microphone-sound distance

[catherine-qian]

Dear authors,

Could you please let me know in the paper, how could you deal with the sound location?
You assume it co-located with the microphone?
Because RIR is related with the microphone-sound path, i didn;t find the information in the paper.
Do you randomly locate a sound?

[nikhilsinghmus]

Hi, apologies for the late response; missed this earlier!

We only have scene-level correspondance, so we focus on late reverberation (which is less sensitive to within-room position and source-mic distance) and don't use other metrics (DRR, EDT) which reflect these other kinds of variation. This is also part of the motivation for using a stochastic mapping. However, you may be interested in this very interesting recent work.

Paper quotes (from ours, describing this) in case they're of interest:
From Data Aggregation (3.1): "Although this dataset contains high variability in several reverberant parameters, e.g. early reflections and source-microphone distance, it allows us to learn characteristics of late-field reverberation."

From Limitations and Future Work (4.4): "Our dataset also contains much variation in other relevant parameters (e.g. DRR and EDT) in a way we cannot semantically connect to paired images, given the sources of our data."