based on the audio control data. Outputting the audio control data may comprise: generating human-readable instructions based on the audio control data; and displaying the human-readable instructions to an operator of the audio processing system. This allows human operators to control, interpret and potentially correct the control actions suggested by the automation. The audio processing context data may comprise audio processing device state data characterising a current state of the audio processing device. The audio processing context data may comprise audio device state data characterising a current state of the audio device. The audio processing system may further comprise an audio processing control device configured to generate audio processing device control data for controlling the audio processing device. The audio processing control device may comprise a keyboard, a button, a switch, a potentiometer, a slider, a mouse, a pointing device, a touch sensitive interface, or a physical sensor. The audio processing context data may comprise audio processing device control data previously generated by the audio processing control device. The audio processing context data may comprise environment data, the environment data characterising an environment of the audio processing system and / or a user of the audio processing system. The audio processing context data may comprise an audio recording and / or an audio measurement. Audio processing context data as defined above enables the predictive model to make more accurate predictions, because it reacts to a more complete description of the audio processing context than predictions based on audio alone or based on previous control commands alone. Furthermore, indications of music instrument types, user identity or desired music style may facilitate appropriate and personalised audio processing control predictions, but may be hard to guess from audio data alone. Inputting such environmental data as context data may enable more appropriate control predictions. Obtaining the audio processing context data may comprise retrieving the audio processing context data from a memory of the computing device or a data bus of the computing device. Receiving the audio processing context data may comprise receiving the audio processing context data via a network interface of the computing device. Obtaining the audio processing context data in this way facilitates more accurate predictions, because the predictive model can utilise context from multiple sources across the whole audio processing system. The audio device may comprise a microphone, a musical instrument, an amplifier, a signal processor, a physical sensor, an audio input port, an XLR plug, a % inch TRS jack, or a pre-amplifier. The method may further comprise storing the audio control data in a memory accessible to the processor. Storing the audio control data in a memory accessible to the processor means that the collected control data can be used to monitor the operating conditions and use cases of the system, e.g. for improvement, debugging or future corrective action, to improve the audio processing system overall. According to a further aspect of the present disclosure there is provided a computer program product comprising computer instructions that, when executed by a processor of a computing device, cause the processor to perform the above-described method. According to a further aspect of the present disclosure there is provided a computing device for controlling an audio processing system, the device comprising a processor configured to perform the above-described method. According to a further aspect of the present disclosure there is provided an audio processing system comprising: the computing device described above; the audio processing device; and the audio device. The audio processing system may further comprise an audio processing control device configured to generate audio processing device control data for controlling the audio processing device. It will be appreciated that the functionality of the devices we describe may be divided across several modules. Alternatively, the functionality may be provided in a single module or a processor. The or each processor may be implemented in any known suitable hardware such as a microprocessor, a Digital Signal Processing (DSP) chip, an Application Specific Integrated Circuit (ASIC), Field Programmable Gate Arrays (FPGAs), a Graphics Processing Unit (GPU), a Tensor Processing Unit (TPU), a System on Chip (SoC) etc. The, or each processor may include one or more processing cores with each core configured to perform independently. The, or each processor may have connectivity to a bus to execute instructions and process information stored in, for example, a memory. The invention further provides processor control code to implement the above-described systems and methods, for example on a general purpose computer system or on a digital signal processor (DSP). The invention also provides a carrier carrying processor control code to, when running, implement any of the above methods, in particular on a non-transitory data carrier - such as a disk, microprocessor, CD- or DVD-ROM, programmed memory such as read-only memory (Firmware), or on a data carrier such as an optical or electrical signal carrier. The code may be provided on a carrier such as a disk, a microprocessor, CD- or DVD-ROM, programmed memory such as non-volatile memory (e.g. Flash) or read-only memory (Firmware). Code (and / or data) to implement embodiments of the invention may comprise source, object or executable code in a conventional programming language (interpreted or compiled) such as C, or assembly code, code for setting up or controlling an ASIC (Application Specific Integrated Circuit) or FPGA (Field Programmable Gate Array), or code