Audio recording/reproduction techniques have been investigated to further the realization of highly realistic communication. However, to achieve ultra-realistic communication, a real-time acoustic rendering technique must be developed along with current recording and reproduction techniques. Acoustic simulation is a promising way to simulate, predict, and render acoustic phenomena. However, it is difficult to apply conventional computational methods to real-time acoustic rendering for large-scale sound fields. Therefore, we propose a new acoustic simulation technique to realize real-time acoustic rendering of a large-scale sound field. The proposed method, based on a lattice gas cellular automaton (LGA), is called sound-field LGA. This paper introduces details of sound-field LGA, including a physical interpretation of the simulation and an application of 1-bit coding for source signal representation. Furthermore, preliminary results are demonstrated and discussed.