Scholarlink

<p>Against the backdrop of the integrated development of digital music creation and Internet of Things (IoT) technology, traditional automatic composition techniques face challenges such as insufficient personalization, delayed multi-device collaborative response, and poor adaptability to musical styles. This study proposes an IoT-supported automatic composition technology incorporating long short-term memory (LSTM) networks, establishing an integrated framework of “multi-source perception–feature modeling–style generation–collaborative output.” The technology collects multimodal data, including ambient sounds and user physiological feedback, through IoT devices, enables intelligent generation of melody, harmony, and rhythm using an improved LSTM network, and achieves real-time multi-device music output via IoT collaboration protocols. Validated on a multimodal dataset comprising 15,000 multi-style musical works, 8,000 environmental audio samples, and 3,000 sets of user physiological feedback, the system achieves a melodic coherence F1-score of 93.7% (a 9.2% improvement over conventional models) and a style matching accuracy of 90.5% (a 7.3% improvement), while maintaining IoT device collaborative response latency below 50 ms. Educational and creative experiments demonstrate that users of this technology exhibit improvements of 22.3% in work innovation, 26.8% in style diversity, and 31.5% in creative efficiency, significantly outperforming traditional automated composition tools. The research contributions include proposing an IoT-LSTM collaborative automated composition framework, innovatively designing multimodal perception and intelligent generation modules, empirically validating the technical application value, and providing a feasible pathway for intelligent and scenario-adaptive development of digital music creation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Application of automatic music composition under Internet of Things and LSTM networks

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月22日

查看全文
  • Zhen Zeng

摘要

Against the backdrop of the integrated development of digital music creation and Internet of Things (IoT) technology, traditional automatic composition techniques face challenges such as insufficient personalization, delayed multi-device collaborative response, and poor adaptability to musical styles. This study proposes an IoT-supported automatic composition technology incorporating long short-term memory (LSTM) networks, establishing an integrated framework of “multi-source perception–feature modeling–style generation–collaborative output.” The technology collects multimodal data, including ambient sounds and user physiological feedback, through IoT devices, enables intelligent generation of melody, harmony, and rhythm using an improved LSTM network, and achieves real-time multi-device music output via IoT collaboration protocols. Validated on a multimodal dataset comprising 15,000 multi-style musical works, 8,000 environmental audio samples, and 3,000 sets of user physiological feedback, the system achieves a melodic coherence F1-score of 93.7% (a 9.2% improvement over conventional models) and a style matching accuracy of 90.5% (a 7.3% improvement), while maintaining IoT device collaborative response latency below 50 ms. Educational and creative experiments demonstrate that users of this technology exhibit improvements of 22.3% in work innovation, 26.8% in style diversity, and 31.5% in creative efficiency, significantly outperforming traditional automated composition tools. The research contributions include proposing an IoT-LSTM collaborative automated composition framework, innovatively designing multimodal perception and intelligent generation modules, empirically validating the technical application value, and providing a feasible pathway for intelligent and scenario-adaptive development of digital music creation.