what do they mean by synthesizing SOUNDS from rigid body simulation? is it like how sound interacts with an object and changes based on whats around it or in the way, like listening through a wall and all?
Staffjamesfong1
@mooreyeel In the earlier slides, modal analysis is introduced as a way to understand how objects will vibrate in response to forces applied to their surface.
When you experience sound, you are sensing vibrations in the air. So to synthesize sound, we simulate how the vibrations of objects convert into vibrations in the air, and how they eventually reach your ear.
I think that generating sound for rigid-bodies is a very interesting idea. However, the examples shown in this video are not very compelling as a realistic simulation of what sounds would be made. One reason I can think of as to why this might sound wrong is because sounds are usually composed of at least 40k samples per second distinguishable to the human ear, while rigid-body simulations generally do not need to run at anywhere near 40k timesteps per second to achieve accurate results. So maybe we are not simulating enough sound making "interactions" to produce a realistic result. Or, maybe the variety of other acoustic effects that are heard in real life but not accounted for in this simulation are the reason for the uncanniness.
what do they mean by synthesizing SOUNDS from rigid body simulation? is it like how sound interacts with an object and changes based on whats around it or in the way, like listening through a wall and all?
@mooreyeel In the earlier slides, modal analysis is introduced as a way to understand how objects will vibrate in response to forces applied to their surface.
When you experience sound, you are sensing vibrations in the air. So to synthesize sound, we simulate how the vibrations of objects convert into vibrations in the air, and how they eventually reach your ear.
The corresponding video for this slide is here: https://www.youtube.com/watch?v=-vo84AxF4R4
I think that generating sound for rigid-bodies is a very interesting idea. However, the examples shown in this video are not very compelling as a realistic simulation of what sounds would be made. One reason I can think of as to why this might sound wrong is because sounds are usually composed of at least 40k samples per second distinguishable to the human ear, while rigid-body simulations generally do not need to run at anywhere near 40k timesteps per second to achieve accurate results. So maybe we are not simulating enough sound making "interactions" to produce a realistic result. Or, maybe the variety of other acoustic effects that are heard in real life but not accounted for in this simulation are the reason for the uncanniness.