search

facebookresearch / sound-spaces

A first-of-its-kind acoustic simulation platform for audio-visual embodied AI research. It supports training and evaluating multiple tasks and applications.
https://soundspaces.org
Creative Commons Attribution 4.0 International
364 stars 58 forks source link

How to generate continuous audio #124

Closed 310mo closed 1 year ago

310mo commented 1 year ago

I have a question about how to generate continuous sounds in SoundSpaces 2.0.

In sound-spaces/soundspaces/continuous_simulator.py, the function crossfade seems to be in charge of smoothly connecting the sound at one point to the sound at the next point. However, crossfade is only applied to the first 50 ms (int(0.05 * sr)). Is this a value specific to the default one-step sound length of 250 ms (config.TASK_CONFIG.SIMULATOR.STEP_TIME = 0.25)? Or can crossfade be applied as is for any step length?

ChanganVR commented 1 year ago

@310mo yes, it certainly depends on how far away the consecutive points are (not only the time) since the acoustics for different points in the space are different. I don't think there is an optimal value for the fading time and you can tune this value to suit your need and application.

310mo commented 1 year ago

Thank you very much for your reply. I close this issue.