Skip to content

Commit 93029c5

Browse files
johnmatthewtennantnpub1mprnacetjua2xx3p5eddmhxyk6wv929ymm5py8kd2xfxurxahspqqlgyta
andauthored
feat(desktop): add Pocket voice settings (#3208)
## Context Pocket TTS can read live agent replies during a huddle, but Buzz does not provide an installation-wide way to turn that behavior off or choose among the bundled voices. The setting also needs a stable voice identity boundary that future backends and per-agent preferences can reuse without treating a display label as identity. ## Summary This adds a compact Pocket TTS voice settings surface with an Agent text to speech toggle, preview playback, and 12 bundled voices. Text to speech remains enabled by default with Mary selected. Disabling it immediately cancels queued and in-flight agent speech. ## Changes - Adds a versioned, device-local TTS settings schema and backend-qualified voice keys. - Bundles Anna, Azelma, Charles, Eponine, Eve, Fantine, George, Jane, Mary, Michael, Paul, and Vera with provenance and deterministic Mary fallback behavior. - Adds voice selection and preview controls while preserving the saved voice when text to speech is disabled. - Labels the visible and accessible selector as `Pocket TTS voice`. - Speaks only new live agent-authored text in active huddles, including kind 40002 agent replies, while excluding history, people, reactions, edits, system or tool rows, and attachment-only content. - Keeps agent speech strictly live-only across reconnects. The `limit: 0` subscription cannot page missed history or reorder replay against new live events. - Removes the unused `speakableAgentText` wrapper and replaces stale replay-override coverage with a live-only no-paging assertion. - Cancels stale, queued, and in-flight speech on barge-in, shutdown, or voice changes without appending stale audio. - Adds content-free route diagnostics, focused Rust and frontend coverage, and a Playwright screenshot flow. - Groups huddle audio settings into a cohesive state boundary and extracts relay timing constants so both existing files remain below their strict file-size ratchets. ## Related issue None found. ## Testing Manual validation in the combined daily-driver build confirmed all 12 bundled voice previews, one-time playback of a new live agent response through the selected output, and cancellation of stale audio after a voice change. ## Screenshots The global toggle, Pocket voice selector, and Preview action share one compact Voice settings surface. A focused screenshot was captured at `desktop/test-results/voice-settings/pocket-voices.png` (SHA-256 `a430ce71e7093612a0e9d27ebd3c6e43b2fa02b081c28376d8292e50f45d0a8b`). It is intentionally not posted. ## Reviewer-reproducible examples Exercise the bundled selector and live-only subscription contract from a fresh checkout: ```bash . ./bin/activate-hermit cd desktop pnpm build:e2e pnpm exec playwright test tests/e2e/voice-settings.spec.ts --project=smoke node --import ./test-loader.mjs --experimental-strip-types --test \ src/features/huddle/lib/ttsLiveMessages.test.mjs \ src/shared/api/relayReconnectReplay.test.mjs \ src/shared/api/relayChannelFilters.test.mjs \ src/features/settings/ui/voiceSettingsLogic.test.mjs ``` The selector renders 12 bundled voices under the accessible `Pocket TTS voice` label. The agent-speech subscription remains `limit: 0` and never requests missed-history paging. --------- Signed-off-by: John Tennant <[email protected]> Signed-off-by: John Tennant <[email protected]> Signed-off-by: John Tennant <[email protected]> Signed-off-by: npub1mprnacetjua2xx3p5eddmhxyk6wv929ymm5py8kd2xfxurxahspqqlgyta <d8473ee32b973aa31a21a65adddcc4b69cc2a8a4dee8121ecd51926e0cddbc02@buzz.block.builderlab.xyz> Co-authored-by: npub1mprnacetjua2xx3p5eddmhxyk6wv929ymm5py8kd2xfxurxahspqqlgyta <d8473ee32b973aa31a21a65adddcc4b69cc2a8a4dee8121ecd51926e0cddbc02@buzz.block.builderlab.xyz>
1 parent 4b9508c commit 93029c5

51 files changed

Lines changed: 4248 additions & 457 deletions

Some content is hidden

Large Commits have some content hidden by default. Use the searchbox below for content that may be hidden.

