System

A system comprising:

• a communication interface coupled to a packet-switched network;
• and
• circuitry configured to:
• receive request information from a terminal device via the communication interface, the request information comprising a prompt and condition information including acoustic environment parameters;
• execute natural language processing on the prompt to extract semantic parameters comprising intensity level, emotional tone classification, and duration specification, and map the semantic parameters to generation control vectors;
• execute a generative neural network model comprising embedding layers, attention layers, and waveform synthesis layers to generate acoustic waveform data conditioned on the generation control vectors, and acquire the waveform data as digital audio data;
• apply signal processing to the digital audio data comprising equalization based on the acoustic environment parameters, multi-band dynamic range compression, noise reduction, and volume normalization to generate processed audio data;
• encode the processed audio data in a format selected based on a bandwidth parameter in the condition information;
• and
• transmit the processed audio data and attribute information comprising codec identifier, bit rate, sampling rate, and channel count to the terminal device via the packet-switched network.

A method comprising:

• receiving, by circuitry coupled to a packet-switched network, request information from a terminal device, the request information comprising a prompt and condition information;
• preprocessing the prompt by executing natural language processing to convert the prompt into a token sequence and a numerical representation, and supplying the token sequence and the numerical representation as generation input data to a generative neural network model;
• executing the generative neural network model to generate an audio signal comprising acoustic waveform data based on the generation input data, and acquiring the audio signal as digital audio data;
• applying a signal processing algorithm to the digital audio data comprising at least one of filtering, dynamic range compression, equalization, noise reduction, and volume normalization to generate processed audio data;
• and
• encoding the processed audio data and transmitting response information comprising the processed audio data to the terminal device via the packet-switched network.