for a hardware description language such as Verilog™ or VHDL (Very high speed integrated circuit Hardware Description Language). As the skilled person will appreciate such code and / or data may be distributed between a plurality of coupled components in communication with one another. The invention may comprise a controller which includes a microprocessor, working memory and program memory coupled to one or more of the components of the system. These and other aspects will be apparent from the embodiments described in the following. The scope of the present disclosure is not intended to be limited by this summary nor to implementations that necessarily solve any or all of the disadvantages noted. BRIEF DESCRIPTION OF THE DRAWINGS For a better understanding of the present disclosure and to show how embodiments may be put into effect, reference is made to the accompanying drawings in which: Fig. 1 shows an audio processing system; Fig. 2 shows a further audio processing system; Fig. 3 shows a computing device; Fig. 4 shows a further audio processing system which is an exemplary implementation of the system of Fig. 2; Fig. 5 shows the training of a machine learning model; Fig. 6 shows an inference process using a machine learning model; and Fig. 7 shows a flow diagram for a method of controlling an audio processing system. DETAILED DESCRIPTION Embodiments will now be described byway of example only. Fig. 1 shows an audio processing system 100. The audio processing system 100 comprises an audio processing control device 102, an audio device 104, and an audio processing device 106. The three devices are connected via a network 110. The audio device 112 may optionally be associated with a user 112. Audio processing device is used herein to mean a device that is configured to process an input audio signal to generate an output audio signal and provide the output audio signal to a further audio device. For example, an audio processing device may be an audio interface, a DSP chip, and / or an audio effects processor. An audio processing device such as the audio processing device 106 may comprise one or several analogue audio input / output (I / O) ports e.g. XLR plugs, % TRS jacks, 3.5mm mini jacks, cinch connectors; and / or may have an analogue-to-digital audio converter (ADC), or a digital-to-analogue audio converter (DAC), or both; and / or may have one or several digital audio input and / or output ports e.g. AES / EBU, ADAT, and / or MADL For example, the audio processing device 106 may comprise an analogue audio I / O port such as the kinds listed above, and an ADC to digitise audio received via the analogue audio I / O port. As a further example, the audio processing device 106 may comprise a digital connection via a point-to-point cable such as AES / EBU or ADAT. As a further example, the audio processing device 106 may be networked, and may therefore comprise a network cable and networked protocol which includes packets addressing and routing. An audio processing device such as the audio processing device 106 may comprise a processor able to perform audio processing (such as digital signal processing or digitally controlled analogue signal processing); such a processor may comprise a microcontroller, a DSP chip, an FPGA, and / or a GPU. The audio processing device 106 may apply audio processing to audio samples received from the device’s own ADC (if present, for example if the audio processing device 106 comprises an analogue I / O port), and / or a data or network port of the audio processing device 106 such as, for example, a digital input and / or output port as described above if present. It may be possible to configure this audio processing remotely, for example via the network 110. For example, this may be possible via an audio processing control protocol such as MIDI, OCA / AES70, OSC, or NKS. Audio processing control device is used herein to mean a device that is configured to control an audio processing device. For example, an audio processing control device may be configured to generate audio processing device control data for controlling the audio processing device. The audio processing control device 102 may comprise human-operable controls such as, for example, a keyboard, a button, a switch, a potentiometer, a slider, a mouse, a pointing device, a touch sensitive interface, or a physical sensor; and may transmit audio processing control information when these human-operable controls are operated. The audio processing control device 102 may, for example, comprise a controller deck, an alphanumeric keyboard, a musical keyboard (e.g. similar to a piano), a beat pad, a personal computer, a tablet computer, and / or a mobile device such as a smartphone. Audio device is used herein to mean any device that relates to audio playback, audio processing, or any other aspect of handling, processing, and producing audio. For example, the audio device 104 may comprise a microphone, a musical instrument, an amplifier, a signal processor, a physical sensor, an audio input port, an XLR plug, a % inch TRS jack, or a pre-amplifier. The user 112 may be a player, operator and / or owner of the audio device 104, for example if the audio device 104 is a musical instrument that the user 112 plays, or if it is an amplification device or an audio effects unit which the user 112 is in charge of. If the audio device 104 is a microphone, the user may own the microphone and / or may intend to produce sound vocally, with an instrument, or in any other manner to be captured by the microphone. As an example of one way in which the audio processing system 100 may operate, the audio processing device 106 may receive an audio signal from the audio device 104, process the audio signal, then forward the processed audio signal back to the audio device 104, or to another audio