I found that each bytes volume can be very different from other bytes in the same feed, which would mean constantly adjusting the volume for each. It would help to have an app wide normalization switch in the settings
From a creator stand point I think it could be something like toggle on the screen where you add your caption. By default the toggle would be on. A user would have to toggle the nose leveling off of they’re doing something specific like music or ASMR or something.