Skip to content

Latest commit

 

History

History
164 lines (127 loc) · 4.52 KB

README.md

File metadata and controls

164 lines (127 loc) · 4.52 KB

MAGIC

[ICASSP'25] Map-Guided Few-Shot Audio-Visual Acoustics Modeling

demo

We visualize the results of eval and provide the corresponding audio files as follows.

The given observations and query location

We provide a house tour gif to help understand the 3D scene and a top-down map indicate the locations of given few-shot observations and the query.

Blue pinpoints indicate the provided viewpoints. The blue arrow represents the direction of the provided viewpoints. The green pinpoint indicates the speaker that emits the audio in the query and the pink pinpoint indicates the listener that receives the audio.

The grouth truth audio

Your browser does not support the audio element.

Few-ShotRIR predicted audio

Your browser does not support the audio element.

ours predicted audio

Your browser does not support the audio element.

demo 2

We visualize the results of eval and provide the corresponding audio files as follows.

The given observations and query location

We provide a house tour gif to help understand the 3D scene and a top-down map indicate the locations of given few-shot observations and the query.

Blue pinpoints indicate the provided viewpoints. The blue arrow represents the direction of the provided viewpoints. The green pinpoint indicates the speaker that emits the audio in the query and the pink pinpoint indicates the listener that receives the audio.

The grouth truth audio

Your browser does not support the audio element.

Few-ShotRIR predicted audio

Your browser does not support the audio element.

ours predicted audio

Your browser does not support the audio element.

The failure case

We visualize the failure results of eval as follows.

The given observations and query location

We provide a house tour gif to help understand the 3D scene and a top-down map indicate the locations of given few-shot observations and the query.

Blue pinpoints indicate the provided viewpoints. The blue arrow represents the direction of the provided viewpoints. The green pinpoint indicates the speaker that emits the audio in the query and the pink pinpoint indicates the listener that receives the audio.

The grouth truth audio

Few-ShotRIR predicted audio

ours predicted audio

The failure case 2

We visualize the failure results of eval as follows.

The given observations and query location

We provide a top-down map indicate the locations of given few-shot observations and the query.

Blue pinpoints indicate the provided viewpoints. The blue arrow represents the direction of the provided viewpoints. The green pinpoint indicates the speaker that emits the audio in the query and the pink pinpoint indicates the listener that receives the audio.

The grouth truth audio

Few-ShotRIR predicted audio

ours predicted audio