device connected to the audio processing device 106. The audio processing control device 102 may be operated by a human operator, e.g. by turning a dial or pressing a key, to generate audio processing control data forwarded to the audio processing device 106 in order to control the parameters of the audio processing applied to the audio signal received by the audio processing device 106. The network 110 may be a wired or wireless network. For example, the network 110 may comprise connections using USB, Bluetooth, Firewire, Ethernet, Wi-Fi, 3G, 4G, 5G, 6G, or any other proprietary wired or wireless connection standard. Audio signals transmitted over the network 110 may be transmitted using an audio transmission protocol. For example, the audio transmission protocol may be an audio-over-IP (AoIP) protocol such as AVB, Dante, AES67, RAVENNA, NDI, MADI, or PCM. Audio processing control data transmitted over the network 110 may be transmitted using a control data transmission protocol. For example, the control data transmission protocol maybe MIDI, OCA / AES70, OSC, or NKS. Furthermore, while Fig. 1 shows the audio processing control device 102, audio processing device 106, and audio device 104 all connected via the network 110, the audio processing device 106 and the audio device 104 may additionally or alternatively be connected via any type of audio connection, including non-networked connections that are not via the network 110. In particular, the connection between the audio processing device 106 and the audio device 104 may be any suitable (e.g. wired) connection between I / O ports of the two devices as described above, if present. For example, if the audio processing device 106 and the audio device 104 both comprise audio I / O ports supporting an XLR connection, the connection between the audio processing device 106 and the audio device 104 may be via an XLR cable. As a further example, if the audio processing device 106 and the audio device 104 both comprise digital input and / or output ports supporting a connection via AES / EBU or ADAT, the connection between the audio processing device 106 and the audio device 104 may be via AES / EBU or ADAT. These wired connections may be in addition to a network connection via the network 110, or may be an alternative to a network connection via the network 110, in which case the audio processing device 106 and the audio device 104 may not be connected via the network 110. While Fig. 1 shows a single audio processing control device 102, audio device 104, and audio processing device 106, the audio processing system 100 may comprise more than one of any or all of these devices. Any or all of the audio processing control device 102, audio device 104, and audio processing device 106 may be integrated into a single device. For example, the audio processing system may comprise an audio processing control device 102, and a further device that integrates the functions of the audio processing device 106 and the audio device 104. For example, this can be a synthesizer which includes a keyboard, an effects processor and a loudspeaker. Or this can be a laptop which includes audio effects which are controlled by a MIDI keyboard connected to the laptop’s USB port. While Fig. 1 shows a user 112 associated with the audio device 104 only, there may be users associated with other devices of the audio processing system 100. For example, there may be a further user operating the audio processing control device 102, additionally or alternatively to the user 112. Fig. 2 shows a further audio processing system 200. The audio processing system 200 contains several like elements to the audio processing system 100 described above, with the addition of an assistive device 202. The assistive device is operable to obtain audio processing context data, process the audio processing context data using a predictive model to generate audio control data, and output the audio control data for configuring the audio processing device 106 (or, if the system 200 comprises multiple audio processing devices 106, for configuring one or more of the audio processing devices 106). It is noted that, given the presence of the assistive device 202, the audio processing control device 102 is optional in the system 200 and need not be present for the system to be implemented in practice. The audio processing context data may provide context regarding the current state and / or preferred configuration of the audio processing system 200. For example, the audio processing context data may characterise a configuration of the audio processing system 200, and / or an environment of the audio processing system 200. Generally, the audio processing context data may be of any form suitable for use by a predictive model to determine how to configure the audio processing device 106. Some or all of the audio processing context data may be received from one or more other devices of the audio processing system 200, such as the audio device 104, the audio processing device 106, and / or the audio processing control device 102 if present, via the network 110 or any alternative or additional connection between devices, for example as described above with reference to Fig. 1. In this case, the audio processing context data may be received via a network interface of the assistive device 202 (such as the network interface 306 described below with reference to Fig. 3). Additionally or alternatively, some or all of the audio processing context data may be retrieved from a memory of the assistive device 202, such as the memory 302 described below with reference to Fig. 3. The following are examples of data that may form part or all of the audio processing context data. As a first example, the audio processing context data may comprise data relating to a current configuration of one or