ICASSP 2026 (oral paper)

Multimodal Room Impulse Response Generation Through Latent Rectified Flow Matching

Figure: Multimodal Room Impulse Response Generation Through Latent Rectified Flow Matching

Ali Vosoughi, Yongyi Zang, Qihui Yang, Nathan Paek, Randal Leistikow, Chenliang Xu

Multimodal room impulse response generation with latent rectified flow matching at 48 kHz (PromptReverb).

My part. Built the RIR preprocessing and dataset-curation pipelines (48 kHz resampling, 5-second standardization, peak alignment, RMS normalization, high-pass filtering, quality heuristics, RT60/DRR-oriented acoustic analysis) and trained and evaluated beta-VAE-style full-band RIR reconstruction models; contributed to the rectified-flow, DiT and prompt stack.

Paper Project page

Multimodal room impulse response generation with latent rectified flow matching at 48 kHz (PromptReverb).

All publications