Basically, I primarily want us to build up a standard collection of room tones as you’d expect: I’d say anything between a minute and a half to five minutes is a sufficient duration. Recordings need to be neutral, in the same way as was required for the City Skylines collection – this is the indoor equivalent. No bumps, bangs, voices, car horns, etc – just neutral indoor air to be used as a sound bed in a scene. There can be a sense of distant traffic but once again, as with the City Skylines collection, it must be a wash rather than contain any specific vehicle details. Level-wise, bear in mind what volume these recordings will be used at – you don’t want to have to pull the volume down loads on the recording every time you use it, but you also don’t want to have to boost it loads either. Think about the volume it plays back at when played at unity. Ideally, when prepping a scene, I like to have to pull atmos beds like room tone down by about 5-10 dB from their recorded level.
Other than that, all I can add is – be as adventurous or as unimaginative as you like! If you can get recordings of unusual spaces then that is fantastic but even if you just get a recording of your living room, bathroom or kitchen then that’s still really useful – all spaces vary slightly in character plus the particular way you record it will give it it’s own feel too (e.g. wider or narrower stereo image, closer to the window, etc. plus everyone’s own ‘outside world’ inevitably influences the indoor tone even when it’s very quiet).
I’ll re-emphasize an important point though: make sure your room tones are quiet! No-one wants to have to go through tracks editing out bumps and bangs, tap drips, car passes, etc. – make sure these recordings are nice and clean! I’ll also add that this is not the ideal theme to use handheld recorders or cheaper equipment for, due to the inherent hiss that will be likely to show up in these quieter recordings.