more devices of the audio processing system. For example, the audio processing context data may comprise data describing a current state of the audio processing device 106 and / or the audio device 104. Data characterising a current state of the audio processing device 106 may be referred to herein as audio processing device state data, while data characterising a current state of the audio device 104 may be referred to herein as audio device state data. For example, the audio processing context data may comprise any current settings of the audio processing device 106 and / or the audio device 104. Additionally or alternatively, the audio processing context data may comprise any other data associated with the audio processing device 106 and / or the audio device 104, such as a network address ora type of musical instrument (if the audio device 104 is a musical instrument or is a microphone which captures an identified instrument). Additionally or alternatively, if the audio processing control device 102 is present, the audio processing context data may comprise current and / or historic control commands and / or control data used by the audio processing control device 102 for controlling the audio processing device 106. This data may be referred to herein as audio processing device control data. For example, the audio processing device control data may comprise a current or past setting of any human-operable controls of the audio processing control device 102, as described above. These settings may be reported by, for example, information encoded with MIDI and / or OCA. Additionally or alternatively, the audio processing context data may comprise data characterising an environment of the audio processing system 200 and / or a user of the audio processing system 200 (such as the user 112). This type of data is referred to generally herein as environment data. For example, the audio processing context data may comprise identification of the user 112 of the audio device 104, such as a name, a face picture, or any other associated ID. Additionally or alternatively, the audio processing context data may comprise either or both of an identification of a type of performance space (such as concert hall, opera house, underground basement, or any other type of performance space); or positions of one or more of the devices of the audio processing system 100, including positions of multiple audio devices 104 if multiple are present. Additionally or alternatively, the audio processing context data may comprise an audio recording or an audio measurement. The audio recording and / or audio measurement may be recorded by one or more audio sensors of the devices of the audio processing system 200. For example, any or all of the assistive device 202, the audio processing device 106, the audio device 104, or the audio processing control device 102 (if present) may comprise an audio sensor such as a microphone capable of capturing audio recordings and / or audio measurements. For example, if the audio device 104 is capable of audio playback, the audio processing context data may comprise a recording of audio playback from the audio device 104. In embodiments, some of the audio processing context data may be directly input by a user (such as the user 112). For example the user 112 may identify themself, their instrument, an intended genre of music, or any other relevant information as part of the environment data described above. The audio control data may be used for configuring the audio processing device 106. In particular, the audio control data may define a configuration of the audio processing device 106 for processing audio input signals to generate audio output signals. As an example, the audio control data may define a change to a setting of the audio processing device 106. Several examples of possible contents of the audio control data are now provided. As an example, if the audio processing context data comprises an identification of the user 112 of the audio device 104, the audio control data may comprise changes to settings of the audio processing device 106 based on known past preferences of a specific user 112. For example, a particular guitar player may prefer their guitar audio to be processed in a particular way. As a further example, if the audio processing context data comprises an audio recording indicating that audio playback from the audio device 104 is currently experiencing an echo, the audio control data may comprise changes to settings of the audio processing device 106 designed to reduce the echo. If the audio device 104 is a musical instrument, the assistive device 202 may additionally or alternatively identify the type of musical instrument from such a recording. As a further example, if the audio processing context data comprises an intended genre of music (such as rock, jazz, folk, or any other genre) the audio control data may comprise changes to settings of the audio processing device 106 to arrive at settings commonly used in that genre. As one example, rock music may be associated with generally increased loudness or a more compressed dynamic range, or with particular guitar effects if the context data further indicates that the processed audio comes from a guitar. As a further example, if the audio processing context data comprises an arrangement of several audio devices 104 on a stage in a particular performance space, the audio control data may comprise changes to settings of the audio processing device 106 to best allow playback from each audio device 104 to be best heard by an audience in the performance space or to avoid acoustic interferences between the audio devices. The assistive device 202 may be configured to implement the audio control data directly, by communicating with the audio processing device 106 to adjust a setting of the