Repository navigation
Calm the voice and land each line on the beat - #11
Merged
Merged
Conversation
Aura-2 has no speed control, so house TTS speaks one line at a time and the deck places those phrases on beats 1 and 3 at 0.9x instead of stretching one breathless take across the bar. Co-authored-by: Kyle Alm <keepkalm@gmail.com>
Deploying with
|
| Status | Name | Latest Commit | Updated (UTC) |
|---|---|---|---|
| ✅ Deployment successful! View logs |
rapbattle | cfa3cec | Sep 30 2026, 05:55 PM |
keepkalm
marked this pull request as ready for review
September 30, 2026 17:50
submit_verse needs both beat_id (pocket bpm) and kind (warmup guard). New intro and warmup synth calls use the calm speaker fallback and the beat's bpm. Co-authored-by: Kyle Alm <keepkalm@gmail.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Spoken delivery was one breathless take, then the deck time-stretched that whole blob (up to 1.12×) to fill a run of bars. That pushed the voice and never put a line on the kick.
@cf/deepgram/aura-2-en) speaks one line at a time. The model has no speed field (speaker, encoding, container, sample rate, bit rate, text). A missing speaker falls back to Odysseus, which Deepgram describes as calm and smooth, instead of Zeus.Test
scripts/pocket.test.mjschecks that phrase starts fall on the kicks, that the browser copy of the scheduler matches, and that a fake AudioContext plays pocket audio at rate 1 from the next downbeat.