Speech to Birdsong Conversion

Great work on this. Hopefully you are able to crack a code that finally allows us to participate in the conversation.

It has always intrigued me that animals recognize when humans are in the area and in general designate a member of the group to monitor us to determine our intentions. I feel like they are experts at human gesture recognition, able to detect intent just from the context of the motions and that some monitor whether we have anything in our hands that could be used against them.

It's hilarious that mockingbirds and other birds will dive bomb people wandering too close to a nest and it's entertaining to think that part of their vocalizing is probably just a series of warnings, epithets and insults as they are forced to take action against the oblivious intruder who bumbled into their territory.

Vocalizations by any animal are just band-limited signals. If we can figure out which frequency components carry the information and derive information about the encoding then we can decipher it.

Sounds produced by birds can be divided into two types: songs and calls. Songs are primarily by males to attract mates, and are sung the same song over and over with minimal variation. Females, in turn, are attracted to the males with highest song stereotypy that sound closest to the "ideal" for the species. Calls are used to communicate to other birds of a certain idea, e.g. a predator nearby, and are also consequently standardized.

So while this is indeed very cool, it's my suspicion it won't be recognized as "authentic" by the birds themselves. You can't just "make up" new syllables or sounds and have it be understood.