audio processing device 106 based on the audio control data. Additionally or alternatively, the assistive device 202 may instruct a human operator (such as the user 112 or another operator such as a sound technician) to implement the audio control data. In this case, the assistive device 202 may generate human-readable instructions based on the audio control data and display the human-readable instructions to the operator. The human-readable instructions may comprise text, pictures, symbols, synthetic vocal instructions, or any other suitable way of communicating information. Additionally or alternatively, the assistive device 202 may store the audio control data in a memory accessible to a processor of the assistive device, such as the memory 302 described below with reference to Fig. 3. The predictive model may comprise a human-produced computer program comprising a set of predefined rules for producing the audio control data based on the audio processing context data. For example, these predefined rules may be handwritten by a user (such as the user 112), a sound engineer, a programmer, or anyone else. Additionally or alternatively, the predictive model may comprise an artificial intelligence (Al) / machine learning (ML) model such as a deep neural network, a decision tree, a random forest, a Bayesian inference model, a Gaussian model, or any other suitable AI / ML model. The details of training and inference of the AI / ML model are described in more detail with regard to Figs. 5 and 6 below. Any or all of the audio processing control device 102, the audio device 104, the audio processing device 106, and the assistive device 202 may be integrated into a single device. For example, the audio processing system may comprise an audio processing control device 102, and a further device that integrates the functions of the audio processing device 106, the audio device 104, and / or the assistive device 202. In this example, some or all of the audio processing context data may be received via a data bus of the assistive device 202 (or, in other words, a data bus of the further device that comprises the assistive device 202). One example of such an integration of devices is a smart instrument (such as, for example, a synthesizer) that functions as the audio device 104, while including an integrated digital audio effects unit (functioning as the audio processing device 106) and a further processor configured to perform the function of the assistive device 202. Fig. 3 shows an example computing device 300. The example computing device 300 may, for example, represent an implementation of the assistive device 202 described above, and / or any or all of the audio device 104, the audio processing device 106, or the audio processing control device 102 (if present) of the audio processing system 200. The computing device 300 comprises a memory 302, a processor 304, and a network interface 306. The processor 304 is communicatively connected with the memory 302 and the network interface 306. The processor 304 may be a processing apparatus comprising one or more processors, e.g. one or more central processing units (CPUs), accelerator processors, application specific processors field programmable gate arrays (FPGAs) and / or graphics processing units (GPUs), and other equipment such as application specific integrated circuits (ASICs). The memory 302 may comprise computer-readable storage in the form of a non-transitory computer-readable medium or media. The memory 302 may comprise one or more memory units employing one or more memory media, e.g. a magnetic medium such as a hard disk; an electronic medium such as a solid-state drive (SSD), flash memory or EEPROM; and / or an optical medium such as an optical disk drive. The network interface allows the computing device 300 to receive and transmit data, for example via a network such as the network 110. In embodiments where one or more devices of the audio processing system 200 are integrated into a single device - for example, wherein the assistive device 202 is integrated with the audio processing device 106 and / or the audio device 104 - the integrated devices may be implemented as a device similar to the computing device 300 but with multiple processors 304, each processor implementing the functionality of a respective one of the integrated devices. Fig. 4 shows an audio processing system 400 which is an exemplary implementation of the audio processing system 200. The audio processing system 400 comprises multiple audio processing control devices 102a, 102b, 102c; multiple audio processing devices 106a, 106b, 106c; and multiple audio devices 104a, 104b, 104c, each associated with a respective audio processing device 106a, 106b, 106c. The description of the audio processing device 106 of the system 200 applies to each of the audio processing devices 106a, 106b, 106c of the system 400, and likewise for audio devices and audio processing control devices. In the system 400, the assistive device 202 comprises a memory 302 as described above, a predictive model 406 (for example, executing on the processor 304 of the assistive device 202), and a network interface 306 as described above. In the example of Fig. 4, the assistive device 202 receives respective data 412a, 412b, 412c from each of the audio processing devices 106a, 106b, 106c via the network 110. The data 412a, 412b, 412c may, for example, comprise audio processing device state data for each of the audio processing devices 106a, 106b, 106c. The data 412a, 412b, 412c may then form part of the audio processing context data 404 stored I the memory 302 of the assistive device 302. As described herein, the assistive device 406 may process the audio processing context data 404 using the predictive model 406 to generate audio control data 408. In the particular example of