The NIP was updated to make kind 1244 for replies. I haven't implemented it yet, but will soon. Kind 1222 is still for root events.
Discussion
Why not just NIP-22 replies?
You can still filter by audio-root.
I'm totally planning audio in "normal" comments.
Because that's what nostr:nprofile1qqsyvrp9u6p0mfur9dfdru3d853tx9mdjuhkphxuxgfwmryja7zsvhqpzamhxue69uhhv6t5daezumn0wd68yvfwvdhk6tcpz9mhxue69uhkummnw3ezuamfdejj7qgwwaehxw309ahx7uewd3hkctcscpyug and nostr:nprofile1qqsrhuxx8l9ex335q7he0f09aej04zpazpl0ne2cgukyawd24mayt8gprfmhxue69uhkcmmrdd3x77pwve5kzar2v9nzucm0d5hsh2c3z4 wanted 😜
It still follows NIP-22 structure though.
Where can I read discussion on this?
https://github.com/nostr-protocol/nips/pull/1923
Remember this is strictly for audio only replies, not just a text reply that includes audio.