现代电影技术Issue(7)：24-32,9.DOI:10.3969/j.issn.1673-3215.2025.07.003
基于大语言模型(LLM)与模型上下文协议(MCP)驱动的智能音频制作系统研发
Research and development of intelligent audio production system driven by large language model(LLM)and model context protocol(MCP)
摘要
Abstract
This study responds to the practical demand for film and game audio production by introducing an intelligent au-dio-production system that addresses the tedium of conventional audio asset retrieval,and the limited automation found in typical digital audio workstation(DAW).Powered by a large language model(LLM)and its model context protocol(MCP)and integrated with REAPER's open scripting framework,ReaScript,and its Python extensions,the system enables intelli-gent audio-asset search and instruction-driven automated operations.With MCP as the foundational layer,the system ef-fectively bridges natural language interaction and automated DAW control.Experiments in the"Summer Rain on Campus"audio scene demonstrate that the system can accurately locate needed audio assets through fuzzy and associative search,and its command-based operations significantly reduce the technical threshold for audio production.Compared to tradi-tional workflows,artists can focus more on creative expression rather than complex operations.In the future,the system will further expand to support multiple types of audio development interfaces and major digital audio workstations,while continuously enriching the audio asset database,thus providing a solid technical foundation for intelligent audio production.关键词
大语言模型/模型上下文协议/影视声音制作/游戏声音制作/REAPERKey words
Large Language Model/Model Context Protocol/Film and TV Sound Production/Game Sound Production/REAPER分类
信息技术与安全科学引用本文复制引用
郑嘉庆,杨璨..基于大语言模型(LLM)与模型上下文协议(MCP)驱动的智能音频制作系统研发[J].现代电影技术,2025,(7):24-32,9.基金项目
北京市高等教育学会2024年专项攻关课题"人工智能在《交互系统设计》课程中的应用研究"(ZX202425). （ZX202425）
