I don't know how diagnostic this is (?), but I'd noticed and remembered the feature from a single short recording, and some time later I successfully identified a singing Robin based on it. I mean the U-shaped hills and valleys, and glissandi in the sonogram like the ones pictured below.
Apart from that, familiarising yourself with the song by listening to it to get the hang of the general timbre and maybe pick up some oft-repeated phrases (possibly with the help of Merlin or BirdNET to confirm the ID in the field) is also a good way of learning to tell the difference, and I'm currently in the process of it.
EDIT:
whoops