This technical report presents AudioWorldSim, an open-source platform designed to generate realistic binaural audio datasets and advance research in audio-based machine learning, particularly world models. Built as a custom extension of Meta's SoundSpaces 2.0 platform, AudioWorldSim leverages their comprehensive acoustics framework, but focuses on the automatic rollout of random agent navigations, as well as implements crucial fixes to how continuous sound is composed. AudioWorldSim is made publicly available to the research community at https://github.com/Luizerko/AudioWorldSim to facilitate reproducibility.
Figures & tables
Figure 1: Visual and auditory breakdown of an agent’s navigation. The physical trajectory along the NavMesh (left) illustrates the agent’s path from its initial state (blue point) to the sound source (red point), guided by intermediate shortest-path waypoints (green points). Green crosses mark the actual discrete spatial steps taken along this route, while the arrows depict the sequence of rotational adjustments made to align the agent’s orientation before it starts with translation again. This physical movement correlates directly with the acoustic energy captured in the left and right ear Mel spectrograms (right). Notice the interaural volume differences, such as the prominent vertical bands visible in the left ear spectrogram, when the agent rounds a corner.
Figure 2: Comparison of output acoustic features segmented by discrete agent actions. A segment of the generated Mel spectrogram (left), which prioritizes spectral frequency resolution, and a segment of the raw STFT output (right), designed to preserve phase information essential for the spatial understanding capabilities of audio-based world models.
Figure 3: Mel spectrogram comparison of continuous audio synthesis. In the original implementation (left), temporal misalignment between unequal IRs causes vertical artifact lines (clicking sounds) between some simulation steps. Our corrected AudioWorldSim implementation utilizes proper zero-padding (right), resulting in a artifact-free audio continuum.
Embodied artificial intelligence has historically very much relied on visual perception, leading to a proliferation of multiple vision-centric world models. However, this reliance fails to capture spatial understanding in its entirety and can even present vulnerabilities in environments with visual occlusions, low-light conditions, or blackouts-scenarios, where acoustic information becomes a critical alternative for spatial awareness and navigation. Despite its potential, research into realistic spatial audio and particularly the development of audio-centric world models remains sparse. In this technical report, we introduce BinauralVAE: a flexible, open-source pipeline (https://github.com/Luizerko/BinauralVAE) that explores multiple models for spatialized audio reconstruction, progressing from fundamental baselines to advanced, mathematically grounded architectures. Our approach evaluates various Variational Autoencoder architectures -- including complex-valued variants -- to learn robust latent representations of binaural signals. Developed alongside AudioWorldSim, our methodology leverages realistic acoustic data captured as a simulated robot navigates an environment. This pipeline establishes a foundation for state representation in a future audio-based world model, designed to map the direct causal connection between navigational actions and their resulting acoustic consequences, and helping to enable sound as an essential complementary modality for spatial knowledge acquisition.
Recent advances in world models have enabled increasingly realistic visual synthesis. However, these generated environments remain largely silent. Bringing sound to world models poses three core challenges: real-time generation to keep pace with interactive video streams, interactive control to respond to mid-stream sound instructions, and spatially aligned stereo to reflect scene geometry and camera motion. To address these demands, we introduce WorldSonus, an interactive video-to-audio framework designed for real-time spatial sound synthesis in world models. For real-time generation, WorldSonus employs a streaming causal autoregressive diffusion architecture that synthesizes audio chunks at a low real-time factor (RTF) of 0.41. For interactive control, we incorporate an audio-centric captioning pipeline with chunk-indexed prompt scheduling, enabling dynamic manipulation of sound events during generation. For spatial alignment, we leverage high-quality stereo supervision curated from diverse stereo and ambisonic data. Extensive experiments demonstrate that while tailored for world models, WorldSonus generalizes effectively to open-domain video-to-audio benchmarks, matching or outperforming state-of-the-art bidirectional models in both acoustic quality and spatial alignment. Project page: https://noizai.github.io/WorldSonus/
Pengjun Fang, Jingyi Fa, Kam Man Wu +9
The Hong Kong University of Science and Technology · Noiz AI · MetaX +1
Text- and image-conditioned world generators can create visually rich 3D environments, yet these worlds often remain silent or rely on soundtracks synthesized solely from text or rendered video. Although such audio can convey what should be heard, it lacks an explicit representation of where sound sources are located and how their perceived sound should vary with listener movement. We introduce Audible World Models, a training-free framework that incorporates sound into the generated world state. Starting from a text prompt, our system constructs a panoramic 3D proxy, separates it into semantic layers, identifies sound-producing foreground objects and ambient background regions, and synthesizes dry audio for each sound label. It then anchors these sources to reconstructed geometry and renders listener-dependent spatial audio using geometric acoustic propagation. By explicitly linking semantics, geometry, and sound propagation, the framework maintains persistent source locations while adapting the rendered audio to changes in listener viewpoint and motion. Experiments across 80 generated scenes demonstrate substantial gains in spatial consistency over text-, video-, and panorama-conditioned baselines, while preserving competitive semantic alignment. VLM-based assessments and human evaluations further indicate that our soundtracks are preferred for their audio-visual consistency, spatial plausibility, and motion-dependent behavior.
Duowen Chen, Jinjin He, Gouthaman KV +2
Georgia Institute of Technology · Dolby Laboratories