Generally speaking, the greater variety of source material and the more atomic and separated it is, the more options and flexibility you will have when creating an Atmos mix. 

As an example, if all of the drums and percussion instruments in your project are mixed into a single stem, you will only be able to create a single object or bed, which limits your creative choices for positioning in relation to your listener. If you had individual tracks for kick, snare, hit hat, chimes, cowbell etc. you could potentially position them in disparate parts of the room to create a sense of space.

More suggestions on preparing source content for Atmos mixing:

  • Using mono stems, rather than stereo stems, will allow for more control of object position.
  • Consider printing out a wide variety of tracks (With and without effects, mono, stereo, bussed etc.) and experimenting in your mix.
    • Printing effects separately from the dry sound can allow you to place the dry source in a different position from the effects, and also allow you to apply different binaural render mode settings for the dry source and effects. This can be especially effective when working with vocals
    • If only stereo stems or stems with effects included are available consider using mid/side processing and position the 'mid' and 'side' separately
    • The binaural render mode settings apply a distance model to the objects for which they are selected. You may want to try these with source content free from other spatial effects (reverb, etc) to see how the settings change an object's sound.
  • The renderer works at 48 and 96kHz. Using source material with sample rates at least this high will yield better results than material that is upsampled.