Fig. 4, the audio control data 408 is applied directly by the assistive device 202 to communicate with the audio processing devices 106a, 106b, 106c and change settings thereof. This in turn affects the processing of input audio signals by the audio processing devices 106a, 106b, 106c to generate output audio signals that are provided to the respective audio devices 104a, 104b, 104c. Fig. 5 shows the training of a machine learning (ML) model 504. Once trained, the ML model 504 may be used as the predictive model described above with reference to Fig. 2, in the case that the predictive model is a machine learning model. The ML model 504 may be trained using training data drawn from a database 502 of audio processing context data (referred to otherwise herein as audio processing context sequences) paired with audio control data (audio processing control sequences). The audio processing context data and audio control data stored in the database 502 may contain any of the exemplary types of data described above with reference to Fig. 2. At step S506, the training of the ML model 504 may comprise deleting parts of some or all of the audio processing control sequences. At step S508, the ML model 504 may be trained to predict the missing part of the audio processing control sequences. The training of the ML model 504 at step S508 may be accomplished by any suitable known method, such as supervised, semi-supervised, active, contrastive or reinforcement learning. As an example, in a supervised training approach, pairs of audio processing context data and audio control data may be extracted from database 502 and grouped into a training set, a development set and a test set. A fully connected deep neural network model or a convolutive neural network model may be specified in terms of several matrix multiplication layers and associated gating functions. A loss function such as the mean squared error (MSE) between training control data and predicted control data may be chosen and differentiated with regards to the model’s parameters. Stochastic gradient descent may be applied, for example based on an optimiser such as the ADAM optimiser, to iteratively adapt the model’s parameters and reduce the loss over the training set and the validation set. The training may be stopped once the loss stops decreasing on the validation set. As a further example, a bidirectional encoder representation from transformers (BERT) model may be defined and pre-trained to reconstruct control data which was deleted from the training sequences, similarly to the masked language modelling technique used to train text predictors. The encoding space may learn an implicit space, a.k.a. embedding, which may encode and generalise the relationship between the control data and the context data. The model may then be fine-tuned on specific audio control tasks which may have fewer training examples, similarly to language modelling techniques, so that it learns to predict the control given the context on more specific control tasks, for example adjusting the relative gains of music instrument sounds input to the audio processing system. Fig. 6 shows an exemplary inference process using the ML model 504. The process shown in Fig. 6 may, for example, be implemented by the processor 304 of the assistive device 202 using the ML model 504. In the example of Fig. 6, the ML model 504 may be used at step S606 to process audio processing context data 602 to generate audio control data. In particular, the ML model 504 may be used to process the audio processing context data 602 to predict a next sequence of audio processing control vectors 604 which serves as the audio control data described above with reference to Fig. 2. At step S608, the processor 304 of the assistive device 202 may then output the audio control data for configuring the audio processing device 106, either by applying the predicted audio processing control sequence 604 to the audio processing being carried out by the audio processing device 106, or by instructing a human operator (such as the user 112 or another human operator) that they should apply the predicted audio processing control sequence 604. Fig. 7 shows a method 700 of controlling an audio processing system. The method 700 may be performed by a processor 304 of the assistive device 204 described above for controlling the audio processing system 200. At step S702, the method comprises obtaining audio processing context data. For example, the processor 304 of the assistive device 202 may obtain audio processing context data. The audio processing context data may be retrieved from the memory 302 of the assistive device 202 and / or received via the network 110 from other devices of the audio processing system 200. At step S704, the method comprises processing the audio processing context data using a predictive model to generate audio control data, the audio control data defining a configuration of the audio processing device for processing the input audio signal to produce the output audio signal. For example, the processor 304 of the assistive device 202 may use a predictive model such as the predictive model 406 to process to the audio processing context data and generate audio control data that defines a configuration of the audio processing device 106. At step S706, the method comprises outputting the audio control data for configuring the 5 audio processing device. For example, the processor 304 of the assistive device 202 may output the audio control data for configuring the audio processing device 106 as described above with reference to Fig. 2. 10 Although the subject matter has been described in language specific to structural features and / or methodological acts, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are disclosed as example 15 forms of implementing the claims.

