Voice Communication Analysis in Esports
Aymeric Vinot, Nicolas Perez
TL;DR
This work investigates how voice communication affects team performance in esports, focusing on League of Legends. It proposes NLP-based metrics to quantify two critical issues: duplicate communications and parasite (unclear/irrelevant) communications, using semantic sentence embeddings and fixed-threshold detection. The authors implement an audio-processing pipeline that transcribes speech with Whisper, diarizes speakers, and aligns text to timestamps, then validate their methods through experiments showing that a general-purpose embedding model ($\text{mxbai-embed-large-v1}$) yields the most balanced performance across tasks, while acknowledging jargon-related limitations. Practical contributions include tools for coaches to profile and improve team communication, with a clear path toward specialized models and multimodal analyses to better correlate communication quality with in-game outcomes.
Abstract
In most team-based esports, voice communications are prominent in the team efficiency and synergy. In fact it has been observed that not only the skill aspect of the team but also the team effective voice communication comes into play when trying to have good performance in official matches. With the recent emergence of LLM (Large Language Models) tools regarding NLP (Natural Language Processing) (Vaswani et. al.), we decided to try applying them in order to have a better understanding on how to improve the effectiveness of the voice communications. In this paper the study has been made through the prism of League of Legends esport. However the main concepts and ideas can be easily applicable in any other team related esports.