‎desktop/playwright.config.ts‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -49,6 +49,7 @@ export default defineConfig({
4949
"**/activity-scope-label-screenshots.spec.ts",
5050
"**/welcome-agent-modal-screenshots.spec.ts",
5151
"**/local-archive-screenshots.spec.ts",
52+
"**/voice-settings.spec.ts",
5253
"**/agent-readiness-screenshots.spec.ts",
5354
"**/agent-error-state-screenshots.spec.ts",
5455
"**/edit-agent.spec.ts",
Lines changed: 35 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,35 @@
1+
# Pocket TTS English VCTK presets
2+
3+
Buzz exposes Kyutai's twelve official English VCTK Pocket presets. The WAV
4+
bytes are unchanged from `kyutai/tts-voices` revision
5+
`323332d33f997de8394f24a193e1a76df720e01a`; only local filenames differ.
6+
7+
| Voice | Upstream asset | SHA-256 |
8+
| --- | --- | --- |
9+
| Anna | `vctk/p228_023_enhanced.wav` | `0a6de25cf12bf1540beb85979f306a92be81fecc051c547c5395e7e5237a3856` |
10+
| Vera | `vctk/p229_023_enhanced.wav` | `309cf91a895830f15842b398f69a4962cb1f7e0bfab10e25dd27838e826c204b` |
11+
| Fantine | `vctk/p244_023_enhanced.wav` | `5f07d4e2a3f20a15572aae885156b43ef3fc12ef3812996fd135680d9956448b` |
12+
| Charles | `vctk/p254_023_enhanced.wav` | `6b681a429198f16e378d53bccb08d06939da7b00144a7696111d4f8f76be7756` |
13+
| Paul | `vctk/p259_023_enhanced.wav` | `7aba504fe0b3b16478b69eb27ce6007e3cb42b0c1915b5f1c6a6024ae37d679b` |
14+
| Eponine | `vctk/p262_023_enhanced.wav` | `a13c27fb47627b05223691a0ef2974358a18c886e6c2f9d2762ff1d02c20926b` |
15+
| Azelma | `vctk/p303_023_enhanced.wav` | `60e3d26cdf2efdec5df712152c839928f4d5522821e6554ae11fd96c57ab1026` |
16+
| George | `vctk/p315_023_enhanced.wav` | `29a41f93bf5236e5b21501091d7774c255d5f3d4e62fa4f9fdf0a92a793c84ae` |
17+
| Mary | `vctk/p333_023_enhanced.wav` | `a35b0468382218e9f37a9a7494d1e4b74deaf18d7ced22265b4e325bb55c183f` |
18+
| Jane | `vctk/p339_023_enhanced.wav` | `2f12e7f155eb3118f55425394f1b049e5b1b67bdc9b3932c8ba4521420aeb84a` |
19+
| Michael | `vctk/p360_023_enhanced.wav` | `b6743e9195e5e3fd34fe9d1633ae93f7ffab787b249e45f6467d7d6f7a6ee6ad` |
20+
| Eve | `vctk/p361_023_enhanced.wav` | `396e7cbd066b0f3fb6d67fa26e7904076958239d736d4390f15b5fe88feb14cd` |
21+
22+
Mary is already installed as the Pocket model's `reference_sample.wav`, so it
23+
is not duplicated in this resource directory.
24+
25+
Source repository:
26+
https://huggingface.co/kyutai/tts-voices/tree/323332d33f997de8394f24a193e1a76df720e01a/vctk
27+
28+
The original recordings are from the Voice Cloning Toolkit (VCTK) corpus,
29+
licensed CC BY 4.0:
30+
https://datashare.ed.ac.uk/handle/10283/3443
31+
32+
The recordings were enhanced by ai-coustics:
33+
https://ai-coustics.com/
34+
35+
Neither Kyutai, the VCTK speakers, nor ai-coustics endorses Buzz.
786 KB
Binary file not shown.
805 KB
Binary file not shown.
624 KB
Binary file not shown.
700 KB
Binary file not shown.
656 KB
Binary file not shown.
659 KB
Binary file not shown.
628 KB
Binary file not shown.
742 KB
Binary file not shown.

0 commit comments

Comments
 (0)