[Cartesia] Upgrading the message cutoff for Cartesia Synthesizer to use timestamps #700
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Overview
Discussed with Ajay over Slack, recapping below.
There's a distinction between Cartesia's version of continuations and Vocode's general expectation for continuations. Vocode seems to expect an
N:N
ratio of senders to receivers, but our continuations is anN:1
ratio of senders to receivers. Vocode'sget_message_up_to
also reflects this expectedN:N
approach.The proposed solution to this is to start storing 2 new variables
self.ctx_message
andself.ctx_timestamps
. The Cartesia TTS now requests timestamps, and those timestamps are used to indicate what message we've gotten up to. In the event that the timestamps aren't available for some reason, or in the event that timestamps are delayed beyond a 2 second gap, we fall back to using an estimated wpm and theself.ctx_message
to get a best approximation.Testing
I set up the
telephony_app
per the Vocode directions. I then adjusted the synthesizer and played around with it locally on my own Vocode deployment to check that it works as intended.