Claims

1. A method of controlling an audio processing system, the audio processing system comprising an audio processing device, and an audio device, wherein the audio processing device is configured to process an input audio signal to generate an output audio signal and provide the output audio signal to the audio device, the method performed by a processor of a computing device and comprising:obtaining audio processing context data;processing the audio processing context data using a predictive model to generate audio control data, the audio control data defining a configuration of the audio processing device for processing the input audio signal to produce the output audio signal; andoutputting the audio control data for configuring the audio processing device.

2. The method of any preceding claim, wherein the predictive model comprises a machine learning model.

3. The method of claim 2, wherein the machine learning model comprises a deep neural network, a decision tree, a random forest, a Bayesian inference model or a Gaussian model.

4. The method of any preceding claim, wherein the predictive model comprises a set of predefined rules.

5. The method of any preceding claim, wherein outputting the audio control data comprises communicating with the audio processing device to adjust a setting of the audio processing device based on the audio control data.

6. The method of any preceding claim, wherein outputting the audio control data comprises:generating human-readable instructions based on the audio control data; and displaying the human-readable instructions to an operator of the audio processing system.

7. The method of any preceding claim, wherein the audio processing context data comprises audio processing device state data characterising a current state of the audio processing device.

8. The method of any preceding claim, wherein the audio processing context data comprises audio device state data characterising a current state of the audio device.

9. The method of any preceding claim, wherein the audio processing system further comprises an audio processing control device configured to generate audio processing device control data for controlling the audio processing device, and wherein the audio processing context data comprises audio processing device control data previously generated by the audio processing control device.

10. The method of claim 9, wherein the audio processing control device comprises a keyboard, a button, a switch, a potentiometer, a slider, a mouse, a pointing device, a touch sensitive interface, or a physical sensor.

11. The method of any preceding claim, wherein the audio processing context data comprises environment data, the environment data characterising an environment of the audio processing system or a user of the audio processing system.

12. The method of any preceding claim, wherein the audio processing context data comprises an audio recording or an audio measurement.

13. The method of any preceding claim, wherein obtaining the audio processing context data comprises retrieving the audio processing context data from a memory of the computing device or a data bus of the computing device.

14. The method of any preceding claim, wherein receiving the audio processing context data comprises receiving the audio processing context data via a network interface of the computing device.

15. The method of any preceding claim, wherein the audio device comprises a microphone, a musical instrument, an amplifier, a signal processor, a physical sensor, an audio input port, an XLR plug, a % inch TRS jack, or a pre-amplifier.

16. The method of any preceding claim, further comprising storing the audio control data in a memory accessible to the processor.5 17. A computer program product comprising computer instructions that, when executedby a processor of a computing device, cause the processor to perform the method of any of claims 1 to 16.

18. A computing device for controlling an audio processing system, the device 10 comprising a processor configured to perform the method of any of claims 1 to 16.

19. An audio processing system comprising:the computing device of claim 18;the audio processing device; and15 the audio device.

20. The audio processing system of claim 19, further comprising an audio processing control device configured to generate audio processing device control data for controlling the audio processing device.

Citation Information

Patent Citations

  • ViewUS2022/0385991A1onEspacenetopensinnewtab

  • ViewUS2025299657A1onEspacenetopensinnewtab